> For the complete documentation index, see [llms.txt](https://help.tellius.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.tellius.com/tellius-6.4/whats-new/release-6.4.md).
# Release 6.4
Tellius 6.4 is here.
Kaiya now works in one chat. Ask a question, follow it step by step, and turn the result into a Business View, an App, a Mission or a file without leaving Kaiya. A user starting with nothing more than a data connection can go from that connection to a published App without leaving the chat.
Mission steps can be edited inline and applied together, keeping the plan yours through every revision.
Kaiya also works the way your team does. Custom skills set Kaiya's tone, report format and domain playbook once, and Kaiya follows them in every analysis, Mission and App. And the same conversation runs in Slack and Microsoft Teams, so work can start in a thread and continue in Tellius.
Administrators can now create custom roles and decide, module by module, what each role can reach. Kaiya usage is reported as cost, and Missions and Apps can read from and write to the systems you have connected. Tellius Mobile for iPhone, iPad and Android gets a new design with a dedicated iPad layout.
Read on to find more:
* **One chat for Kaiya.** Questions, Business Views, Apps and Missions all start from Kaiya chat, and Kaiya carries the context with the work as it changes direction.
* **From a data connection to a running App.** Connect a source, build and publish a Business View, then build, check and publish an App, in one conversation.
* **Files from any analysis.** Turn a result into a PowerPoint deck, a PDF, a Word document, an Excel workbook or a CSV file, change a single slide, repeat the analysis for another region in the same format, bundle the files into a zip, or send them by email.
* **Missions from the conversation.** Turn an analysis into a scheduled Mission, add steps and branches by asking, and publish it from the chat.
* **Missions and Apps that work together.** Build several Missions and the App that reads their results in one conversation, and change a metric across all of them with one request.
* **Custom skills everywhere.** A skill set once shapes every answer, Deep Insight, export, Mission and App.
* **Kaiya in Slack and Teams.** Each thread is a full Kaiya conversation, and View in Tellius opens the same conversation on the web.
* **Deep Insights as a conversation.** Start with one question and drill in turn by turn. Kaiya checks the data before it plans, asks when the data doesn’t support the question, and builds each answer on the last.
* **Inline editing of Mission steps.** Revise several steps and apply them in one pass.
* **Tellius as an MCP client.** Kaiya can query the MCP servers you have connected and save the result as an input step in a Deep Insight, Mission, or App. Missions and Apps can also write to connected systems such as Snowflake, under controls you set.
* **Snowflake Cortex inference.** Kaiya can run on models hosted in your own Snowflake account, so LLM consumption draws on your Snowflake credits.
* **Custom roles.** Create your own roles and set the access level for each module.
* **Kaiya usage shown as cost.** Kaiya usage is reported as cost, worked out from each model's rates.
* **Kaiya Context Model.** A versioned model of a data source that you build in the conversation, and that Kaiya, Apps and Missions answer from. It is built on live data sources, and support for non-live models and more sources is planned. It is rolling out gradually.
* **Voice mode.** Talk with Kaiya in the chat and hear short spoken answers, with the full answer saved in the conversation.
* **Skills from the conversation.** Create, edit, test and share custom skills by asking Kaiya, without leaving the chat.
* **Tellius Mobile.** A new design for iPhone, iPad and Android, built for iPad as well as phones.
Further detail on the rest of the release follows below.
## **🚀 New features**
### **One space for Kaiya**
Asking a question, building a Business View, creating an App, and building a Mission now happen in one chat. Kaiya keeps the context of the whole conversation, so a question can become an App, and the App a Mission, without starting again elsewhere.
Ask a question. When the answer is worth keeping, ask Kaiya for an App or a Mission. It opens in its own conversation in the Mission or App builder, with a summary of the context it carries over. To keep working in the same conversation, ask Kaiya to do so. You do not need to choose the surface up front.
After an answer, Kaiya offers the follow-up questions and actions that apply to that specific analysis, such as drilling into the underlying data, creating an App, creating a Mission, or exporting to PDF or PowerPoint. In an App build conversation they appear as two groups, follow-up questions and actions. Selecting a question asks it in the conversation, and selecting an action starts it. Follow-up questions show a short description and give you space to type your own answer.
The Missions and Apps pages remain, so you can still browse what exists. Editing a Mission or an App opens it in the chat. Opening a draft Mission from the Mission Library or the Home page takes you to the conversation it was built in, a published Mission opens its own page, and on an App's page the chat icon opens the App's conversation.
Once an analysis is done, ask for the same one on a different region or segment. Kaiya runs the same steps and returns the same outputs, in the same format. Where the value you name is not in the Business View, Kaiya asks which one you mean.
* **Clarifying questions when they matter.** Where a question is ambiguous, such as a quarter with no year, Kaiya asks before it answers, and you reply in the same conversation.
* **Answers in the right shape.** A one-line question gets a one-line answer. Kaiya returns a number, a sentence, a table or a chart, whichever the question needs, and checks the data first when the question calls for it.
* **Follow-ups that build on the last answer.** When you push back or add a condition, Kaiya adjusts the answer it already gave, so follow-ups return faster.
* **The Business View you chose.** Kaiya works only with the Business View you selected. When a question needs another Business View, Kaiya asks whether to switch to it or stay on the current one, and switches only after you confirm, unless the view it needs is soft-linked to the current one, in which case Kaiya adds it without asking. After a switch, the Business View picker shows the new view and follow-up questions use it.
* **Finding the right Business View.** When several Business Views could answer, Kaiya checks the Business Views you can access before asking you to choose, and avoids offering unrelated ones. Typing the name of a Business View you can access selects it, just as choosing it from the list does.
Share a conversation, and a colleague can continue the analysis with Remix. Super Users and Content Admins can open another user's conversation, including its files and Rich Summary, and select **Continue this conversation** to carry on in their own copy. The files also open for users the conversation is shared with.
**Conversations from earlier releases**
Conversations from before 6.4 open in Kaiya chat. Earlier answers, including charts and Deep Insight results, keep the format they were answered in and appear read only, with a link to the classic view for anything Kaiya chat does not display. New questions in the same conversation are answered in the 6.4 format, and Kaiya can still draw on the context of the earlier messages. Opening a Mission or App built in an earlier conversation continues it in Kaiya chat with its history, and a conversation linked to a Mission or App has an **Open Mission**, **Open App** or **Show Artifact** action to return to it.
Chat history for Kaiya, Missions and Apps is held in one view, so what you did in one place is available when you pick it up in another.
### **Plans and versions in Kaiya Apps**
Kaiya Apps show you the plan before anything is built, and keep the versions you publish.
#### **The Plan, laid out**
* The **Plan** tab presents an App as pages, sections, and analyses, with a running count of each. Pages can be expanded to show the sections they hold and collapsed again, and the analysis steps that feed the App are listed underneath. Each round of iteration updates the counts, so the effect of a change is visible before the App is built.
* The **Plan** also shows which analysis powers which part of the App. Each section is tied to the analysis step behind it, so before anything is built you can see where a given number on the page will come from.
* While you are revising an App that has already been published, the **Plan** on screen is the active version's. A new draft version is created at the point a new **Plan** is written, not before.
#### **Reviewing a plan**
A plan review has three outcomes:
* Approve it and the build proceeds.
* Cancel it and every tier is restored together (plan, data, views, and code) so an app is never left with a plan describing one thing and data describing another.
* Ask for a change, and the plan is revised and re-shown with the change named.
#### **Apps get publish-based versioning**
Changes are collected in a draft, and a version is created when you publish, either from chat or from the interface. A version can also be given a name of your own in place of the date-and-time label, which makes a particular version easier to identify later. Any version, active or not, can be renamed from its version menu, and it keeps its published state.
The most recently published version is the **Active** one. Moving to an earlier version is an explicit switch, and that version then becomes the active one shown in the App chat.
Kaiya prompts you to publish at logical checkpoints. The interface also offers the option once an App has produced its first screen, so a working state is captured rather than overwritten by the next edit.
### **Conversational Deep Insights**
Deep Insights now works as a conversation. Start with one question, then follow it where it leads: ask whether a change is worth worrying about, test what is behind it, or ask for a rough split of the causes. Each answer builds on the one before it, so you go deeper turn by turn instead of writing one long prompt.
* Where automatic Business View selection finds several strong candidates, Kaiya narrows the choice and continues the conversation with that Business View in context.
* Where a question calls for a forecast or another calculation, Kaiya runs it in Python and shows the assumptions behind it.
* At any point you can ask Kaiya to summarize what you found, export it, or repeat the same analysis for another region.
* A Rich Summary appears as a **Visual summary** card in the conversation. While the summary is being built, the card reads "Preparing summary…", and you can select it as soon as the summary is ready. The side panel opens only when you select the card, including for export cards and when you return to a conversation.
* Deep Insight progress is shown in one place while the analysis runs.
**Before planning.** The agent establishes what the data actually holds. It checks whether the columns the question needs are present, and what values they carry, and for a comparison it checks that the data being compared exists. This step appears in the progress messages. Where the data is missing, or several columns could satisfy the question, it asks you clarifying questions rather than choosing on your behalf.
**While planning.** The plan is built one step at a time and appears as each step is ready, rather than arriving complete at the end. A longer analysis therefore starts showing its shape sooner.
**Before running.** The finished plan is reviewed for analytical soundness. Where a problem is found, it is revised and reviewed again before anything executes.
**While running.** Where a step fails, only that step is rewritten. The rest of the plan is kept and the analysis continues from there.
**After running.** The results are reviewed against the question that was asked. Where they do not answer it, the analysis is revised.
### **Exports from Kaiya and Missions**
* Ask Kaiya, in the conversation or in a Deep Insight, to export the analysis as a PowerPoint deck, PDF report, Word document, HTML page, Excel workbook or CSV file. If you don't name a format, Kaiya asks which one you want. Decks, PDFs and Word documents carry your configured logo.
* When a result holds more than one table, an Excel export puts each table on its own sheet, and for a CSV Kaiya asks which table to export.
* CSV and Excel files open in an in-app preview. Excel previews show sheet tabs and zoom. CSV previews show the first 1,000 rows, with a note to download the full file.
* Markdown and text files that Kaiya creates, such as an exported Mission plan, also preview in the app, and you can download them.
* A Mission can include several export steps, so one run can deliver a PDF report together with several CSV files. Export steps produce PowerPoint, PDF, Word, Excel or CSV files, and Mission Preview shows every file the run produced.
* Export is offered once any clarification is answered and the Deep Insight completes. While an export is being prepared, the Export action is disabled, so each request is exported once.
### **Custom skills shape every part of Kaiya**
A custom skill now applies to everything Kaiya does: answering questions, Deep Insights, exports, Business Views, Missions and Apps. A team writes a skill once, and Kaiya follows it in every conversation without the instruction being repeated.
Teams use custom skills to:
* **Set tone and audience.** For example, "Write for business users. Keep it short and leave out technical detail."
* **Fix a house report format.** For example, an executive summary first, then the findings, then next steps.
* **Add a domain playbook.** For example, "When TRx drops, check payer mix, then territory, then top HCPs."
* **Decide how Kaiya handles ambiguity.** Ask Kaiya to always ask clarifying questions before it assumes, or to proceed on stated assumptions and let you iterate.
* **Tune or replace a default.** Replace the default Deep Insights skill with your own approach.
Where a custom skill and Kaiya's default behavior differ, the custom skill takes priority. A **Global** custom skill applies to every user in the workspace, and a **User** custom skill applies only to the person who created it.
#### **Create and manage skills in the conversation**
You can now create and manage custom skills by asking Kaiya in any conversation, without leaving the chat. Describe the rule in plain words. Kaiya asks a few short questions, drafts the skill and shows it on a card before it saves anything.
* **Create, edit, test, unpublish and delete.** Ask Kaiya in your own words, or send a short request such as `/skills`. Unpublishing a skill keeps it as a draft but stops Kaiya from loading it. Deleting a skill is permanent.
* **Personal skills.** A skill you create for yourself applies only to you.
* **Global skills.** Users whose role has Full Access to Manage workspace, or any access to Content management (by default Super Users, Admins and Content Admins), can make a personal skill global, so it applies to everyone, or make a global skill personal again. Before you approve, Kaiya says who gains or loses the skill.
* **Domain skills.** Create a domain skill for a Business View from the chat. A domain skill applies only to questions about the Business Views that use it.
* **Every change asks first.** Kaiya shows each change on a card and saves it only after you approve. Saved changes apply from your next question.
* **A description that says when to use it.** Kaiya decides which skills to load from their descriptions. A skill's description must say when the skill applies. If it doesn't, Kaiya suggests better wording and doesn't save the skill.
See [Kaiya Skills](https://help.tellius.com/kaiya/kaiya-skills) to create one, and [Managing skills in the conversation](https://help.tellius.com/kaiya/kaiya-skills/managing-skills-in-the-conversation) to create, change, unpublish, promote and demote skills without leaving Kaiya chat.
### **Kaiya in Slack and Microsoft Teams runs the same conversation**
Each Slack or Teams thread with the Kaiya bot is a full Kaiya conversation, with the same skills, context and capabilities as Kaiya on the web.
* **Work back and forth in the thread.** Ask follow-ups in the thread. Clarifying questions arrive as cards that you answer in place.
* **Get files where you asked.** Ask for a deck or a report, and the file is attached to the thread in Slack. In Teams, the file opens from a link.
* **Build from the thread.** Build, test, publish and schedule Missions and Apps without leaving Slack or Teams.
* **Continue in Tellius.** **View in Tellius** opens the same conversation on the web. When you come back to the thread, Kaiya posts a short summary of what was done on the web.
* **Ask about a Mission's output.** When a Mission sends its output to the bot, ask follow-up questions about it in the same thread.
* **Bring the bot into a channel.** Add the bot to a channel or group, and teammates can ask it to check a number on the spot.
With an MCP server connected for your account, Kaiya can also draft an email or create a calendar event from the thread.
To set up the bot, see [Slack Integration](https://help.tellius.com/settings/administration/slack-integration) and [Teams Integration](https://help.tellius.com/settings/administration/teams-integration).
### **Voice mode in Kaiya chat**
Talk with Kaiya in Kaiya chat. Voice mode opens over the conversation, listens for your question and answers aloud, using the Business View selected in the chat.
* **Short spoken answers, full answer in the chat.** Kaiya speaks a short summary with the key numbers and offers a follow-up. The full answer, with its charts and **How Kaiya got this**, builds in the conversation at the same time.
* **Follow-ups keep the context.** Ask a follow-up and Kaiya builds on the answers before it.
* **Mute when you need to.** Mute the microphone while Kaiya answers, and unmute it to ask the next question. A transcript shows what you said and what Kaiya said.
* **Every turn is saved.** Select **Exit Voice Mode** to return to the chat. Each spoken question and answer stays in the conversation as a normal message, so you can read the full answers later.
Voice mode uses OpenAI voice. It appears in Kaiya chat when an administrator turns on the Voice Assistant with OpenAI as the provider in **Settings → Intelligence → Kaiya & AI → Kaiya Configuration**, using an OpenAI key with Realtime access. Apps and Missions can't be created by voice, so type the request in the chat instead.
### **Missions and Apps that work together**
In one conversation you can build several Missions and an App that reads their results. Each Mission runs on its schedule and saves its output to a table. The App reads from that table, so it shows the latest results after each run.
* **One request updates every part.** Ask Kaiya to change a metric, for example "use Gross Sales instead of Sales", and it updates each Mission and the App that uses it, changing only what needs to change.
* **Every run stays in the conversation.** Scheduled runs are recorded in the same conversation, so you can ask, "Compare this week's report with last week's."
* **Start small or go big.** Begin with one Mission feeding one App, or combine several Missions that each own one part of a larger job.
### **Build, check and refine an App in the conversation**
An App is now built in the same conversation as the analysis behind it. You can start from a question, from an existing analysis, or from a goal with nothing more than a data connection.
* **From a goal to a published App.** Describe what you need. Kaiya connects to your warehouse from the chat, asks whether the data should be live, builds and checks the Business View, and builds the App on it.
* **Check the numbers before you share.** Ask Kaiya to verify the App, and it checks each figure, reports any problem it finds, and fixes it in the draft.
* **Checks the App the way you see it.** After Kaiya saves an App, it views the App screen by screen and tab by tab, and checks for blank values, charts that drew nothing, overlapping charts and content that overflows the page, so it can correct what it finds in the same turn.
* **Match a design you already have.** Attach an image or an HTML file, and Kaiya restyles the App to match it. Ask it to reconcile the App's numbers against the prototype, and it explains where and why they differ.
* **Explain the App.** Ask for a summary, and Kaiya writes a guide to the App's tabs, what each one shows, and how to drill down.
* **Turn an App into a Mission.** Ask for the same analysis on a schedule, and Kaiya builds a Mission that delivers it as a PDF.
* **Query the data directly.** Kaiya can run its own queries to explore what the data holds or test an assumption, so a question about your data can be answered without building anything.
* **Knows when not to build.** Kaiya holds guidance on what a Live App is and when one is not the right choice, so it can raise a concern rather than build something unsuitable.
* **Builds in small steps.** Kaiya builds and updates an App in small steps, so a single request can add new parts and change existing ones at the same time.
* **Repairs what it can.** Where a step fails (in the Deep Insights analysis or in code generation) Kaiya attempts to resolve it before involving you. When an App preview can't render or load its data, Kaiya tries to fix the error on its own a limited number of times. If it still can't, it shows **Unable to automatically resolve this issue.** Select **Fix with Kaiya** on the banner to ask Kaiya to fix it in the conversation. It works on the draft, or on the active version when the App has no draft.
* **The finished App, on its own.** When Kaiya finishes building an App in the conversation, the side panel shows the App on its own. A later change that creates a new version shows a notice with an option to view it.
* **Apps you already have.** On the page of an App built in an earlier release, select **Open app assistant** to keep working on it in Kaiya chat, with the App beside the conversation and its earlier chat history carried over. Your edits are saved as a draft version, and the published version stays live until you publish.
### **A new workspace for building, testing, and publishing a Mission**
Building a Mission now happens in a single, redesigned screen. Kaiya sits on the left as a conversation, and the Mission is on the right across four tabs: **Steps, Preview, Triggers,** and **Schedule.** Kaiya can now also detect the right Business View from the request itself, and the selector remains available when you want to set it yourself.
* The **Steps** tab shows the plan Kaiya built from your request as ordered SQL, Python and summary steps you can edit, in two views. **Outline** lists the steps in order, which suits reading a plan from start to finish.
* **Diagram** lays the same plan out as a map, drawing an arrow from each step to the steps that consume its result and naming the output each one produces. A Mission that branches, or one where a Python step draws on two earlier queries, is more readily followed in this form.
* The **Preview** tab goes further than showing the finished artifact. It runs the plan and shows what each individual step returned, before you publish anything. Each planned step shows its title, SQL or Python code, status, row and column counts, a five-row preview and run time, and the generated summary appears underneath in full. Files a step exports, including HTML reports, appear as cards in that step, open in the Artifact view, and stay available after you reload the page. You can therefore check that a Mission produces the right answer while it is still a draft.
* **Citations in Preview.** Open citations to see where each number in the output came from or how it was derived, with each figure marked **Verified** or **No source**. Citations are generated only when you request them in Mission Preview, never during scheduled runs.
* The name, description, triggers, and schedule can all be edited from the conversation rather than the tabs.
* Asking Kaiya to send the report to a colleague adds that recipient, and asking it to also send to Slack or Teams enables that channel. Kaiya confirms each change in the thread. The description stays auto-generated and current as the Mission evolves.
#### **Build and change a Mission in the conversation**
* **Build a Mission from an analysis.** When a set of questions is worth repeating, ask Kaiya to turn it into a Mission. It carries the steps over and adds an input step where you want to choose a value at run time, such as a region.
* **Change it by asking.** Ask for a new finding and Kaiya adds the steps it needs. Edits touch only the steps that changed.
* **Branch on a result.** Ask for a different summary for each outcome, such as below 70% of goal, between 70% and 90%, and above 90%.
* **Check before you schedule.** Ask Kaiya to confirm the numbers before the Mission goes live. Kaiya runs its own queries to check them and traces where a difference comes from. Where a step looks wrong, Kaiya explains what it found, offers a fix, and can update the steps so the output matches what you asked for. You can ask Kaiya to check the numbers in any answer the same way.
* **Export in your format.** Describe the layout in the conversation, attach a screenshot of it, or use a custom skill such as a house deck format, and the Mission's export follows it.
* **Schedule and publish from the chat.** "Schedule it for Fridays at 4 PM and publish" does both.
* **Handle a renamed column.** When a source column is renamed, the Mission re-reads the schema and adapts, or tells you what changed.
#### **Missions you already have**
* Missions built in earlier releases keep working without being rebuilt. Their schedules keep running, and their run history stays on the Mission page.
* **Configure** on a Mission's page opens the Mission in Kaiya chat, with its earlier conversation and the Mission beside it.
* Changes are made in a draft. The active published version shows as **Locked** and can't be edited. Until you publish the draft, every run and schedule uses the active version.
* Publishing makes the draft the active version. The Mission's existing schedule uses it from the next run, so you don't need to create the schedule again.
* New Missions end with a Rich Summary by default. A run gives a short summary in the conversation and the full visual report as a **Visual summary**. For a Mission that ends with a text summary step, we recommend asking Kaiya to replace that step with a Rich Summary.
{% hint style="info" %}
**Export step or custom skill?** Use a custom skill for a house style that every export should follow. For a file your team relies on, write its storyline, contents and design into the Mission's export step, so every run produces the same file. When the export step's instructions differ from a custom skill, the export step takes priority.
{% endhint %}
#### **Debug the Mission from the conversation**
The Debug Assistant helps you work out why the Mission you are building behaves the way it does. In 6.4 you reach it inline: ask a debug question directly in the Mission conversation, without opening a separate mode.
* Replies in debug mode are not saved to your Mission and do not affect the planner, so a draft can be examined without altering it.
* Debug assistant opens with a set of starting questions covering the checks most often needed: whether the output is correct and which step is responsible, why a run is slow, whether the figures can be relied on, and how Kaiya approached the analysis. You can also ask your own question about the draft.
* Asking a question of a finished Mission run shows the working behind the reply. It lists the steps that ran, the query behind each one, and what it returned. A response that does not match expectations can be traced to the step responsible.
#### **Every published version is kept**
* A Mission keeps every version you publish.
* A selector in the header lists them with the date and time each went live, and marks one as Active.
* You choose which version is active, and that's the one every run uses, whether you start it yourself or it runs on a schedule. Edits are made on the draft. Because earlier versions are retained rather than replaced, you can see how a Mission changed over time and which version produced a given briefing.
* A branched Mission can end with a summary or an artifact for each path, so each branch reports its own finding.
* Run history is available from the Mission header. It lists every run with its date, duration, and how it was started, alongside the total number of runs, the success rate, and the average duration.
### **Tellius as an MCP client**
The Tellius MCP server, introduced in 6.3, lets an external client such as *Claude Desktop* or *Cursor* call into Tellius. 6.4 adds the opposite direction: Tellius is now an MCP client. Kaiya can query the MCP servers you have connected and bring their data into an analysis.
* **An MCP server as an input step:** Until now, the only way to bring data into a Deep Insight, Mission, or App was a SQL query against your data sources. A connected MCP server can now do that job as well.
* **Kaiya queries the server directly and saves what comes back.** This saved result becomes a step in the plan, and the steps after it read from that step. Saving the result, rather than holding it in the conversation, is what allows a large set of records to be used. A Mission can pull the open opportunity list from your CRM, and a later step can summarise that list or write it into an export.
* **Missions work with your connected systems.** Missions can read from and write to the systems you have connected, such as Snowflake, Slack, Salesforce, Gmail, Notion, Microsoft 365 or an MCP server your administrator adds. A connected-system step sits in the Mission plan beside SQL and Python steps. It can pull live data into the Mission or write results out, for example by inserting rows into a Snowflake table. Each step card shows what was asked, the calls that were made, and whether the write succeeded.
* **Writing back:** In the conversation, every write is presented for review and executed only once approved. In a Mission, writing is off by default, and the Mission owner turns it on for each step in **Manage access** with **Allow it to send, post and create**. **Run as** sets whether the step uses the connection of whoever runs the Mission (the default) or always the Mission owner's. In 6.4, scheduled runs can read from connected systems but can't write to them yet, and a step set to run as the owner works only when the owner runs it.
* **Apps work with your connected systems.** Apps can read data from your connected systems and write back through buttons in the App, for example to create a record or a calendar event. Each write is confirmed on screen before it runs, and an App action that runs under its creator's connection needs approval first. Apps can also use unstructured data, such as indexed documents, as a step.
* **Deep Insights read from connected systems.** Deep Insights can also read from connected systems, and the **MCPs** tab under **Settings → System → Administration → Integrations** shows the callback address to register with your provider.
* **Microsoft 365 with one switch.** On the **MCPs** tab, the **Microsoft 365** connector turns on Mail, Calendar, Teams, SharePoint, OneDrive, People and Search together. It doesn't need a Microsoft 365 Copilot license.
* **Clear handling when a server is not connected:** Asking Kaiya to search a system you have not connected or authorised returns a message asking you to connect it.
### **Snowflake Cortex inference**
Kaiya runs on models from OpenAI, Anthropic, Google and other providers. It now also runs on models hosted in Snowflake through Cortex Inference.
For teams that would rather keep model execution inside their own Snowflake account, Kaiya can be configured to use it. LLM consumption then draws on Snowflake credits already committed, rather than a separate model spend.
### **A redesigned Context page**
The Kaiya Context page is organized into two layers, **Skills** and **Memory**, each shown as a card with its count. A panel shows how many phrase learnings have moved to Memory and how many remain. Beside it, **Support details** lists the last attempt, the number of attempts, the reason and a reference, with **Copy details** and **Contact support**. A read-only **Legacy learnings** tab appears while phrase or query learnings remain outside Memory.
### **Kaiya Context Model**
A Kaiya Context Model is a versioned model of a data source that Kaiya answers from. It holds the source's tables, the relationships between them, metrics written in the source's own SQL, business vocabulary, rules such as a default period, and reusable analyses. The Kaiya Context Model is rolling out gradually.
* **Built in the conversation.** Ask Kaiya to build a context model. Kaiya asks which data source and tables to use, studies the data, and shows you a plan to approve before it saves.
* **No publish step.** A context model doesn't need a publish job. A save is live straight away, and Kaiya answers on it from your next question.
* **Versions.** Each save creates a new version, and you can compare a version with the one before it to see what changed.
* **Apps and Missions.** Apps and Missions can use a context model as their source.
* **Learns from conversations.** As people ask questions, Kaiya notes the assumptions it had to make, such as what a metric means. Once a day, Kaiya turns them into context model suggestions, and the model's owner, or someone who can edit the model, approves the suggestions and publishes them as a new version. Kaiya uses only published changes.
* **Where to find it.** Context models are listed on the **Context Models** tab of the Data module, where **New context model** starts one. You can also pick a context model in the data picker of the Kaiya chat box and of the home page ask box.
* **Who can use it.** Each role decides, with the **Context Models** module in its Kaiya group. **Read Access** lets people open context models and ask questions on them. **Full Access** also lets them create, edit, share and delete models and decide their suggestions. The module needs **Chat** at **Read Access** or higher. It replaces the **Enable Kaiya Context Models** setting of earlier Tellius 6.4 builds, and each role keeps the access that setting gave it. Suggestions are off by default, and a Super User turns them on in Kaiya Configuration with **Learn from conversations (context model suggestions)**.
In Tellius 6.4, a context model is built on a live data source. It queries the data source directly, so answers reflect the current data. Support for non-live models and more sources is planned. Vizpads don't use context models. See [Kaiya Context Model](https://help.tellius.com/data/kaiya-context-model) to build and manage one.
{% hint style="info" %}
**Context model or Business View?** Use a context model when Kaiya chat, Apps and Missions answer from your data and you want the meaning of the data kept in one reviewed place. Keep using a Business View for Vizpads.
{% endhint %}
### **Content Admin role**
A Content Admin role gives designated users standing read and edit access to all Datasets, Business Views, Vizpads and Kaiya assets, including content that already exists.
A Content Admin can always see and manage content across the platform, so when someone leaves or changes teams, their content is reassigned and keeps running.
Content Admins can also create folders and import Business Views and Datasets, use dataset tools such as scripting, data fusion, schedules, and Python and SQL transforms on other users' datasets, export those datasets, and see all users' content on the Connect and Refresh Pipeline tabs. "Author" filters are renamed "User".
### **Custom roles**
Administrators can now create their own roles and set, module by module, what each role can reach. Roles are managed under **Settings → Workspace → Users & Access → Roles**. Select **Create role**, enter a **Role name** and **Description**, and set the access level for each module.
* **No Access.** The module is not directly available to the user.
* **Read Access.** The user can view and use existing content without modifying it.
* **Full Access.** The user can use all supported functionality in the module.
Modules are grouped into Kaiya (Chat, Apps, Missions, Context, Context Models), Analytics (Vizpad, Predict, Insight, Search, Feed), Projects, Collaboration (Sharing), Data (Datasets, Business Views, Data download, Connect source) and Administration (Own settings, Manage workspace, Manage system, Content management). Sharing, Data download, Connect source, Manage workspace, Manage system and Content management offer No Access or Full Access only. Own settings is fixed at Full Access for every role.
* A role's module access can only restrict. It narrows what the role can already do and never adds to it.
* When you turn on a module that needs another module to work, the Roles screen raises that module to the level required and locks it, with a note naming the module that needs it. For example, turning on Missions raises Chat, and Chat in turn raises Datasets and Business Views.
* Assign a role from the **Role** dropdown when you edit a user, which lists custom roles alongside the built-in ones, or assign it to a group. A new group starts with no role.
* When a user receives access from more than one assignment, such as their own role and a group's role, the highest level for each module applies.
* The Roles page lists each role with its type, how many modules it restricts (or **All modules**), and how many users and groups hold it. Built-in roles show as **System**.
* The **View effective access** dialog shows each module's level for a user and the role that granted it. Open it from **My Profile**, beside **Roles in effect**, or from a user's row menu on the Users page.
* A custom role can be the default role for single sign-on. Who can share content and who can use Kaiya Context is set for each role with the Sharing and Context modules, which replace the earlier Kaiya Context settings. A new custom role starts with No Access to both, so give it access as needed.
* Changes to a role apply the next time the user loads a page.
* The built-in Interactive, Explorer, Discoverer, Explore & Discover and View roles don't include **Data download**, so their users can't download the raw data behind a dataset or Business View. Exporting a chart, Vizpad or Insight isn't affected. An administrator can turn **Data download** back on for a role.
### **Kaiya usage shown as cost**
Kaiya usage is now reported as cost in place of credits. Tellius records the model and token counts for each Kaiya call and works out cost from each model's rates when you view the report, so a corrected rate also updates past cost. Find it under **Settings → Intelligence → Kaiya & AI → Kaiya Cost**.
* The **Monthly Usage** tab shows total cost, cost by area and by model, and lists any models that have no price yet, so you can add their rates.
* On the **Model Pricing** tab, Super Users can add a model, set its input, cached input, cache write and output rates per million tokens, and add other names for the same model.
* Tellius includes default rates for common Anthropic, OpenAI and Gemini models.
### **Cross-datasource swap**
Datasource swap now works across datasource types, such as Snowflake to Redshift, Redshift to Snowflake, or MySQL to Microsoft SQL Server, as well as between sources of the same type, and is designed to accommodate further combinations.
Schema compatibility is validated before the swap completes. Where the target lacks a required column or a data type differs, a clear message describes the mismatches so they can be resolved. A force-swap path is available where a mismatch is understood and accepted.
Tellius asks for the details the target source needs. For a database source, you choose a load method: DB Table with the table name, or Custom SQL. For AWS sources, you enter a storage path. A tooltip notes that Vizpads, Insights and Predict models built on the dataset may be affected.
### **Tellius Mobile**
Tellius Mobile is the Tellius app for iPhone, iPad and Android. Sign in to your Tellius workspace with your username and password or your company's single sign-on, then ask Kaiya about your data and follow your Missions and Apps from your phone or tablet. The app also supports Face ID or fingerprint app lock, voice dictation, sharing with people and groups, and reading recent chats while offline.
This release brings the following to Tellius Mobile.
* **A new design.** The app is organized into three tabs: **Today** for what's new, **Library** for Missions and Apps, and **History** for every conversation. Loading placeholders appear while content arrives, and a confirmation appears when you save.
* **Built for iPad.** On iPad, a sidebar or icon rail sits beside every screen. Missions and Apps open next to their list, answers open on a canvas beside the conversation, and History uses a table layout.
* **Five languages.** The app is available in English, Spanish, German, French and Portuguese (Brazil). It follows the language set in your profile and updates every screen at once. You can edit your name and language from your profile in the app.
* **Today.** Today shows your latest briefing first, tells you when you're caught up, and lets you pick up where you left off. One ask box handles questions and builds, with suggestions to start from. Choose **Quick answer**, **Deep Insight**, **Build a mission** or **Build an app**, and choose the data to use.
* **Talk with Kaiya.** Kaiya reads its answer aloud, then listens for your next question, so you can keep the conversation going hands-free. Mute, stop Kaiya speaking or end the conversation at any time. Speech is processed on the device.
* **Answers.** Switch a chart between Bar, Line, Pie and Table. Wide tables scroll sideways rather than squeezing their columns. Visual summaries, Kaiya's interactive chart pages, open in the app.
* **Files from answers.** Reports, decks and spreadsheets Kaiya creates appear under the answer that made them, with their versions, such as "PDF export · Version 2". Every file in a chat is also listed in **Files in this chat**.
* **Attachments.** Photos and files you attach to a question appear as pictures and file cards, including when you reopen the chat. Tap one to open it.
* **Briefings.** Each Mission run has its own briefing page with Kaiya's thought process, the summary, an interactive visual summary and every file the run produced. Open it from Today, from a Mission's run history or from a notification. A run picker lets you move between runs and shows the success rate and average run time.
* **Missions.** The Library lists Published and Draft Missions, with search, sort and an access filter. A Mission's page shows its status and stats, with **Run now** and **Publish** or **Unpublish**, and lists its steps so you can see what it does.
* **Scheduled missions.** See every schedule you created or that is shared with you in one place, filter by **All**, **Created by me** or **Shared with me**, and search. You can subscribe, edit or delete a schedule. Open it from **Library → Missions** or from **See all** on a Mission.
* **History.** History lists the same conversations as the web: chats, App builds, Mission builds and Mission runs, newest first. The **Show** filter offers **All chats**, **Favorites**, **My chats** and **Shared with me**, the **Type** filter offers **Chat**, **Mission** and **App**, and one tap resets both. A chat you star shows as a favorite everywhere.
* **What Kaiya remembers.** You can now edit the text of a memory, as well as add, turn off or forget one.
* **Workspaces and sign-in.** Stay signed in to several workspaces and switch between them. Password managers, such as Bitwarden, iCloud Keychain and Google, can fill in your sign-in.
* **Links.** On Android, shared Tellius links to chats, Missions, Apps and Home open in the app. On iPhone and iPad, paste a Tellius link into the ask box to open it. When a link belongs to another workspace, the app offers to switch workspaces, open the link in your browser or cancel.
* **Account.** You can request deletion of your account from the app.
## 📈 Enhancements
### **Kaiya**
1. Summaries answer the question asked. The Summary Agent now answers what was asked and stops. Where data is not found, that is stated clearly at the top of the response rather than within it. Database statistics, SQL fragments, and format analysis are kept out of end-user summaries.
2. Metadata questions without a Business View selected. Questions about the data itself are now answered when no Business View has been selected. Asking what data you have access to returns a summary of the Business Views available to you. Asking what metrics a particular Business View holds returns those metrics, without needing to select it first.
3. Upload a document or image to Kaiya chat. A file can be attached to a Kaiya conversation and asked about directly.
Kaiya decides where each question about the file should go (structured, unstructured, or a Deep Insight) based on what is being asked, so an uploaded file and your Business View data can be reached in the same conversation.
Every place you can attach files in Kaiya accepts the same file types, including the chat composer, the Home page composer, the Insight Apps sidebar and Kaiya Architect chat: images, PDF, Word (DOCX), Excel (XLSX and XLS), PowerPoint (PPTX), and CSV, TXT, LOG, JSON, HTML, HTM and MD files, up to 20 MB each. Each file you attach appears in the message you sent, including after you reload the page or reopen the conversation. Kaiya reads the text in attached PDF, Word, PowerPoint and HTML files, and reads each sheet of a spreadsheet separately.
4. Deep Insights across multiple Business Views. A Deep Insight scoped to several Business Views runs across all of them in one analysis.
5. Row-level permissions across multiple group memberships. A user can belong to several groups, and those groups can each set the same Security Tag to a different value. That combination now resolves to a query that runs, instead of one that stops before it reaches the data.
6. Multi-fact Business Views on In-Memory. An In-Memory Business View is no longer flattened into a single joined table when it is published. Its datasets are kept as separate tables and joined at the point a question is asked.
Because the datasets stay separate and are joined per question, a Business View holding several fact tables can now be used in memory.
Existing In-Memory Business Views are converted as they refresh. Nothing changes on upgrade; a Business View moves to the new shape the next time it is published or refreshed on its normal schedule.
7. Time filters reach the query. A date range detected in a Deep Insights question is now passed through to the steps that follow. The query therefore filters on the range you asked about, rather than one fixed when the plan was written.
8. Kaiya Credits replaced by Kaiya Cost. In 6.4, Kaiya Credits and My Kaiya Credits are retired and replaced by Kaiya Cost, and links to the old credit pages open Kaiya Cost. Cost limits count usage for the calendar month to date.
9. Memory is saved when you choose to save it. Kaiya's memory remains explicit, with nothing stored in the background. A correction or a clarification in chat now offers a card to save it, so you decide what is kept. You can also ask Kaiya to remember something directly. Duplicate entries are consolidated rather than accumulating, the memories applied to a given answer are shown in **How Kaiya got this**, and a memory agent can report what is stored. Kaiya chat uses your saved memories across conversations: with each message it looks for relevant memories, and it asks for your approval before saving a new one. Kaiya can also suggest a takeaway from the conversation to save to Memory.
10. Recent items reflect active work. The Recent list is built from meaningful interactions such as editing or actively working with a conversation, rather than from every conversation that was opened and viewed.
11. Built-in instructions shown as skills. The System Prompts settings page is retired. Kaiya's built-in instructions are shown, read-only, as Unified Chat Skills on the Skills tab of the Context page, where Super Users can see them. Custom prompts are not listed in the interface.
12. Concurrent Skills edits. Two people, or one person in two browser tabs, can now edit Skills on the Context page without losing work. If a second save is based on an older version of the skill, it is reported rather than replacing the newer one.
13. Stop and continue. Stop a response while Kaiya is working, and the message shows **Response stopped**. You can continue the conversation later.
14. Message times and copy. Hover over a message to see when it was sent, such as "7 minutes ago", and hover over the time to see the exact time. A copy icon sits beside it for your own messages and for Kaiya's answers. Copying an answer copies all of it, and the icon briefly shows a check mark.
15. How Kaiya got this. The Thought Process card in Kaiya chat, including App and Mission conversations, is renamed **How Kaiya got this**. While Kaiya works, the card stays closed and shows one live status line with a step count. When Kaiya finishes, it shows **Done** and the total time. Open it to see Kaiya's reasoning and the steps it took in plain words, such as reading the data's structure, querying the data and reviewing the result. Switch on **Detailed view** to see every step with its time, including the schema it read, the skills it loaded, the queries it ran, and the reviewer's verdict on the result with a **Read full review** link. Your personalization no longer adds a notice under every answer. When your preferences changed an answer, **Detailed view** explains what changed.
16. Refined how Kaiya writes its answers, so notes about its next steps stay out of the answer and the answer starts with the results.
17. Periods from your calendar. In Kaiya chat and in Missions, Kaiya works out periods such as "this week" or "last month" from the calendar set on the Business View, including fiscal and custom calendars, and measures them from the latest date in the data. No new setup is needed. If the calendar doesn't cover the period you asked about, Kaiya says so in the answer, with both date ranges and the week convention it used instead.
18. Weekly results label each week with its week-ending date, such as Dec 01, 2024, and list weeks in date order. To see ISO week numbers such as 2024-W01, ask for them or add a custom skill.
19. Citations in plain language. Select a citation number in an answer to open its details. Each citation starts with a plain-language summary of the figure: the metric and how it is calculated, the filters, the level of detail and the source. Below the summary, **How it was calculated** shows the SQL or Python code, followed by the data behind the figure.
20. Personalization. Administrators set up Personalization Questions under **Settings → Intelligence → Kaiya & AI → Personalization Questions**. Add a question by describing it, then edit its header, helper text, options, pre-ticked answers, order and follow-up conditions, or switch a question off. After a few turns in the chat, Kaiya shows the questions on a card, and you can answer them, choose **Not now**, or choose **Don't ask again**. Your answers apply to the whole conversation as background context, separate from skills and memory. See [Personalization Questions](https://help.tellius.com/settings/kaiya-and-ai/personalization-questions).
21. Ask across your documents. Any Kaiya conversation, not only a Deep Insight, can search your indexed document workspaces alongside your Business View data. When it needs more, Kaiya searches again or reads a whole document. It cites each claim to the passage it came from, and every search uses your own document access, so answers draw only on documents you're allowed to see. See [Ask across your documents](https://help.tellius.com/kaiya/ask-across-your-documents-unstructured).
### **Missions**
1. Briefings for manual runs. Briefings now appear for manually triggered runs as well as scheduled ones. This covers a run started from the Run button and a run started from a Kaiya conversation. A Mission's output history is therefore complete regardless of how each run began.
2. Provenance in chat. When a Kaiya answer came from a Mission, the answer names the Mission, the version that ran, and how it was started (Scheduled, Chat, or Run manually). The Mission name and an arrow control open that Mission directly.
3. Inline editing of Mission steps. Every part of a Mission plan can be edited in place, and your edits are applied together when you are ready rather than one at a time as you type.
Collecting the edits and applying them in a single pass keeps the Mission under your control while you shape it. A plan needing many small corrections is revised as one change, and the apply cycle runs once rather than once per field.
4. Input values at run time. When a Mission with input steps runs from the conversation, Kaiya asks you for each value, such as a region or a threshold, and offers the step's default as the first option. Selecting **Run** still asks for these values. Scheduled runs use the value saved with the schedule, or the default.
5. Run a Mission by asking Kaiya for it by name, or select **Run** on the Mission listing or detail page, the Briefing page or a Mission Library card. **Run** opens a new conversation with the Mission details and runs the published Mission straight away. While it runs, the conversation shows **Working…** with live progress.
6. Mission schedule times are shown in your browser's time zone, including when Kaiya confirms a schedule or reports a run time in the conversation.
7. On the Scheduled Missions page, the Mission and User filters load more results as you scroll, and search and A to Z sorting apply across all Missions and users. The Mission filter lists scheduled Missions, and choosing a user narrows it to that user's Missions. The Mission picker in the schedule form also loads more Missions as you scroll.
8. When you create a schedule from **Missions → Schedules**, the Mission dropdown lists published Missions from the Modern view, both your own and those shared with you.
9. In a Mission's Business View dropdown, the selected Business Views appear at the top.
10. In the classic Mission compose view, **Approve & Run** stays available after the plan is generated, so you can run a preview and then publish.
### **Kaiya Architect**
Kaiya Architect works inside the Kaiya conversation. Connect a source from the chat, choose live or not live, and Architect classifies facts and dimensions, checks join coverage, and flags joins that would inflate totals, such as fan traps and chasm traps, before it builds the Business View.
1. Multiple CSV upload. Several CSV files can be uploaded at once, creating a separate dataset for each.
2. Incremental thinking. Reasoning streams in progressively, the accordion expands automatically, and tool status is highlighted as work proceeds.
3. Join validation for datasets that are not live. Datasets that are not live are loaded temporarily while the conversation is in progress, so join coverage is checked against the data for live and non-live datasets alike before you reach the join decision.
4. Suggested questions as links. Suggested questions render as clickable links.
5. Ask analytical questions before the Business View is published. Once Architect has produced a Business View plan, you can query it directly without publishing first. Questions run against the drafted model and return charts and tables in the analytical panel alongside the conversation.
Architect also generates a set of business metrics from the plan automatically, running a few analytical queries over the model so you can see representative KPIs as soon as the plan exists.
Because results are visible at plan time, a modelling problem surfaces before publication. Asking why a figure looks wrong prompts Architect to investigate the joins behind it, explain what it finds, and set out the options for fixing it, including restructuring the plan into separate Business Views with a soft link where a single joined view would inflate the totals. The revised plan replaces the original, and the same question can be re-run against it.
A final health check runs at publication, executing the metric queries once more against the published Business View.
6. A structured view of the model. A structured explore view now lists the Business Views with whether each is live and how many columns it holds. Opening one shows its columns with display name, data type, role, and source dataset, alongside calculated columns, date mappings, and dynamic parameters. A Datasets tab lists the connected datasets with their source, whether each is a fact table, and whether it is live. Where Architect has not yet produced a Business View, the connected datasets are shown on their own. The existing raw view is unchanged.
7. Several sources in one Business View. A Business View that joins several sources, including two fact tables with their dimensions, can be built and validated in the conversation. When the sources are not all joined, Kaiya Architect names the datasets in each separate group so the joins can be fixed, and a direct join between fact tables on a unique key is allowed with a note.
8. When two or more fact tables share a dimension in one model, Kaiya Architect stops before publishing, shows the shared dimensions and asks how you want to resolve it.
9. Kaiya Architect has a refreshed design across its chat, sidebars, tabs and Browse Data dialog.
### **Kaiya browser extension**
1. Automatic routing to Deep Insights. Questions needing multi-step analysis route to Deep Insights automatically. A running status appears in the current view, with a link to open the full report when ready.
2. Charts in the panel. Charts render and persist in the extension. Where a chart is too large for the embedded view, a link opens it in a full tab.
3. Query context. The selected Business View domain is passed by default for both automatic Business View selection and user query context. Uploaded images are treated as additional context for follow-up questions about a specific chart.
### **Kaiya Apps**
1. A single planner. Live and non-live planning run through one planner, so a non-live App refresh executes its existing plan rather than regenerating the SQL step.
2. Stable table names across refresh. Business Views carry a stable view name that is repointed at the latest underlying table on each save. Plans, queries, and Apps referencing a Business View continue to resolve after a refresh.
3. Where a remixed App came from. A remixed App links back to the App and version it was remixed from. If that source or its version has since been removed, the link explains why it is unavailable.
4. Narrative length is configurable. The maximum length of a generated narrative in an App is set through a runtime configuration option with its own interface control. This option was retired in 6.4.
5. Clearer App creation flow. Once an App is generated, the interface says so and offers a link to open it alongside an option to create another, rather than repeating the original call to action.
6. The App page has a refreshed design, including the Plan and Analysis tabs, the Publish dialog, and the version and refresh schedule menus.
7. Long App titles stay on one line with an ellipsis, so the creator and last-updated time stay visible on App cards.
### **Data and administration**
1. Business View CSV download. A download that ends early now explains why. The notification offers a Download action in place of View, and states that the exported file is retained for five hours, after which it is cleared.
2. Snowflake key pair guidance. The Public Key field carries a tooltip and supporting documentation explaining how to register the key in a Snowflake account, following Snowflake's policy of collecting public keys rather than private ones.
3. Usage alerts. In 6.4, Kaiya Credits are replaced by Kaiya Cost, and administrators set spending limits on its **Limits** tab. An amber banner appears when usage reaches 80% of a limit. A red banner appears at 100%, and sending to Kaiya and exporting are blocked. Alerts appear as banners in Tellius rather than as emails.
4. Multiple help videos and articles. Settings supports configuring several help videos and articles rather than one of each, with independent titles and URLs, surfaced on the Help & Support page.
5. Left navigation. The left navigation shows up to six primary modules, which an administrator chooses under Enabled Modules. Everything else is under More. "Kaiya Architecture" is corrected to "Kaiya Architect", and labels are aligned with the pages they open.
6. Your own model gateway. Kaiya uses the model provider and settings configured for your workspace. A custom endpoint or custom headers set on an Anthropic, OpenAI, OpenRouter or Gemini model connection are applied to Kaiya, so requests go through your own gateway. Add the endpoint in the **Model Parameter** JSON and any headers in **Default Header**.
7. Kaiya Semantic Views can be managed from the interface.
8. When you import content and map a dataset or Business View to an existing one, the mapping is accepted when the target holds every source column by name, even if it has extra columns. Column order doesn't matter, and Tellius names any source columns the target lacks.
### **Performance**
1. Long-running summary steps stay contained. A Summary step working over a very large result set is bounded, so it completes without holding up other work running at the same time.
2. Streaming Excel export for pivot tables. Pivot table Excel export writes rows to the workbook as it goes rather than holding the whole set in memory first, so large pivot exports complete without straining the server.
3. Dataframe loading. Dataframe loading performance is improved, including in the App archive and unarchive flow.
4. Live Narratives are cached and cleared. Narratives generated live in an App are stored, then removed automatically once they age out, so repeat views render without regenerating while storage stays bounded.
5. Faster tab and view switching in Vizpads. Switching tabs, changing view, entering edit mode, and saving are quicker, with background work removed from those interactions.
### **A redesigned home page**
The home page now opens onto the work already in progress rather than an empty starting point.
1. The standalone Kaiya page has been retired.
2. Your Missions and Your Apps. **Your Missions** shows today's completed Mission briefings first, then Missions that have not run today. **Your Apps** shows your most recently updated published Apps, including Apps shared with you, with a count of any more.
3. Suggestions that follow your work. **Suggested for you** is built from what is already there: a scheduled briefing due, then existing Missions, then the option to create a new one, with existing Apps listed the same way. The most recently updated items come first, and each card opens the matching Mission or App. The notification count shows on the icon when the sidebar is collapsed.
4. A new page lets you browse your saved queries, filter them by category or Business View, and load one into the composer.
5. One box for Kaiya and search. The Home page ask box handles both questions to Kaiya and search. Press **⌘+K** or select the search icon to switch it to search across Tellius. Results are grouped into tabs for Kaiya conversations, Vizpads, charts, Insights, Business Views and settings, and show only content you can access. Selecting a setting opens it and scrolls to it.
6. A start for new users. When you have no Missions or Apps yet, Home shows a card for each module that explains what Kaiya can do there, with a sample prompt. Selecting a card picks the matching mode and fills in the prompt.
### **Platform**
1. Expanded automated testing for Kaiya, Deep Insights and Kaiya Architect, added load testing for many concurrent conversations, and refined the infrastructure and usage tracking behind them.
2. Report a problem. **Report a problem** replaces **Provide feedback**. Select your name at the bottom of the sidebar, then **Report a problem**. Under **How bad is it?** you can choose **Blocking**, **Annoying**, **Cosmetic** or **Idea**, then describe the issue in **What happened**. **Screenshot of this page** attaches a screenshot for you, and you can remove it before you send. Your name, the web version, your browser and operating system, and the page you were on are sent with the report. After you select **Send**, a confirmation shows the report's ID.
## 🔻Deprecations
### **The Classic Editor for Missions is retired**
A Kaiya Mission is an analytical workflow that runs on a schedule or on request. That has not changed, and Missions are not going away. What changes in 6.4 is how you build one.
In 6.3 a Mission could be authored two ways. You could describe it to Kaiya in conversation, or use the form-based Classic Editor, filling in each part on its own. An administrator setting decided which of the two a user saw. 6.4 settles on the conversational path. Every new Mission is now built by describing what it should do, and the Classic Editor is retired.
Missions built in the Classic Editor remain in place and keep running on their schedule. Opening one shows its name, description, Business View, triggers, and steps as read-only reference, so the record of its configuration is preserved. These Missions are filtered out of the Mission Library by default, and a Type filter brings them back into view when you need them. That list is scoped to the people who own those Missions rather than shown to everyone. The "Classic Editor" and "Mission Editor" labels are removed throughout, and trigger phrases in Kaiya now resolve against Missions built in the current editor.
### **Learnings is being retired in favour of Memory**
Learnings and Memory both hold context Kaiya carries between sessions. Learnings has been read-only for some time and is now being withdrawn. The integration is removed from the modules that used it, and a note in the interface marks it as deprecated ahead of removal in the following release. Phrase Learnings move into Memory automatically, so nothing is lost. Query Learnings stay read-only.
### **Vizpad Classic mode is retired**
The Switch to Classic Mode option is removed from both the Vizpad interface and the Vizpad Configuration page. Vizpads always open in Modern mode, and Modern mode downloads are enabled by default.
### **Agentic Workflows is retired**
"Agentic Workflows" is no longer used. Where the term appeared alongside Missions, Missions now covers it. There is no separate Agentic Workflow concept to configure or migrate.
## **🛠️ Minor fixes**
### **Kaiya**
1. Improved the Plan view so it uses the full available width.
2. Improved PowerPoint export from a Kaiya conversation in three ways. Text now renders inside the slide boundaries. The number of slides is matched to the length of the conversation rather than to the template, so the deck length reflects the conversation. Where the available data does not support what was asked for, that is reported before the deck is generated.
3. Improved the Jump to Latest control so it navigates to the most recent query. Scroll behaviour during Deep Insight and Mission runs was also refined, so that newly generated steps do not pull the view away while earlier content is being read.
4. Improved conversation sharing so conversations shared before the current sharing model continue to appear for their recipients.
5. Improved how Kaiya joins tables on keys that span more than one column, so counts stay accurate.
6. Ensured that Custom Skills are applied when Kaiya runs a query.
7. Resolved an issue where an earlier message could keep showing **Working…** after you sent a new one, so each message shows its own state.
8. Ensured that each result shows the name of the Business View its query used, including after you switch Business Views and after a new Business View is created and published in the conversation. The name appears in the "Based on…" label right away, without a page refresh.
9. Ensured that results show the Business View you selected, rather than "Auto-detect".
10. Resolved an issue where Kaiya could report a Business View you can access as unavailable, for example when it has no description. Kaiya now finds Business Views based on your access, so it uses them without asking you to enable access.
11. Refined the review status so it shows the final result once an automatic retry succeeds.
12. Improved the elapsed-time counter so it keeps counting from the start of the response when you collapse and expand progress.
13. Improved SQL output so it displays in a readable code format, and **Copy Code** confirms when the copy succeeds.
14. Refined chart rendering so charts display without flickering and resize with the layout.
15. Refined chart sorting so values follow the measure on the Y-axis rather than the alphabetical order of the labels.
16. Removed the **Open reviewer session** option from reviewer output.
17. Refined how Excel and CSV exports display in the side panel, and ensured that image previews display. Follow-up questions and actions match the standard Kaiya layout, and HTML files can be previewed.
18. Ensured that the export preview card appears in the conversation once a Deep Insight or export completes.
19. Refined the export view so its title sits clear of the action icons and the view can be closed.
20. Improved scrolling when loading older messages, so the view stays at the message you were reading instead of jumping to the top or bottom.
21. Ensured that conversations opened from the **Recent** list in the sidebar open in the current Kaiya chat.
22. Opening a conversation that has not been shared with you shows a message naming its owner, so you can ask them to share it.
23. Refined the alignment of the arrow in the plan approval input, and the spacing between the Assistant and Send controls.
24. Improved how formatted text in Kaiya responses and citation details is handled, so the page keeps showing your content.
25. Resolved an issue where images attached in the conversation or in Apps were ignored by models from providers other than Anthropic and Amazon Bedrock, such as OpenAI, Azure, Gemini and OpenRouter models.
26. Improved how the Mission page handles an interrupted query, so the page stays available. When Kaiya is briefly unavailable as you start a new conversation, your message goes back into the chat input with your selected Business Views, and a notice asks you to try again in a moment.
### **Deep Insights**
30. Ensured that a long Deep Insight shows its progress without a generic "contact support" message while the analysis is still running, including when it pauses to ask a clarifying question.
31. Improved how Rich Summary pages load, so a Deep Insight's Rich Summary displays when it is ready and shows the selected Business View name.
32. Refined Rich Summary layout. Every page has the same even padding in the side panel, and new Rich Summaries use the same text sizes as the conversation.
33. Ensured that the "Based on" label uses the same capitalization throughout the Thought Process.
### **Missions**
34. Improved the layout of the Configuration surface. The header, banner, tab strip, titles, and step list now share the same left alignment. Each tab's view switch sits on the same row as its title rather than dropping to its own line. The artifact panel is labelled for the artifact it contains rather than repeating the tab name above it.
35. Renamed the Visuals tab to Artifact, and the tab is presented only when an artifact exists.
36. Improved the Mission listing and detail actions, with Create Mission as the primary action and View Schedules alongside it. On a Mission's page, Run and Configure are buttons in the header, and Create Schedule is in the three-dot menu with the other Mission actions.
37. Applied the current design system to Mission briefings and conversation components.
38. Improved the Started At time on the run page so that it shows the moment the run was triggered. Query metadata was also refined, so that a query's updated timestamp changes only when the query itself is edited.
39. Ensured that creating a Mission from a Deep Insight conversation creates a single draft Mission.
40. Refined Business View selection when creating a Mission, so the prompt clears once you choose and the processing steps appear.
41. Improved the Mission conversation so you can add more context as soon as a response finishes or is stopped, without refreshing the page.
42. Improved access handling for Mission queries that join two Business Views you can access.
43. Ensured that each chart in a Mission run appears once.
44. Ensured that a Mission published from the conversation shows as Published in the conversation, in its configuration and in the Mission Library, where its version is marked Active.
45. Improved publishing a Mission from the conversation. Kaiya asks you to confirm with **Publish** or **Not now**, and Python steps run on scheduled runs too, so they don't hold up publishing. After a successful preview, an unchanged plan is not run again, and Kaiya adds or changes a schedule only when you ask.
46. Refined spacing between SQL steps in the Mission view.
47. Ensured that a Mission built in Kaiya chat opens from **Edit Steps**.
48. Improved consistency between a published Mission's report and its step outputs when the Mission runs in a new conversation. Saved Python steps always run their saved code.
49. Refined the message shown when publishing after editing steps that have already run.
50. Refined padding on Mission Briefing reports, so content stays clear of the card edges.
51. Ensured that deleting a Mission built in a conversation keeps the conversation.
### **Apps**
49. Improved version handling so removing a KPI from a tab does not remove the tab, and analysis steps remain visible and current across versions.
50. Improved the App tab so it reloads after an App is published or its active version is switched.
51. Resolved an issue where an App could render incorrectly immediately after a code update and correctly again once the page was refreshed. The generated code and the view definitions it renders against are now confirmed to be in step before the App is displayed.
52. Resolved an issue where the plan review card could fail to display when an App contained an unnamed section.
53. Improved the App conversation so it stays active while the App is being built.
54. Ensured that an App Kaiya reports as built and saved displays in the interface. When Kaiya retries the start of an App build, it continues on the same App, so the App you open holds the finished build.
55. Kaiya confirms an App plan update once the saved plan has changed, and tells you clearly when an update is not supported.
56. The Plan tab on older Insight Apps shows the plan as formatted text, with headings and bullet lists.
### **Vizpads**
57. Extended number formatting in Excel export to calculated columns, so decimals, currency, percentage, separators, prefixes, suffixes, and trailing zeros are carried over as they are for regular columns.
58. Improved table export so the compare column is filled in for a compare-filter measure, including in PowerPoint and PDF exports.
59. Improved Vizpad stability so chart positions are retained and filters remain in place.
60. Improved column header behaviour so a tooltip shows the correct column name where a display name matches another column's original name.
61. Improved the quarter-to-date time filter so that it selects the expected number of days.
62. Improved column renaming so a new name is reflected when switching to table view and in the chart title, without a page reload.
63. Improved null-value filtering so it behaves consistently across the Equal and IN operators.
64. When a Vizpad export to Excel or CSV is interrupted by a temporary issue, a dialog offers to retry it once.
### **Data**
65. Improved connector search when connecting a Databricks Delta Lake datasource.
66. Improved dataset import so a Snowflake source using OAuth or key pair authentication can be created as a new source on import.
67. Improved the datasource delete dependency view so it lists Business Views, Vizpads, and insights from consolidated datasources alongside their datasets.
68. Improved notification accuracy so a completed data load is not reported as still running.
69. Improved decimal handling so high-precision numeric values, including values with more than 15 digits, are loaded and stored at their exact value rather than being narrowed to a floating-point approximation.
70. Improved CSV upload so a single upload reports a single completion notification.
71. Improved folder deletion on the Prepare page so the confirmation dialog can be dismissed, and ensured the confirmation appears on the Business View page.
72. Resolved an issue with creating a Snowflake data source connection that uses OAuth.
73. Ensured that filtering sync jobs by **Queued** in the Notifications list returns queued jobs.
74. Ensured that the **User** filter on the unstructured connector list shows the selected user's connectors. Super Users can view every user's unstructured connectors.
### **Administration and access**
75. Improved SSO logout so a user returning to the login page is redirected to the configured SSO provider where automatic SSO login is enabled.
76. Improved routing so an unrecognised login URL resolves to the login page, and an embedded link that carries no campaign source parameter resolves to its intended page.
77. Improved project sharing so the project owner retains access to a Vizpad moved into the project by a shared user.
78. Improved the conversation window so it can be reopened after it is closed.
79. Improved custom view sharing so a shared view carries edit access by default, simplifying the share and revoke flow.
80. Improved Vizpad and Business View access so a Vizpad shared in edit mode stays reachable when the original recipient is removed and access is reassigned.
81. Improved session handling so signing out and back in from one browser tab keeps the other open tabs working.
82. Improved the Kaiya Apps chat panel so the bottom-right controls collapse into an overflow menu when the panel is narrow.
83. Improved App chat so history is held for the session and earlier messages load immediately.
84. Improved the archive flow so a remixed App can be unarchived in the same way as any other App.
85. Improved conditional and gradient formatting so both are carried into PDF, PowerPoint, and Excel export, including individual chart, tab-level, and Custom View exports.
86. Improved pivot table export so a requested row range beyond the available rows returns a clear message rather than an empty file.
87. Improved pivot table grand totals so null values are handled according to the include-nulls setting.
88. Improved the Share control so it is presented to users with the Discover role.
89. Improved job handling so a job that stops part-way through is reported as such rather than leaving a progress indicator running.
90. Improved Python step repair so a step that can't be repaired automatically is reported clearly in the conversation.
91. Improved Redshift write-back for columns holding longer values.
92. Improved usage records so sign-in activity is attributed to the correct user.
### **Interface**
93. Refined listing pages so they display steadily when you scroll to the end of the list.
94. Card titles for Vizpads, Insights, Projects, Predict and Kaiya Apps, along with listing page headers, use a heavier font weight and the primary text color.
95. Removed the repeated description on the My Kaiya Credits page.
96. In the Home page composer's Business View dropdown, long names are shortened so the checkmark for the selected Business View stays visible.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.
Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:
```
GET https://help.tellius.com/tellius-6.4/whats-new/release-6.4.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
