Start free

Analytics AI — part of the Brainis AI Suite

Reports on your live company: questions across teams, alerts, honest numbers, and analyses that show their method.

Dashboards · reports · BI · Analysis Runtime. I do the work, you have the final say, and the work gets checked.

The real product

The product, screen by screen.

The screens Analytics AI really runs, the same ones my agents work in.

Analytics AI · Dashboard

sketch

A sketch of the screen, not a screenshot. Real captures come with the recorded demo.

What it runs

Real plays, not a mock-up.

This part of the suite is newer. Its plays aren’t in the catalog yet, so I won’t list ones I can’t back up. The same missions and checks below still apply.

Step by step

Answer the churn question with method disclosure

Not a chat guess: the cohort analysis runs in a checked environment and returns findings that name their method, sample and limits.

  1. you

    Ask the question — natural language or the query builder, joined across domains

  2. agent

    Run the ready-made cohort analysis in a checked environment

  3. Brainis

    Findings arrive with their method, and a small-sample warning where the data is thin

  4. check

    Numbers checked to trace back to live data

  5. you

    Set the threshold alert — the metric watches itself from now on

Each step shows who does it. The 2 steps marked "you" are your calls. Analytics AI handles the rest.

What’s in it

The full feature list.

The boring list buyers check. No label means it's live today. Anything else says so on its own line.

21 features listed · 19 live · 1 beta · 1 coming

Dashboards

5
  • Custom dashboardsDrag-to-build with grid snapping and team-level sharing.
  • Dashboard templates
  • Widget libraryKPI cards, sparklines, funnels, cohort charts — drop them anywhere.
  • Annotations & favorites
  • Executive briefing view

Query & modeling

7
  • Visual query builderNo code required; full SQL underneath when you want it.
  • Natural-language queries
  • Cross-domain joinsPeople + Revenue + Work in one chart — one picture underneath, so the numbers really line up.
  • Semantic layer & glossary
  • Data catalog & lineage
  • Segment builder
  • Data transforms & quality checks

Reports & alerts

5
  • Report designer
  • Scheduled reportsDigests delivered on schedule — PDF, CSV, or live links.
  • Threshold alertsPinged when a metric crosses your line, not when a vendor feels like it.
  • Export controls & data masking
  • Role-scoped access

Analysis Runtime

4
  • Curated analyses with method disclosureRegressions, forecasts, and cohort analyses that name their method and sample — never a chat guess.
  • Small-n warningsThin data says so, instead of performing significance.
  • Custom analysis authoringComing
  • Calibration feed to VerifyBetaThe views that check how sure predictions were use this same data.

How it connects

How Analytics AI fits in.

Analytics AI uses the same picture of your company as the rest of the suite (see Company State). Its work turns into missions, gets checked by Verify, and only goes as far as you let it.

In detail

What it reads
the live company, not exports — so a number on a dashboard is the number in the product.
What it gets from me
regular reports become scheduled work instead of a monthly scramble.
What gets checked
each number traces back to live data; each analysis shows its method as part of the work.
Works with
every module feeds it; threshold alerts open decisions and missions in the module that owns the metric.
Honest numbers
each insight shows where its number came from, how fresh it is and how sure I am.

Agents at work

What the agent can touch, and what it can’t.

My agent works inside Analytics AI, one job at a time. Its access is like a keycard: it opens only what the job needs, and nothing marked "never", however the agent asks. How the keycard works.

Agent keycard · Analytics AI

ANALYTICS AIKEYCARDone job onlyexpiresREAD:the notes for this jobWRITE:drafts you reviewNEVER:your database

Drag sideways to see the rest

checked again on every step · nothing extra gets through

UnderstandDecideDeliverVerifyLearn

Analytics AI is one of the real products I do the work in.

FAQ

Questions people ask first.

Is this a real BI tool or AI summaries of our data?

A real one: dashboards, a visual query builder with full SQL underneath, a semantic layer, lineage, scheduled reports, alerts. I run analyses on top — and they show their method instead of guessing.

What makes the numbers honest?

They come from live data with labels: source, freshness, how sure. Analyses name their method and sample size, and thin data gets a small-sample warning instead of a confident answer.

Can I write my own analyses?

Ready-made analyses ship now; writing your own keeps its "coming" label until it lands. What exists is never blurred with what's next.

Plans

Which plans include it.

On every plan

Every plan runs it, Free included, with every source you connect.

Prices live on one page: see the plans.

See the AI suite How it all works