Agentic BI in Slack: Run Analytics with /knowi Command | Knowi

Agentic BI in Slack lets you run your entire analytics workflow from any Slack channel using the /knowi slash command. Type a request in plain English and Knowi’s AI orchestrator handles the rest: finding the right dataset, running the query, building dashboards, setting up alerts, and scheduling reports – all without leaving the conversation.

Quick Summary (TL;DR)

• The /knowi slash command connects directly to Knowi’s full AI agent layer from any Slack channel or DM where the bot is installed.
• You can query data, create dashboards and widgets, set threshold alerts, schedule recurring reports, search assets, and get trend analysis – all through natural language.
• When a request is ambiguous, the bot surfaces clarification options as buttons rather than guessing, maintaining conversation context across turns.
• Results come back in the right format automatically: data tables with chart images, dashboard links with action buttons, plain text for insights, and confirmations for alerts and schedules.
• Follow-up questions work within a channel – “break that down by region” re-queries using the context of the previous result without re-stating the original request.

Table of Contents

What Is Agentic BI in Slack?

Agentic BI in Slack is a native integration that connects Knowi’s AI agent layer to Slack through a slash command interface. Unlike standard Slack bots that surface static reports or pre-built notifications, the /knowi command routes each request through an AI orchestrator that interprets intent, selects the correct dataset, runs a live query, and takes action – creating dashboards, setting alerts, or building schedules as needed.

The integration works in any channel or DM where the Knowi bot is installed. There is no separate interface to learn and no need to open Knowi in a browser to get a data answer.

How to Use the /knowi Command

Type /knowi followed by your request in plain English. The bot processes the request, runs the appropriate agents, and posts the result back in the channel. Examples of supported requests:

/knowi what is total revenue this month?
/knowi create a dashboard with revenue by region
/knowi alert me when sales drop below 50k
/knowi send this to #finance every Monday
/knowi what trends do you see in the data?
/knowi find dashboards about revenue

/knowi what is total revenue this month?

/knowi create a dashboard with revenue by region

/knowi alert me when sales drop below 50k

/knowi send this to #finance every Monday

/knowi what trends do you see in the data?

/knowi find dashboards about revenue

Requests do not need to follow a specific syntax. The AI interprets the intent and routes to the correct agent. “Show me last month’s churn” and “what was churn in March?” are treated the same way.

What Can You Do from Slack?

The /knowi command supports seven categories of actions, each handled by a different agent in the Knowi orchestrator:

How Do Responses Work?

The bot returns different formats depending on what was requested. Each format is designed to match what you actually need from that type of response rather than returning everything as raw text.

Data Results

Query responses return a formatted data table of up to 10 rows with a chart image when available. Action buttons below the result let you download the full dataset as a CSV or open it in Knowi to explore further. For larger datasets, the download button retrieves all rows, not just the 10 shown in Slack.

Dashboard or Widget Created

When you ask Knowi to create a dashboard, the bot returns a chart image of the newly created visualization along with four action buttons: View, Schedule, Share, and Create Alert. You can take any of these next steps without leaving Slack.

Insights and Confirmations

Trend analysis, alert confirmations, and schedule confirmations return as plain text. These responses do not include a chart image – the information is the result, not a dataset.

Clarification

When a request matches multiple datasets or is not specific enough to act on, the bot asks a follow-up question rather than guessing. Options are presented as clickable buttons. Select one and the bot continues with full context from the previous turn.

How Does the Clarification Flow Work?

Clarification prevents the most common failure mode of data bots: returning results from the wrong dataset without telling you. When the AI finds multiple plausible matches for a request, it stops and asks before proceeding.

Here is an example of the clarification flow in practice:

You:/knowi show me sales data
Knowi: I found 3 sales datasets. Which would you like to use? 
[Product Sales] [Regional Sales] [Online Sales]
You: Click [Product Sales]
Knowi: Returns chart and data table from Product Sales

Conversation context is maintained across the clarification exchange. A follow-up like “now break that down by quarter” will continue to use Product Sales without requiring you to re-specify it.

How Does Conversation Context Work?

By default, the bot maintains conversation context within a channel. Follow-up questions reference earlier results without you needing to repeat the original query. This makes the interaction feel like a data conversation rather than a series of isolated commands.

If you want to start a fresh session – clearing previous context – add -n to your command:

/knowi -n what is total revenue?

This is useful when switching between unrelated data questions in the same channel or when a previous conversation has accumulated context that is no longer relevant to the current request.

How Is Agentic BI in Slack Different from a Standard BI Bot?

Standard Slack integrations for BI tools typically push pre-built reports to channels on a schedule. The /knowi command is different because it connects to a live agent layer that can create new assets, not just retrieve existing ones.

How Do You Enable Agentic BI in Slack?

Agentic BI in Slack requires three things to be in place before the /knowi command becomes available:

• Knowi Slack bot installed in your workspace. A workspace admin completes this setup through the Knowi admin panel.
• Your Knowi account linked to your Slack user. Each user authenticates individually through Knowi’s user settings.
• AI Agents Access enabled on your account. Navigate to Settings, then User Settings, then AI Settings, and toggle on AI Agents Access.

Agentic BI in Slack is enabled by default for Knowi cloud accounts. For on-premises deployments, contact your Knowi administrator to enable the integration. Full setup documentation is available in the Knowi AI tools docs.

To understand how the same agent layer works through other interfaces – including Claude and other MCP-compatible tools – see the Knowi MCP server guide.

Frequently Asked Questions

Ready to run your BI workflow from Slack? Start a free trial or talk to our team.