Introducing Kimo v2: ask anything, trust the answer

Kimo v1 was a good dashboard tool. Teams connected Postgres or Stripe, dragged a few fields onto a canvas and shared a link. But in hundreds of customer calls we kept hearing the same thing: the chart was never the hard part. The hard part was knowing which table to trust, what “active customer” meant this quarter, and whether the number in the board deck matched the number in the CRM.

Kimo v2 is our answer. It is a rebuild of the engine underneath every product, and it changes how you get from a question to an answer you can defend in a meeting.

Ask Kimo: answers you can audit

Ask Kimo lets you type a question the way you would ask a colleague: “What was net revenue retention for annual plans last quarter, split by region?” Kimo resolves every word in that sentence against your semantic layer — the shared definitions your team has agreed on — and only then writes the query.

That ordering matters. Most AI analytics tools generate SQL directly from your schema and hope for the best. Kimo refuses to answer when a term is ambiguous, and tells you why. Every answer ships with a trace: the measures it used, the filters it applied, the generated SQL, and the freshness of each source.

A semantic layer you will actually maintain

Semantic layers have a reputation for being a six-month project owned by one heroic analytics engineer. We designed ours to be maintained by the people who own the definitions. Measures and dimensions live next to the data model, are versioned, and can be edited from the UI or from YAML in your repository.

12345678910111213model: subscriptions
source: stripe.subscriptions
measures:
  - name: mrr
    label: Monthly recurring revenue
    sql: sum(case when status in ('active', 'past_due') then amount_monthly end)
    format: currency
    owner: finance@yourco.com
dimensions:
  - name: plan_interval
    sql: plan_interval
  - name: region
    sql: customer.billing_region

Once mrr exists, every dashboard, every alert and every Ask Kimo answer uses the same definition. Change it once and the change propagates — with a diff in the activity log so nobody is surprised on Monday.

Live dashboards, not snapshots

Dashboards in v2 are wired to the incremental sync engine, so they refresh as soon as new rows land rather than on a nightly cron. Each tile shows its own freshness badge. If a source is late, the tile says so instead of quietly showing yesterday’s number.

• Freshness on every tile — green under the SLA you set, amber when late, with the exact last-sync time on hover.
• Drill to rows — click any bar to see the underlying records, filtered exactly as the chart was.
• Alerts from any tile — “tell me when weekly signups drop more than 15% week over week” in two clicks.
• Share with context — links carry filters, comparison period and a frozen definition version.

Three products, one engine

The same engine now powers three products, each tuned for a different kind of team. Kimo Marketing ties SEO, ads and product data into one funnel. Kimo Business Intelligence gives finance and RevOps governed metrics and board-ready reporting. Kimo Defense Intelligence runs fully on-premise — even air-gapped — for teams fusing sensor, OSINT and cyber feeds.

What’s next

Over the next quarter we are focused on three things: write-back to your warehouse so models can be materialised where your other tools can read them, a public API for embedding Ask Kimo in your own product, and scheduled narrative reports that explain why a metric moved, not only that it did.

Kimo v2 is available today for every workspace. Existing dashboards migrate automatically; your definitions are imported into the semantic layer as drafts for you to review. If you want a guided tour, open the live demo — it runs on simulated data, so click everything.