Revenue Operations War Room

The Revenue Operations War Room blueprint demonstrates one of the most ambitious patterns on the platform: multi-agent time-scale architecture. Three agents operate on different cadences - trigger-based, daily, and weekly - sharing a persistent revenue knowledge base through a common Space. Each agent has its own skillset with role-specific abilities, and they share a dedicated Revenue Knowledge Base skillset that any agent can install on demand to read from and write to the same set of shared files.

The Problem

Sales managers and RevOps teams spend their days bouncing between HubSpot, Stripe, email, calendar, and Slack. They manually prep for calls, chase payment follow-ups, update pipeline forecasts, and write status updates. These are high-value activities that follow predictable patterns - exactly the kind of work that benefits from agent automation.

Three Agents, Three Cadences

The blueprint is built around the idea that different revenue operations tasks have different time horizons:

• Pipeline Watcher (every few hours) - the always-on sentinel. It polls HubSpot for deal stage changes, stale deals, and new opportunities. When it detects something noteworthy - a deal that has not moved in 14 days, a new high-value opportunity, a sudden stage regression - it updates the shared pipeline state file and posts an alert to Slack. This agent uses HubSpot read abilities (deal list, deal search, contact fetch) and Slack messaging to keep the team informed in near-real-time.

• Meeting Prep Agent (daily, morning) - the context builder. Each morning it reads Google Calendar for the day's meetings, cross-references attendees against HubSpot contacts and deals, pulls recent Stripe payment history for those accounts, and assembles a briefing document. The briefing lands in Slack before the first call. This agent synthesises data from four different services into a single actionable document - the kind of prep work that takes a human 20 minutes per meeting.

• Revenue Analyst (end of day or week) - the strategist. It reads the pipeline state file, reviews payment data from Stripe, considers meeting outcomes, and produces a narrative revenue report. It detects trends: deals stalling in a particular stage, revenue acceleration from a segment, at-risk accounts with overdue invoices. The report goes to Slack and gets saved to the shared space for historical reference.

Shared Revenue Knowledge Base

The agents communicate through a shared Space containing evolving files:

• pipeline-state.md - current snapshot of deal pipeline, updated by the Pipeline Watcher
• daily-briefings.md - rolling log of meeting prep briefings
• revenue-reports.md - weekly narrative reports from the Analyst
• contact-intelligence.md - accumulated notes on key contacts and accounts

This file-based coordination pattern is simple but powerful. Each agent reads what the others have written, building on accumulated context rather than starting from scratch every time. The file abilities live in a shared Revenue Knowledge Base skillset that each agent installs on demand via a conversation/skillset/install[by-id] ability - keeping the per-agent skillsets focused on their role-specific tools while avoiding duplication of file access logic.

Why This Pattern Matters

Most agent blueprints are single-bot, single-task designs. This one demonstrates that ChatBotKit can orchestrate multiple specialised agents that collaborate through shared state. The Pipeline Watcher does not need to know about calendar events, and the Meeting Prep Agent does not need to track deal stage changes over time - each agent focuses on what it does best, and the shared files provide the coordination layer.

Market Context

HubSpot's own "Breeze" AI agents (launched 2025) prove massive demand for CRM-native AI automation. Revenue intelligence tools like Gong and Clari are valued at over $1B. This blueprint brings that same capability to any team using ChatBotKit, with the added advantage of being fully customisable and composable.

Getting Started

1. Configure secrets for HubSpot, Stripe, Google Calendar, Google Mail, and Slack.
2. Each bot is pre-configured with its own skillset and a shared Revenue Knowledge Base skillset that can be installed on demand. Deploy the blueprint and the three agents are ready to operate.
3. The Pipeline Watcher and Meeting Prep Agent can be triggered on schedules. The Revenue Analyst can run on demand or at end of day.
4. All three agents share the Revenue Knowledge Base space - check the files to see the accumulated intelligence grow over time.

Backstory

Common information about the bot's experience, skills and personality. For more information, see the Backstory documentation.