Wobby

Wobby is an agentic data analyst built for the modern data stack, created by an early stage startup in Antwerp, Belgium. It lets a company deploy configurable AI analyst agents on top of the data warehouse and dbt models it already trusts, so business users can ask questions in plain language and get charts, summaries, and structured reports without writing SQL. The whole product is built around governance rather than open ended experimentation, using only structured, governed data and preserving context across a conversation. In December 2025, HCLSoftware announced it would acquire Wobby into Actian, its data and artificial intelligence division, and the product is being rebranded as Actian AI Analyst.

The heart of the platform is its governed semantic layer and a companion called the Steward Agent. Before any questions are asked, the Steward Agent structures models, metrics, glossary terms, and relationships from a company's existing data assets, so business logic is defined and agreed up front. It then keeps watching definitions and data connections, surfacing reviewable action plans when something needs attention. Because every question is answered against this layer, a metric means the same thing every time, and the analyst agents produce results that are consistent across conversational analysis, recurring monitoring, and executive reporting rather than inventing a new interpretation each time.

Trust comes from visibility and control. Every answer can be opened up to show exactly how it was built, including the joins, filters, and metric calculations behind it, and each query is logged and auditable down to who asked what and when. Data teams set the rules, sign off on query templates, and scope an agent's access to specific models, dimensions, and measures to protect sensitive data. Queries run in place against the warehouse with no data copied out. Business users reach the analyst through a web application and directly inside Slack and Microsoft Teams, where context is preserved within a thread so a team can explore together.

For everyday questions the agent returns fast, direct answers, while a deeper mode engages more advanced reasoning that looks across many tables and runs multiple steps to investigate why a number moved. Beyond answering on demand, Actian AI Analyst continuously monitors key metrics, trends, and anomalies and proactively surfaces important changes on its own, then packages findings into structured reports that can be refined in conversation and shared. Pricing is self serve and metered by message volume, with a fourteen day free trial, monthly limits that reset without overage charges, and higher enterprise tiers, so it fits data teams that want governed, self service analytics on their existing warehouse.