I automated my own search firm. Here's what I learned.

Solo founder. Small team. Two arms, two legs. If you've run a boutique search firm, you know the maths.

You're trying to do justice to a live retained search — briefing, market mapping, long list, screening, shortlisting, the closing dance — and keep BD warm at the same time. Both jobs need your A-game. Neither rewards a half-effort.

So small firms slide into feast-or-famine. Heads down on delivery for three months while BD goes cold. Then heads up on BD while delivery slips. Rinse, repeat, burn out.

If I was going to grow — or even survive — I had to amplify my productivity. Not by 20%. By a multiple.

The old playbook said hire another recruiter. But hiring is slow, expensive, and gets you linear gains at best. I needed a moonshot.

Here's where it cracked open.

I wasn't on the LinkedIn Recruiter licence — couldn't justify the spend as a solo founder. I was on Sales Navigator, doing what every under-licensed recruiter does. Copy, paste, send. Copy, paste, send. Personalising tokens by hand.

Friday afternoon. End-of-week beer waiting. Copy-pasting the same message to fifty people on LinkedIn. Asking myself: what am I actually doing?

That's the moment. Every founder I talk to has one. Yours might be the 11pm Sunday spreadsheet, or the brief you've rewritten three times at 30,000 feet. The shape changes. The frustration is identical.

First answer: Dripify — a LinkedIn outreach automation tool. Godsend. Saved me hours a week. But the longer I used it, a different question started forming: why am I paying for someone else's tool when I could build my own?

Not because Dripify wasn't good. Because once you taste automation, you start seeing every other piece of your week the same way. Outreach. Screening. Reporting. Pipeline updates. All manual. All automatable.

So I stopped buying tools and started building a system.

So I did what every 'techie' would do before automating an entire firm.

Pen and paper.

Yes, I bought a giant drawing pad, a bunch of post-it notes, and some coloured markers, and blocked out three days.

Then I tore my own firm apart on it. Every workflow, hand-drawn. Every process, broken into micro-steps. The systems used, the triggers, the inputs, the outputs, the handovers, the pain points. Where the time leaked. Where the quality dropped. Where I was doing the same thing twice because two tools didn't talk to each other.

Then I scored every step on a value matrix — what's high-value-only-Peter, what's high-value-but-anyone-could-do-it, what's pure admin drag, what could be automated tomorrow.

I'd essentially audited my own business.

What I drew on those sheets of paper is exactly what I now run for other recruitment founders. I call it the AI Performance Audit™.

The build that came out of that audit had four rules I refused to break:

Rule one — first principles, before tools.

Identify the biggest problem you're trying to solve. Describe what good looks like before you build a thing. No automating low-impact admin for the sake of it. No buying a tool because it's trending on LinkedIn. Only the highest-leverage problems earned a build.

Rule two — live searches only.

Every workflow had to survive a real candidate, a real client, a real deadline — with client data protected throughout. No sandbox. No demo data. If it didn't work on a paying search, it didn't ship.

Rule three — one module live before the next.

Start small. Iterate in production. Each module shipped, ran on a real search, produced an artefact, then informed the next. No parallel builds. No building ahead of validation.

Rule four — Customer One from day one.

Designing for one firm is a dead end. Designing for two is a Plugin. So from module one, I built with a Customer One in mind — a second search firm who'd use it. Even if for free. Even if only part of it. The moment another founder tries it, you begin to see what really works.

Four rules. One discipline. That's why Customer Zero isn't a marketing concept — it's the operating system of the firm.

The Metamorph OS runs on eight tools, glued together so they actually talk to each other. Every one of them earns its seat.

Six modules. One learning loop.

The Metamorph OS isn't a tool. It's a recruitment firm in a box — six departments, end to end. Plug it in, and you've got the operations of an AI × Human firm running underneath you.

The learning loop was the discipline. Ship. Run live. Capture the artefact. Inform the next. That's how all six came online — and how they keep getting sharper.

Twelve months in. Here's where it stands.

What surprised me.

The magic isn't any single tool. Any firm can buy AI tools. The magic is wiring them together so they move as one. Honestly, it's unbelievable. You trigger it on a live search and just… the computer doing what used to be your week. Long list, profiles, shortlist, outreach — all running while you make coffee!

Every founder I've shown gets the same reaction: it rewires what you think is possible.

The second-order gain: admin drag, gone. The hours back didn't go into more sourcing — they went into BD, client conversations, forecasting, running the firm.

Where I had to backtrack.

Four rebuilds along the way. CRM moved from GoHighLevel to Attio for an AI-native foundation. Candidate reports went from automated PowerPoints to live dashboards. Automations shifted from n8n to Claude as the models got good enough. And the website itself — rebuilt from scratch in Claude Code (the one you're reading). All four are paid-up assets now. I'd take the same trades again.

What I'd do differently.

If I were starting again, I'd open the doors sooner. Two or three friendly founders reviewing as I shipped would have sharpened things faster than my own eyes. Rule four was on the wall day one. I still didn't move fast enough.

What this means for you.

Every recruitment founder I talk to says the same: too much time working in the business, never enough on it. The automation flips that.

Doing it yourself is doable. Just hard — and it doesn't end at twelve months. Every skill, every workflow, every model update needs ongoing work.

You could hire in-house. Find someone with AI expertise, recruitment context, and a salary you can justify. Then carry that salary every month, whether the work is there or not.

Or do what your own clients do every time they brief you — outsource the expertise. Stay focused on what you're an expert at: recruitment.

This is one founder, twelve months in. Picture a similar OS across your team — five recruiters doing the work of fifteen. AI doing the heavy lifting, recruiters doing what they do best. That's the AI × Human firm. This is the floor, not the ceiling.

What you've just read is the Metamorph OS playbook. Now I run it for other recruitment and executive search founders. Not as theory — as the lived, every-day-I-fix-a-bug practice.

Three services:

• Strategy — AI Performance Audit™. From $5,000. Know where AI fits and what to automate first, before you spend on implementation.
• Process — Implementation Sprint. From $10,000. 6–8 weeks of hands-on build. The right stack for your firm — defined in the Audit, built in the Sprint.
• People — Flow Optimisation. From $2,400/mo. Because AI tools without AI-capable people is expensive shelf-ware.

You already know AI is rewriting recruitment. You can feel it on every brief, every Friday afternoon, every late-night spreadsheet. I work best with founders who've decided not to be the ones who got left behind. The ones who'd rather be the case study than read one.

What recruitment firm will you wish you'd built five years from now?