Bring the SQL, the chatbot and the dbt project you already have
Jinja report queries become governed data APIs as written; a coding agent provisions them through the API; an ask-sql chatbot moves without changing its client; suggestions come from dbt and query history.
For data platform leads planning a migration · 10 min · 14 Oct 2026 EngineResolve the name before writing the SQL
A misspelt client in a question becomes the right id before any SQL is written — and when five records share the name, the engine asks which one. The names come from an entity catalogue you curate.
For heads of data and chatbot teams · 8 min · 13 Oct 2026 ProductFind anything by name — even spelt wrong
A search API over the warehouse's entities: names fuzzily from three letters, ids exactly, active records first, related records attached, and only within the caller's scope. Defined without SQL.
For product owners and engineers · 9 min · 12 Oct 2026 ProductBuild the query, then the chart — without writing SQL
A visual query builder that works out the joins, and a chart builder with shelves, Show me and a sentence box. Both compile to your warehouse's dialect; every chart added to a dashboard is a published data API.
For analysts and analytics managers · 8 min · 11 Oct 2026 IntegrationA home page for every person, in one call
A 360 view asks for one person's home and gets every section — their book, accounts and brokers — as verified insight cards, computed overnight. Personas are configuration, not code.
For product teams building a 360 view · 8 min · 7 Oct 2026 ProductAn insight that shows its work
An agent proposes the questions your warehouse can answer. You build on one, every figure in the story is checked against the rows it came from, and what you publish is a versioned endpoint — not a screenshot.
For heads of data and analytics leads · 10 min · 3 Oct 2026 ProductDescribe a dashboard. Watch it build.
One sentence becomes a filter bar and a page of panels that fill in live as each query is written and proven. Add a panel by describing it, drop in an insight, or make the whole page take a broker group as a required parameter.
For analytics managers and product teams · 9 min · 3 Oct 2026 GovernanceA catalog with a ledger
Everything published in one list with its usage, every call recorded, an approval before a publish, grants by API key or by Okta group, column masking and daily quotas — governance as the product, not a policy document.
For heads of governance, platform and security · 10 min · 3 Oct 2026 EngineeringTeaching the engine your vocabulary
"Flagship stores" is a size code. "Bound" is three status values. How the glossaries you already wrote become rules the SQL engine enforces — and what a machine is never allowed to decide on its own.
For analytics leads and data governance · 9 min · 29 Sep 2026 GovernanceWhere did this number come from?
A KPI on a dashboard should answer for itself: the definition that computed it, its expression and traps, the glossary term it defines. Three links make that true, and the cache becomes a governance feature.
For heads of governance, BI leads and risk · 9 min · 29 Sep 2026 StrategyYour dashboards are a dead end for AI
Every BI tool locks its numbers behind its own screen. That was tolerable when the only consumer was a person reading a report. It stops being tolerable the moment the next consumer is an application, an agent or a customer portal.
For CIOs and heads of data · 8 min MigrationFrom a Power BI screenshot to a live dashboard
An agent reads a report page, fans out one sub-agent per panel to write the SQL against your warehouse, and hands back a working dashboard — with an honest report of what it could not answer, and why.
For analytics managers and tech leads · 9 min GovernanceOne number, one definition
Storing SQL an AI wrote, exposing it as an endpoint, letting applications pass filters into it — that is a governance question before it is an engineering one. Here are the specific answers.
For heads of AI, risk and engineering · 10 min EngineeringEmbedding insight in the apps people already use
Two calls: the layout once, then rows as filters change. Your application receives a render-agnostic specification and draws it with its own components. No iframe, no second design system.
For engineering leads and architects · 8 minWhere to start
If you are deciding whether this belongs on your roadmap, read the strategy piece first, then how an insight is built and verified. If you have already decided and are sizing the work, start with describing a dashboard and the migration piece. If your question is "how does it get the number right?", read the vocabulary piece and then the lineage piece. If your question is "who may call it, and who approved it?", read the catalog piece.