As summer winds down and school resumes, Domo is gearing up to head out for its annual Domo Connections AI + Data Tour.
Across stops this year, Domo is giving customers a chance to sharpen their skills and see how other teams are using AI agents and workflows. Plus, a chance to glimpse what’s coming next.
If you’ve already attended a stop, you may have heard some of the updates already. If you have an event on your calendar, you’ll get to explore them in person. And if you haven’t been able to make it out, we’ve got you covered in here.
Despite the growth in BI adoption (from about 8 percent to roughly 20–30 percent of employees over the past two decades) questions about value linger: Only 49 percent of data teams are confident their companies benefit from their work. We believe conversational AI offers more people a way to ask questions and interpret the answers. The challenge is to create that trustworthy experience where answers are grounded in governed data and business context.
Domo’s latest updates address this challenge by extending Domo’s governed data foundation into more AI tools, applications, and workflows. Here’s a look at the features the tour has highlighted, including what’s available now, what’s in beta, and what’s coming.
Extending Domo’s data foundation
1. AI Library and reusable toolkits
The AI Library gives you one place to build and manage agents and their reusable toolkits, with the same governance you already apply to your data. A toolkit brings together the content an agent can draw on and the tools it can use, along with skills—instructions that guide how it behaves.
A support team, for example, could define that context once and share the toolkit across agents with the appropriate access. You can combine Domo-provided tools with custom tools built in Code Engine, giving agents capabilities suited to your team’s work.
When configuring an agent, you can attach a toolkit, add further knowledge, and customize its chat interface. A testing panel lets you ask the questions your team would ask and refine the agent’s responses before sharing it.
Status in September 2026: Beta
2. Use Domo agents and toolkits outside Domo
A custom agent built in Domo can be embedded in a company portal or another setting where people already work. Teams can also connect a Domo toolkit to an external AI client such as Claude, Gemini, or Codex through the Model Context Protocol, or MCP. Choosing “Use MCP” from the toolkit’s menu provides the connection details.
The toolkit’s knowledge, tools, and skills remain managed and governed in Domo, so you can reuse the business context you’ve already established. Calls to Domo use the permissions of the person connecting to Domo. For example, someone with access to one region’s data remains limited to that region when asking questions through an external client.
A Domo toolkit can also work alongside other MCP connections. A support team could use its Domo toolkit together with a separate connection to Jira to investigate an issue from the same AI client.
Status in September 2026: Beta
3. More Domo work through MCP
Domo’s initial MCP beta, released in May, let agents query DataSets and search Domo content. The September release adds Domo-provided toolkits in AI Library that let external AI clients create and manage selected Domo content.
These toolkits extend MCP access to cards and Workspaces, along with user administration and Sandbox. Builders can use their preferred AI tools to help create content their teams can use in Domo, with that work governed by their existing Domo permissions.
Additional capabilities for App Studio, Magic ETL, publishing, and embedding administration are on the roadmap. Automation capabilities involving Workflows, Code Engine, Tasks, and Forms are also planned.
Further capabilities for App Studio, Magic ETL, and Domo Everywhere publishing and embedding administration are planned in the future. Anticipated automation tools will extend that work into running business processes through Workflows, Code Engine, Task Center, and Forms.
Status in September 2026: Beta
4. Use Domo data to build working apps
The expanded MCP capabilities also support a more operational use of AI. A builder can connect an external AI client to Domo data, identify what information is available, and create an app that combines existing records with information captured during the work.
A collections team, for example, could bring together customer, invoice, and payment data while adding a form for recording promises to pay. Once a representative records a customer’s commitment, the app can track whether the promise is kept and flag the account for follow-up if it is missed.
This kind of workflow gives teams a way to capture information that previously lived in email, notes, or someone’s memory and make it available for later analysis.
Status in September 2026: Beta
Building and delivering applications
5. An AI assistant for App Studio
The upcoming App Studio AI assistant lets you describe changes in a chat right inside your app. You can ask it to update the theme across the app or add a page by explaining what it’s about and who it’s for. The assistant lays out the page, selects visuals, and connects them to your data, giving you something to review and refine in the same conversation.
The assistant is designed to reduce the manual configuration involved in building an app while keeping the builder in control of the result. You’re building on the same governed data foundation already in place in Domo, so faster app development carries that foundation’s controls forward.
Status in September 2026: Upcoming
6. More flexible App Studio components
App Studio is also gaining components and sharing options for different audiences.
A new KPI widget provides a reusable way to display a metric, while conditional formatting lets components and header KPIs respond to defined conditions. An AI summarization component is planned to add a written explanation alongside the numbers.
Dynamic page permissions will let builders share individual pages with specific people or groups instead of maintaining separate copies of an app for each audience. Scheduled PDF reports will deliver app content by email to colleagues who rely on regular reports.
Status in September 2026: Upcoming
7. Ask more complex questions in AI Chat
AI Chat is being updated to handle multipart questions and draw on multiple DataSets in one conversation. Instead of breaking a complicated request into separate prompts, you can ask several related questions at once and receive a combined response.
When a question involves more than one DataSet, AI Chat queries each DataSet separately and synthesizes the results. It does not join or union the DataSets. That means you can combine findings in an answer without creating a new DataFlow for every cross-functional question.
You can also configure a custom chat experience for an App Studio app, with the knowledge and tools selected for that audience.
Status in September 2026: Beta
8. Bring documents and visuals into AI Chat
AI Chat will also be able to use information beyond structured DataSets. Support for Domo Documents will allow stored files to contribute context to a conversation. You’ll also be able to upload a file or paste an image directly into the chat.
When an answer would be clearer as a visual, AI Chat can create a card from the analysis and add it to the current page. You can direct a conversation toward a particular source by mentioning a DataSet, while Domo’s Knowledge Base can provide instructions for completing tasks without leaving the current workflow.
Status in September 2026: Beta
Preparing data for AI
9. Carry context through Magic ETL
Context added to a DataSet should remain useful as the data moves through a DataFlow. New connector settings will allow column descriptions and other metadata to be brought into Domo from a source system.
An upcoming Magic ETL beta will allow AI-ready metadata to travel through transformations. At the output, builders can review the inherited context, add more information, or resolve conflicts when multiple sources contribute definitions.
The result is more information about what fields mean, even after data has been combined or transformed. That context can then support both human analysis and AI-assisted work.
Status in September 2026: Upcoming
10. Use new AI tiles and Document Collections
Magic ETL’s existing AI Text Generation tile will support both batched and non-batched processing. New Sentiment Analysis and Classification tiles are designed for more specific tasks, such as identifying the tone of support tickets or sorting documents and images into categories defined by the user.
Document Collections support, coming to beta, will extend Magic ETL to unstructured content. Information from PDFs, images, spreadsheets, and Office files can be converted into structured DataSets for use in a DataFlow and further transformation.
Status in September 2026: Upcoming
11. Get guidance while building DataFlows
A planned Magic ETL AI assistant will help users understand unfamiliar data and build a DataFlow. You could ask what kinds of analysis might be useful for a DataSet and then ask how two DataSets should be joined. The assistant can analyze the available information and provide step-by-step guidance for configuring the join.
The initial version focuses on advice. Later enhancements are expected to configure actions on people’s behalf and create documentation for the DataFlow. The release is expected to go to beta later this fall.
Status in September 2026: Upcoming
12. Troubleshoot DataFlows more easily
Two DataFlow improvements are moving into general availability with the September customer release.
First, execution heatmaps show what happened at each step of a run, including changes in row counts and resource use. Builders can use different views to identify where rows were dropped or multiplied and investigate unexpected results.
Second, you can now inspect earlier versions of a DataFlow without making one of those versions current. After reviewing an earlier version, you can save it as the current version or as a separate DataFlow.
Status in September 2026: Live
Connecting business logic and controlling change
13. Bring Snowflake Semantic Views into Domo
Now in beta, Domo’s integration with Snowflake Semantic Views lets you reuse the business logic and table relationships you’ve already defined in Snowflake. You can register a Semantic View through Cloud Integrations much as you would connect a Snowflake table.
Domo reads the table and column metadata, including comments and synonyms, and makes it available in the DataSet schema and AI Dictionary. The relationships between tables also carry across into the Data Model editor.
The result works like any other Domo Data Model, ready to use in cards, filters, dashboards, and apps. You can build on the definitions you’ve established in Snowflake without recreating those relationships in Domo.
Status in September 2026: Beta
14. Track AI spending and write data back to external systems
New administration connectors for Anthropic, Cursor, and OpenAI bring billing and usage information into Domo. Teams can use that information to monitor AI spending, build dashboards, or create alerts as usage changes during the month.
The JSON writeback connector also adds OAuth and API key authentication. That expands the types of external APIs to which Domo can write data.
Both updates are available now, closing gaps we heard about from you, the community.
Status in September 2026: Live
15. Reuse connector configurations and track changes
Upcoming connector improvements will make it easier to reuse configurations and understand changes over time.
Export and import support for JSON connector configurations will allow teams to save a working setup to a file, share it with a colleague, and reuse it with the appropriate credentials. Planned schema-change alerts will notify teams when a DataSet’s structure changes in a way that could affect downstream assets.
Connector notes and configuration history will provide additional context about how a connection was set up and what has changed.
Status in September 2026: Live
Organizing everyday work
16. Organize everyday work with navigation and Workspaces
The new left navigation and Workspaces are already generally available in Domo. The latest updates build on what they can do.
And coming to beta in October, enhanced pinning will let people create pins, folders, and subfolders more directly within the navigation. Smart folders will gather content according to criteria such as content type, ownership, or recency, while respecting permissions.
Also, permissions-aware navigation configurations will let administrators create a navigation structure that adapts to what each person can access. A new Workspace tray will let users browse folders and open content without losing their place in the navigation.
Status in September 2026: Upcoming
Explore the updates
If one of these updates caught your attention, the Domo Drop playlist on YouTube is a good place to take a closer look. Its feature walkthroughs explain how the latest additions work and where they could fit into your day.
For features available in the September and upcoming betas, joining Domo’s beta program gives you a chance to try them with your own work and share feedback with the product team. Your experience can help shape how these capabilities develop, just as customer feedback has helped shape the updates covered here.
.jpeg)