I’m 46. I spent 30 years on the tools. I joined Nate’s AIS Skool community in October last year. This morning a UK construction MD emailed me four words I’ll remember for the rest of my life:
“Can we proceed with phase-1 please.”
Here’s how it happened.
Today, the MD of a UK construction firm emailed me four words I’ll remember for the rest of my life.
“Can we proceed with phase-1 please.”
£4,500 Phase 1 build. 50% deposit on start, 50% on delivery, two-week go-live guarantee. Signed verbally Tuesday evening, confirmed in writing this morning.
This is not a victory lap. This is a story for anyone in this community sitting where I was sitting back in October when I first showed up. If I can do this, you absolutely can.
Here’s how it actually happened.
Where I came from
30 years on the tools. 20 of them co-founding a construction firm that grew from 2 people to 70. I’m 46. I started teaching myself AI in early 2025. Properly self-taught. No degree, no tech background, just YouTube and reading and breaking things and fixing them.
The timeline that matters for this room:
- 23 June 2025 I joined Liam Ottley’s AAA Skool community. Still a member today.
- 15 September 2025 I incorporated ConstructionX AI Ltd at Companies House.
- October 2025 I joined Nate Herk’s AIS Skool community.
- November 2025 I joined AIS+.
- Late January 2026 I stepped up into Liam’s paid AAA Accelerator tier. Still a member today.
- 13 February 2026 the workspace went live and I committed to two posts a day.
- 14 May 2026 (today) the first £4,500 build signs through the productised funnel.
The brand had been quietly forming for months before the daily build started. Domain registered, soft conversations, early demos for friends. The visible push did not really begin until February.
So:
- ~14 months teaching myself AI
- 8 months since incorporation
- 7 months in this room
- 6 months in AIS+
- 3 months of daily public building
- Today, the first £4,500 build
Not slowly, not quickly. Just relentlessly.
The work that built up to today
While I was learning, I was building. Out loud.
- Two posts a day on LinkedIn, X, Facebook, Skool, every weekday plus weekend Groundwork. About 3 months of daily consistency on top of the lighter posting before that.
- 357 warm contacts in a tracker. Every reply logged, every follow-up scheduled.
- One YouTube long-form a week since February.
- An internal Opportunity Map on my own business (the diagnostic I now sell to others).
- A pilot Opportunity Map for a family member’s business (free, validated the format).
- An entire “Front Door Stack” of services built and dogfooded on my own site: a voice agent on my homepage, a WhatsApp voice-to-quote system, an internal CRM stack replacing parts of my GoHighLevel setup, a content pipeline that drafts posts overnight.
- Five public Claude Skills built for UK construction workflows.
- A flagship client called Steel Serenity already on the books from before the productised funnel existed. A website bringing 13 independent local businesses together under one roof. 5-star testimonial. Proof point #1.
Stack underneath all of it: Claude API (Sonnet + Haiku), n8n cloud, Google Workspace, Supabase, Wispr Flow, Netlify, Mac Mini M4 Pro always-on node. No third-party black boxes I cannot inspect.
I tell you the workload because that’s the part nobody posts about. The grind under the wins.
The Opportunity Map
This is the thing.
About 6 weeks ago I locked in the productised version of what I sell. Every paid build now starts with an Opportunity Map: a 45-minute paid diagnostic where I sit with a business owner, audit every recurring task across 9 areas of their business, score each task for automation potential and value impact, and deliver a visual heat map plus a phased build plan.
The pitch I lead with:
“Before you spend a penny on automation, you need to know what to build. Not what some generic consultant recommends. What’s right for your specific business, your specific bottlenecks, your specific team. That’s what the Opportunity Map gives you.”
Lite tier £497 (single process). Full tier £1,500 (whole business). Both credit in full toward the build.
The Map becomes much easier to sell once a business owner sees their own numbers laid out on screen.
The session that became Phase 1
Last Friday, 8 May, I drove to a UK construction firm I knew through my old firm. Pre-existing relationship. What we call a warm prospect.
I want to tell you about the drive over because it’s the bit nobody talks about.
I’d practiced the live demo on my own business and on a family member’s pilot Map. Both times the room was friendly, expectations were low, and a stumble would have been fine. This was different. Four stakeholders waiting at the table: MD, estimator, project manager, trainee. A real working construction firm. People who have spent decades watching tech vendors over-promise.
I rehearsed every failure mode the whole way over.
What if Wispr Flow does not pick up his voice over the noise in their meeting room. What if their wifi drops in the middle of the live build. What if Claude takes 30 seconds to respond instead of 10 and the silence kills the moment. What if the AI hallucinates a price or fabricates a detail and I have to explain it on the fly. What if the foreman just folds his arms and says “this is the same crap the last lot tried to sell us in 2019”.
I have 30 years of construction reputation sitting behind me. I would rather shut this agency down than build it on smoke and mirrors. That weight does not let go on the drive over. This was not a sales call. It was a builder asking another builder to trust him with the next chapter.
Pulled in. Sat in the van for a minute. Walked in.
The Map itself took most of the session. Standard structure: walk through 9 areas of the business, score every recurring task on automation potential and value impact, sketch the heat map live as we went. By the end the four of them were looking at their own week laid out in colour-coded boxes for the first time.
Then the demo moment.
First demo: pre-baked.
I’d spent 30 minutes that morning researching a real job they’d actually delivered. Built them a custom 1-page job pack for that real job in their own brand voice. Saved it locally before I left home. When the demo moment came I opened it on screen as if I’d built it that morning.
“Here’s what last week’s paperwork could have looked like.”
They leaned in.
Second demo: live.
This is the bit I had actually been nervous about. I handed the MD my phone with Wispr Flow open.
“Talk me through a job. Any job. Make it up if you want.”
He started slow, then warmed up. Site address. Scope. Hazards. Materials. Dates. Names of his lads.
The dictation finished. The Claude prompt fired. I watched the progress bar crawl across the laptop screen, aware of every second the room was waiting in silence. Then the PDF appeared on the second screen. In his voice. In his format. On a job that did not exist before he said it out loud about 90 seconds earlier.
The MD read it. His team read it. The conversation that followed in the room was theirs, not mine.
That was the moment.
I closed the laptop. We talked through the build options for another twenty minutes. They walked me out.
The close call
Four days later, Tuesday evening 7pm, I had the close call.
One hour before that call, from 6pm to 7pm, I sat with a sales coach I work with as part of an agency programme I’m in. He walked me through the structure of the call: disarm-isolate-challenge, the “would you be opposed” commitment question, locking the diary date live on the call rather than letting the prospect take it away to think about.
I want to be honest with this community. The call closed because the relationship carried it more than my close did.
I went with a softer version of his structure (“let me send the summary over and see what you think”) instead of his sharper version (“would you be opposed to getting started next week”).
With this prospect it worked. With a colder one it wouldn’t have.
That’s a real lesson, captured in the post-call review I wrote that same night before the proposal email went out.
But the verbal yes in principle landed.
Two and a half hours later, 9:43pm Tuesday, the proposal email went out. Phase 1 build. £4,500. Four paths beyond Phase 1. Two-week go-live guarantee. Travel windows openly disclosed (JSAI in Tokyo, US AI/business visits later in the summer).
Two days later, this morning, the reply landed:
“Looks good. Can we proceed with phase-1 please. I’d like to get phase-1 up and running and get the lads used to using it. I’d like to get this bit well oiled before adding too much more. I’m back on Monday so happy to catch up then?”
What helped
A stack of things, layered, with deep gratitude to this room over the last 7 months.
- @Nate Herk and the AIS community. Specifically the 5 Levels of Claude framework that’s seeding the long-form I’m recording next Friday on adapting the levels for UK trades. The Level 4 plan-mode-as-default discipline from your walkthroughs is now baked into our process-compliance rule. Both pulled directly from you.
- The AIS community itself. Every builder in this room posting what shipped, what broke, and what they tried next. That public-build standard is what kept me from drifting into closed-doors agency mode through the soft brand-build months before the visible ramp started.
- AIS+ specifically. The deeper builds and walkthroughs in that tier raised the technical bar I held myself to.
- A peer in the trades-adjacent space who told me over coffee in a Norfolk village deli that the authenticity moat is real, and the writing has to sound like me, not a language model. Validated every anti-slop discipline I’ve built since.
- The sales coach who spent an hour with me before the close call. His structure shaped the conversation even when I softened his sharper close into my register.
- Every other builder in AIS who quietly posted their own grind through the last 7 months. Watching you all kept me showing up.
Massive thank you to Nate and the whole AIS community. The depth of build talk in this room is rare. Keep it going.
What I’d say to anyone earlier in the journey than me
The same things you’ve heard everyone in here say. They are all true.
- Consistency compounds. Three months of daily posts looks like nothing on day 30 and like a brand on day 90.
- Sell clarity, not tools. The Opportunity Map sells because it’s a diagnostic, not a pitch.
- In-person beats Zoom for trades. The room does the work.
- Get coached. The hour before the close call was the hour that mattered.
- Be honest in your post-call reviews. Your win is not the same as your lesson.
- Don’t fake the voice. If you can’t write it like you, write less of it.
- Soft start is fine. The company existed for 5 months before I started building publicly. That groundwork mattered.
If you’re earlier in the journey than me, staring at the dashboard, doing the warm outreach, posting into what feels like silence, wondering when the first proper paid one lands, please believe me.
It lands.
Show up tomorrow.
Right. Back to building.
Phase 1 kicks off w/c 9 June after JSAI Tokyo. Will share back through the build what works and what breaks.