Slack-native business intelligence agent

Your company data, answered like a conversation.

Operator connects to the systems your team already uses, understands how the business works, and answers operational questions in Slack with sources, context and guardrails.

10 min

first setup

Read-only

by default

Sources

on every answer

Company Agent

Connected to 6 business systems

Live

Question

Why did qualified pipeline drop this week?

Answer

Pipeline fell 18% mainly from fewer demo-booked leads in Colombia. HubSpot shows a 31% decrease in SQL conversion, while Mixpanel activation stayed flat.

HubSpot
Snowflake
Mixpanel
Customer.io
Connect the operating stack

Plug in the tools. Keep the context.

The agent maps schemas, events, objects and campaigns into a business-aware catalog, then asks for approval before using discovered metrics.

Slack

HubSpot

MySQL

Snowflake

Mixpanel

Customer.io

Postgres

Stripe

Slack

HubSpot

MySQL

Snowflake

Mixpanel

Customer.io

Postgres

Stripe

Ask from Slack

No new dashboard. No training session. Anyone can ask a business question where work already happens.

Queries real systems

Connect warehouses, CRMs, product analytics and lifecycle tools through read-only connectors.

Explains the why

The agent breaks questions into steps, checks multiple sources, and returns the evidence behind the answer.

Permission aware

Access is scoped by role, channel and source. Sensitive fields can be masked before the model sees them.

Built for trust

Not another reckless chatbot.

Operator is designed around narrow scopes, auditability and reviewable work. It can answer fast without pretending the model should have unlimited access.

Read-only connectors by default
Role and channel-based access
SQL validation before execution
PII masking and query logs

// query policy

allow: SELECT

block: INSERT, UPDATE, DELETE, DROP

limit: 5,000 rows

timeout: 30 seconds

mask: email, phone, address

// every answer returns

answer + method + sources + confidence

Setup that feels obvious

From disconnected tools to answerable business questions.

1

Install the Slack app

Add the agent to selected channels and choose who can ask company-wide questions.

2

Connect data sources

Use OAuth, service accounts or read-only database users to grant narrow access.

3

Approve discovered metrics

Review the tables, events and definitions the agent discovered before launch.

4

Ask your first question

Get an answer with sources, methodology and the next useful follow-up.

Give every team a business analyst that already knows where to look.

Start with Slack, HubSpot and your warehouse. Add more systems as the agent earns trust.

Request access