First User Ran a Real Campaign with the Outbound Playbook MCP — Here's What Happened

Two days ago I shipped the Outbound Playbook MCP server — an alpha that packages Prem Kumar's "-1 to 0" founder outbound methodology into Claude Code tools. I wrote about the build. Today I'm writing about something more interesting: someone else used it for real and got a reply.

The Message That Made My Morning

In our AIBoomi Playbook Roundtable group (about 50 founders and builders), Sri dropped this: "@Mahesh CR — Tried my first campaign today and it works like a charm."

Not "tried installing it." Not "played around with it." He ran a live campaign. On Windows, with Sonnet 4.6. Against real prospects.

What Sri Actually Built

His use case was immediately more specific than anything I'd tested with. He sells Arete, an AI training companion for hospitality staff — think practice real guest scenarios on your phone instead of sitting through classroom training. The vertical: hotel chains in Chennai.

He created a campaign called "Hotel Chains — Replace Classroom Training with Arete" and loaded five Tier 1 prospects:

  • Vinod at Taj Coromandel
  • Murali at Radisson Blu
  • Anirudh at Hilton
  • Lakshmi at Hyatt Regency
  • Saravanan at ITC Grand Chola

The playbook generated a full 14-day outbound sequence for each: Day 1 email plus LinkedIn connection, Day 3 LinkedIn DM, Day 5 follow-up call and second email, Day 7 third email, Day 10 call with voicemail script, Day 14 breakup email. All personalized. All signal-based.

Here's the part that matters: the subject lines weren't generic. For Lakshmi, who'd recently been promoted, the system generated "new chapter at hyatt" — lowercase, casual, the kind of subject line that actually gets opened because it reads like it's from someone who knows you.

The Reply

Lakshmi replied with interest.

And this is where the methodology layer proved itself. The playbook didn't just track the reply — it guided the next step. "Don't pitch. Book the meeting." That's Prem's framework encoded as a tool response, not just documented in a PDF somewhere. The difference between knowing the methodology and having it actively shape your workflow in real time.

Sri's prospect tracking table showed all five contacts at different stages: some waiting for LinkedIn acceptance, one reply received, sequences in progress. The whole pipeline visible inside Claude Code.

Why This Matters More Than the Build

I've shipped tools before. The build is the easy part — especially with Claude Code doing the heavy lifting. What's hard is the jump from "works on my machine" to "works in someone else's hands, on their problem, with their data."

Sri didn't need hand-holding. He installed the MCP server, pointed it at his real business problem, and ran it. The methodology transferred. The tool worked outside my terminal. That's the validation that matters for an alpha.

What the Community Wants Next

The feature requests came fast:

Email/read receipts — know when prospects open your emails. Makes sense for timing follow-ups.

AI avatar calls — automated voice outreach. Ambitious, but the bones are there in the sequence engine.

Mini CRM — prospect state management beyond what the current JSON tracking handles. Another group member immediately offered to open-source a lightweight CRM as an additional MCP tool. That's the kind of contribution that makes open-source worth doing.

Where This Goes

The immediate roadmap writes itself: npm distribution so installation is a one-liner, CRM integration (likely the community-contributed lightweight version first), and better prospect state management. The alpha proved the concept. Now it needs to be installable without reading a README twice.

The bigger idea hasn't changed. Methodology-as-tooling is an underexplored pattern. Every founder, every consultant, every practitioner has frameworks they've refined over years. Most of those frameworks live in slide decks and PDFs. MCP servers can turn them into active guidance — tools that don't just automate steps but encode the thinking behind when and why to take each step.

Sri's campaign worked not because the emails were well-written (though they were). It worked because the sequence, the timing, the response handling — all of it followed a methodology that Prem spent years refining. The MCP server just made it executable.

That's the pattern worth paying attention to.