Operations

AI Business Analyst

Turns a table of numbers into the three things that matter

Version 1Published Oct 19, 2026
OperationsExtractSummarizeclaude-opus-5Read-only · autoResult
4output fields
A structured result, not a paragraph you have to re-read
≤7steps per run
Bounded work: at most 6 tool calls, enforced by the runtime

Read-only by design. This agent has no path to sending, spending, publishing or committing anything. Where that limit is the product, removing it would remove the reason to trust the output.

Live wiring for version 1. Tool nodes are the capabilities this agent is registered to call — nothing else can run.

Overview

Give it a dataset and the question you are actually asking. It returns the findings that matter, the plausible drivers behind them, and what to do next — each with an explicit confidence level.

It is instructed to distinguish correlation from cause and to state when the data cannot answer the question, which is more often than most dashboards admit.

Benefits

3 findings

Analysis, not a dashboard

Arrives at a conclusion instead of more charts.

Testable next steps

Each recommendation includes how to validate it.

States what the data cannot answer

Every run returns its limits — the questions the dataset does not support — which is the half a dashboard never shows you.

What it does

  • Question-led analysis

    Analyses toward your question, not a generic summary.

  • Driver hypotheses

    Proposes plausible causes and how to test them.

  • Confidence labelling

    Every finding carries a confidence level.

  • Explicit limits

    States what the data cannot support.

What this changes

What this changes about your week

It takes over
Pulling the numbers together and working out what actually moved.
Done today by
The owner, in a spreadsheet, on a Sunday.
How often
Every review.
Typical manual effort
A monthly pack assembled by hand is a day.

Why Operations in particular. The analysis usually exists in the data already. What is missing is the day nobody has to go and find it.

  • Runs unattended

    Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.

  • The same standard every time

    The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.

  • It cannot act on its own

    AI Business Analyst has no path to sending, spending or committing. That limit is why its output is safe to act on.

How it runs

This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.

  1. 1InputValidated against the agent input schema
  2. 2Modelclaude-opus-5 with automatic fallback
  3. 3Toolsextract_structured_data, summarize_text
  4. 4PermissionsRuns READ actions automatically; anything higher waits for approval
  5. 5OutputStructured result matching the output schema
  6. 6LoggingTokens, duration and every step recorded

Demo

Demo dataIllustrative sample output.

Input

{
  "dataset": "week,visits,signups\\n1,4200,310\\n…",
  "question": "Why did Q3 conversion fall?"
}

Output

{
  "limits": [
    "No channel dimension in the data, so acquisition mix cannot be ruled out"
  ],
  "drivers": [
    {
      "test": "Compare page-level conversion before and after",
      "hypothesis": "Pricing page change in week 6"
    }
  ],
  "findings": [
    {
      "finding": "Conversion fell 18% from week 6",
      "evidence": "Signups fell while visits held",
      "confidence": "high"
    }
  ]
}

Requirements

Integrations

No integrations required.

Before you start

  • A dataset as CSV text or JSON
  • The business question you want answered
  • Requires the All-Access plan or higher

Use cases

Monthly business review

Turn the numbers into a narrative.

Anomaly investigation

Work out what changed and why.

Pricing

All-Access

Billed monthly through your PROMIVO subscription. Cancel at any time.

Runs consume your plan allowance for agent executions and tokens. See plan limits

FAQ

Can it connect to my warehouse?

Not in this version. Supply data as CSV or JSON, or use a workflow HTTP node to fetch it from an internal API. Warehouse connectors are a custom agent request.

How do I get my data in?

Paste it as CSV or JSON, or attach the file — a CSV export is read on upload and discarded after the run. Spreadsheets need saving as CSV first, which is one menu item in Excel.

Who pays for the AI usage?

You do, directly to your provider on the key you connect. The size of the dataset you paste is what moves the cost, so trim to the columns your question needs.

Reviews

Reviews open once customers have run this agent.

Need this agent to work differently?

Tell us what to change and our team will scope a customised version for your business.

Customize this agent