Revenue AI Lab. Book the audit
All work
CASE 01 // MSP / CYBERSECURITY // SHIPPED Q4 2025

A 2-founder security firm, rebuilt as an AI-run revenue engine.

MSP Pentesting ran on cold outbound and MSP referrals. Every hour the founders spent on marketing was an hour not spent on pentests. We rebuilt the inbound engine from zero — programmatic SEO, AI outbound, two internal agents — and handed the code to the team. Thirty hours of founder time back per week, and the system is still running.

Timeline

6 months

Pages shipped

420+

Reply rate (outbound)

6.8%

Founder hours back

30 /wk

/ The problem

Every marketing hour was an hour not billing.

MSP Pentesting is a two-founder security consultancy running network pentests and red-team engagements for mid-market companies and the MSPs that serve them. Their highest-leverage hour is a billable one — running an engagement, reviewing findings, briefing a client. Every hour they spent writing cold emails or pushing out a blog post was an hour not on the clock.

By Q3 2025, the pipeline was one of three things: referrals (unpredictable), cold outbound the founders wrote by hand (brittle), or the occasional inbound lead from a blog post they'd written months earlier (rare). Growth was real, but it was bottlenecked by the founders' calendar.

The ask wasn't "generate us more leads." It was: make the revenue engine something that doesn't eat our week.

IMG / PLACEHOLDER

Pre-engagement: the outbound spreadsheet that ate Friday mornings

Swap this placeholder with a real screenshot of the founder-maintained prospect sheet, with names blurred.

/ What we built

Four systems, wired as one engine.

Scope was fixed up-front. The team owns the code, the DNS, the inboxes, and the agent definitions. Nothing we built is locked behind our accounts.

  1. 01 / RANK

    Programmatic AI SEO

    420 pages across three page types — "pentest for [industry]", "MSP security for [state]", and a long-tail comparison grid against the 15 biggest security firms. Every page is human-reviewed, schema-marked-up, and internally linked. First organic pentest inquiry landed in week 9.

  2. 02 / WRITE

    Founder-voice content pipeline

    Every Friday the founder records a 20-minute audio memo on whatever they learned that week. A transcription → outline → draft agent pipes that into a 1200-word post in the founder's voice. They spend 15 minutes editing. One post ships every week without fail.

  3. 03 / REACH

    AI outbound with signal triggers

    Three warmed sending domains, rotating inboxes, SPF/DKIM/DMARC done right. A research agent watches funding announcements, compliance events (SOC2, HITRUST renewals), and MSP hires. When a signal fires, the agent drafts a message referencing the specific event and sends it from the matching inbox. 6.8% reply rate at steady state.

  4. 04 / AUTOMATE

    Two internal agents

    A research agent that takes a prospect's domain, reads their exposure surface, and drafts a pre-pentest brief. A proposal agent that takes the brief + call notes + scope doc and outputs a signable SOW. Proposals that used to take two hours now take twenty minutes of review.

/ System architecture

How the four systems talk.

MSP-PENTESTING / REV-ENG ARCHITECTURE v.04 · LIVE INPUT SIGNALS PROCESSING OUTPUT OWNED BY Funding rounds (Crunchbase) SOC2 / HITRUST events MSP hires (LinkedIn) Org growth signals Founder Friday memo AGENTS Research agent (Claude) Outreach writer (GPT-4) Content pipeline Proposal drafter Warmed inboxes → prospect /blog/* (programmatic) /services/* (long-tail) Slack → founder review HelloSign → signed SOW MSPP MSPP MSPP MSPP MSPP 5 SIGNALS 4 AGENTS 5 OUTPUTS ∑ 1 ENGINE
IMG / PLACEHOLDER

Week 20: the organic traffic chart after the programmatic SEO went live

Drop in a Google Search Console or Plausible screenshot showing the compounding traffic curve.

/ Results

Six months in.

Inbound leads / month
17 from 0 at kickoff
Outbound reply rate
6.8% industry avg ~1.5%
Founder hours back
30 /wk per founder
LLM citations
42 tracked in ChatGPT + Perplexity for "MSP pentesting" queries
Code ownership
100% handed off to MSP Pentesting's GitHub

"Our marketing went from a side-job to infrastructure. I check the dashboard once a week and the rest of my time is on engagements."

Connor Cady · Founder, MSP Pentesting
/ Takeaways

What this engagement proved.

  1. 01

    An AI system that actually compounds is four systems wired together — not one "AI tool." The moat is the wiring.

  2. 02

    Founder voice is defensible. AI that writes like the founder — sourced from the founder's own raw thinking — reads as authentic. Generic AI content does not.

  3. 03

    Owned infrastructure beats SaaS. The cost of running this system is a fraction of the Outreach + Apollo + Clay + 11x stack it replaced.

  4. 04

    The goal of an engagement like this isn't "leads." It's getting the founder's highest-leverage hour back. Everything downstream of that is easier.

Book the audit

Want the same engine?

Thirty minutes. We audit your current revenue stack and send a written scope within 48 hours.

  • Free
  • No deck
  • Written scope in 48 hrs