Business Intelligence Reporting Agent

Wrap your reporting API as MCP tools

Create MCP tools that map to your most common reporting queries. You don't expose raw SQL — instead you expose safe, pre-defined query templates as named tools.

Tools to expose via MCP:
- get_revenue_by_period(start_date, end_date, group_by)
- get_top_products(period, limit, region?)
- get_customer_acquisition(channel, period)
- get_support_metrics(period)
- get_inventory_levels(category?, warehouse?)

Secure your MCP server

Your MCP server runs inside your network and only exposes the specific query tools you define. No raw database access is given to the AI — it can only call the tools you whitelist. Add authentication so only SentientOne can call your MCP endpoints.

Register in SentientOne

Add your MCP server URL in the SentientOne dashboard. The platform auto-discovers all your tools and makes them available to assign to agents.

Create the BI Reporting Agent

Create an agent with a system prompt that defines the business context and output format preferences.

System Prompt:
You are an internal business intelligence assistant for [Company].
You have access to live business data tools. When answering questions:
- Always fetch live data using the available tools
- Format numbers clearly (e.g. $1.2M, not 1234567)
- If comparing periods, calculate % change and highlight trends
- Keep summaries concise — lead with the key number, then context
- If a query is ambiguous, ask one clarifying question before proceeding

Attach your BI MCP tools

Connect your MCP server to this agent. All your reporting tools are now available. The agent intelligently selects which tool to call based on the question.

Set access controls

Generate separate API keys for different teams (Sales, Ops, Finance). Each key can be scoped to a specific agent version — so Finance gets the full agent while Sales gets a version without sensitive cost data.

Slack integration (most popular)

Create a Slack bot that passes messages to SentientOne and returns the response. Your team types questions directly in a #bi-assistant channel.

// Slack bolt app handler
app.message(async ({ message, say }) => {
  const res = await fetch("https://app.sentientone.ai/api/chat", {
    method: "POST",
    headers: {
      "X-Agent-Id": "YOUR_BI_AGENT_ID",
      "X-App-Key": "YOUR_APP_KEY"
    },
    body: JSON.stringify({
      message: message.text,
      sessionId: message.user
    })
  });
  const { reply } = await res.json();
  await say(reply);
});

Internal portal integration

Embed a chat widget in your internal admin portal or intranet. Route queries to the BI agent. The sessionId ties follow-up questions to the same context — so 'break that down by product' works after a revenue question.

Scheduled reports (optional)

Call the agent on a schedule (cron job) with predefined questions like 'Give me a daily trading summary.' The agent fetches live data and returns a formatted summary you can send via email or Slack.