Proactive Intelligence and Wider Support
MAY 15, 2026 – After several weeks since general availability, MyDataWork v1.1 is now live. Analysts, BI developers, data engineers, data scientists, and the leaders who manage them benefit from a meaningfully more proactive workspace, broader file format coverage, and clearer accountability around how AI is used.
For readers new to MyDataWork: the platform is a workspace where data professionals catalog the analytical work they do — the files, scripts, dashboards, notebooks, and queries scattered across the tools they use every day — and connect that work to the use cases, stakeholders, and outcomes that give it business meaning. GA established that foundation. v1.1 makes the workspace start to do work for you.
This post covers what’s new in v1.1 and what is still ahead. The summary is that the workspace is no longer just a place that organizes your data work — it now reviews it.
What’s new in v1.1
The Workspace Agent. A new Suggestions tab in the Analytics section gives you a single button — Analyze — that sweeps your entire workspace across six checks and surfaces findings grouped into three categories: Cleanup, Activity, and Insight. The agent looks for use cases missing stakeholders, active use cases that have gone stale, recently-added assets that aren’t yet linked to anything, removed assets still depended on, high-value use cases tracking below estimate, and assets that have quietly become “hidden infrastructure” by being linked across many use cases. Each suggestion is one-click navigable, and the agent auto-resolves it once you address the underlying condition.
The Workspace Agent is the first capability in MyDataWork that operates without being asked. Every other AI surface — the Assistant, AI Recommendations, the Leverage modes — answers questions you bring to it. The agent looks at your workspace on its own and tells you what it noticed. You’re charged one credit per suggestion surfaced; a run that finds nothing costs nothing; your first analysis is complimentary.
Expanded Connector file types. Three additions to the formats the Connector recognizes: ThoughtSpot (.tml), Dataiku DSS project exports (.zip), and Looker LookML (.lkml). These join the existing coverage of Power BI, Tableau, Alteryx, SQL, Python, R, Excel, CSV, and notebook formats. Place files in any folder the Connector already scans and they appear on the next pass.
Asset metadata. The asset detail panel now shows each file’s size and creation date alongside its existing metadata. Useful when you’re sizing up an unfamiliar folder or recalling when a file first appeared.
Notes URL on use cases. Each use case now carries a Notes URL field — an HTTPS link where you keep the running notes for that work. Paste a Google Doc, Notion page, Confluence space, or any other web destination, and an “Open notes” button appears in the use case panel. Use cases stop being islands; they connect to the working notes that surround them.
AI credit transparency. Every AI-spend trigger in the app now states its cost at the moment of clicking. Generate buttons show “(1 credit)” inline. The Assistant chat panel notes “1 credit per message” under the input field. The Workspace Agent’s Analyze button discloses its variable ceiling. You can see exactly what each action will cost before you click it.
Refund on AI call failure. If an AI generation fails — a timeout, a service error, any technical issue — the credit is automatically refunded and the response message confirms the credit wasn’t charged. You never pay for a result you didn’t receive.
Tab visual grouping. The main workspace tabs now carry subtle color tinting grouping them by function: Content for assets and use cases, Analytics for insights and AI surfaces, Foundation for people and setup. The active-tab styling is unchanged — this is a quiet visual cue that makes the navigation feel less like a flat list.
What’s still ahead
Several capabilities remain on the roadmap and are worth naming so users know what to expect.
Mac and Linux Connectors. Still Windows-only. Mac remains the higher priority of the two and continues to be the next major Connector platform target.
Additional cloud source integrations. v1.1 doesn’t add to the GA set (GitHub, dbt Cloud, Databricks, Snowflake). Looker, Sigma Computing, Hex, BigQuery, Tableau Cloud, and Power BI Service remain candidates for future releases based on user demand.
Workspace Agent expansion. v1.1 ships with six checks. The architecture is designed to grow — additional rules covering more patterns of analytical work are a natural next step, and the categories (Cleanup, Activity, Insight) are deliberately broad so new rules can join an existing structure rather than introducing new surfaces.
Connector upgrade experience. The Connector still requires manual reinstall to pick up changes. A polished in-place upgrade path remains on the post-GA list.
Mobile experience. Desktop-first remains the design center. Mobile-friendly views remain a longer-term consideration based on observed customer usage.
What this means for you
GA established that your analytical work could be cataloged and connected. v1.1 establishes that someone can look at it for you.
For analysts who want their workspace to surface what’s worth attending to: the Workspace Agent makes the catalog active. You don’t have to remember to look — Analyze does the looking. The findings are framed as insight and prioritization, not alarm. Run it weekly, address what matters, and your workspace stays current without manual upkeep.
For teams that want shared visibility without surveillance: the agent operates on the workspace, not on individuals. Findings about hidden infrastructure, stale work, or unrealized value reflect the work itself, not who did or didn’t do something. Team leaders see patterns; individual contributors see suggestions on their own work.
For organizations thinking seriously about agentic AI: the Workspace Agent is a deliberate example of how an agent can be useful without being unpredictable. The decision logic is explicit and auditable — six rules with named thresholds. The language model handles only the phrasing, not the judgment. This is the shape of agentic capability that earns trust at scale.
The fastest way to see v1.1 is to use it. Existing users see the new tab on their next login. New users can sign up at mydatawork.com, import the demo workspace, and click Analyze to see the agent in action within the first session.