Founder's Story

Why I Built WisyPlan

It didn't start with a business plan. It started with an empty chair at a school concert.

The night this became personal

My son's school band concert. He played beautifully — the whole orchestra did.

Then, just after the applause, another boy started crying.

His dad wasn't there.

I knew his dad, Martyn. I texted him straight away: why aren't you here?

The reply was simple, and it has stayed with me ever since:

"I'm still at work. I'm trying to figure out who approved a change three months ago."

Three months of emails, RFIs, Teams messages, revised programmes and monthly reports — each one telling a slightly different version of the same story. A system full of data and empty of answers.

Martyn wasn't careless. He wasn't disorganised. He was trapped.

And while he scrolled and searched, his son walked onto a stage for the first time, and the chair next to his wife stayed empty.

That night I didn't sleep. A few days later I started experimenting — not with pitch decks, with a question:

What if AI could turn project chaos into clarity, automatically?

That was the moment WisyPlan was born.

What I actually saw

I've spent fifteen years in this industry — engineering, digital delivery, major infrastructure, standardising workflows across six global regions. I know how good these people are.

So the conclusion I reached wasn't about competence:

"Projects don't fail because people are incompetent. They fail because their data doesn't talk to each other."

A project manager spends three to four hours a day triaging email, and roughly one full day a week manually pulling data together just to answer questions the project already knows the answer to. Eighty percent of large construction projects run over budget. Not for lack of information — for lack of a way to make that information speak.

Every tool in this industry stores something. Almost none of them connect anything.

I wrote about this in detail: Your PMs are losing 1 full day a week. Here's the fix. →

Before WisyPlan, I built something I had to stop

Ten months earlier I built a different product. It was called Will & AI — an AI assistant that helped parents understand what their children were doing online.

It won the Microsoft for Startups AI Hackathon in San Francisco. Parents loved it.

Then my son worked out what it was.

He stopped using it that day. Not because it didn't work — because it worked exactly as designed, and once he understood the design, he didn't consent to it.

That was the most valuable thing that happened to me that year.

I'd built AI that watched people on behalf of someone else. The parents were the customer. The child was the subject. And the moment the subject understood the arrangement, it collapsed — because it should have.

I learned something I now treat as non-negotiable: if AI is going to work alongside humans, the ethics have to be in the architecture, not in the marketing. Not a policy page. Not a compliance checkbox. A design constraint from line one.

So when I turned back to my own industry, the question wasn't "where can AI extract the most value?" It was: what would an AI look like if it were built to protect people's judgement rather than replace or monitor it?

That's the difference between WisyPlan and a surveillance tool wearing a productivity badge. WisyPlan reads what the project already produces and gives the answer back to the person accountable for the decision. It doesn't score people. It doesn't rank them. It gives them their afternoon back.

What WisyPlan is

WisyPlan is the AI operating system for construction projects.

It connects the systems a project already runs on — email, documents, programmes, site data, the tools you've already bought — and turns them into one layer that can actually answer a question. Who approved that change. What moved this week. What's about to go wrong.

It doesn't replace Procore, Aconex, SharePoint, Outlook or ACC. It sits above them, neutrally, and does the one thing none of them can do alone: make them talk.

On why building your own is usually the wrong answer: The "Build vs Buy" Debate Is the Wrong Question for Construction AI →

What I got wrong

I targeted the wrong customer first.

We went straight at large enterprises. And it worked — in the worst possible way. Everyone was interested. Every conversation was warm. Every demo landed.

And nothing closed.

Not because of the product, and not because of the price. We couldn't clear IT review. In this industry the constraint on an early-stage company isn't whether the value is obvious — it's whether procurement and security can say yes to you at all. I was selling to people who wanted to buy from a vendor they weren't allowed to buy from yet.

So I rebuilt the sequence instead of the pitch. ISO 27001 certification in progress with an accredited auditor. CASA Tier 2 verification. Penetration testing completed. Enterprise-grade infrastructure. And a deliberately narrow entry point — one agent, one workflow, one project — instead of a platform-wide ask that triggers a twelve-month governance review.

I lost months learning that. I'd rather write it down than pretend the path was straight.

The part nobody tells you

Everyone warns you about the money, the hours, the uncertainty. Those are real, and they're survivable.

Here's the one that actually matters: if you don't love the problem itself, you quit within a few months.

Not because it gets hard. Because it gets boring — the tenth security questionnaire, the silence after a good meeting, the week where nothing moves. Passion for the idea evaporates around month three. What's left has to be the problem.

I left a secure job. I've put in my own money, my time, and time that belonged to my family. Eight months in, against a difficult market, I'm still here — because I've seen the empty chair, and I'm certain this gets solved. It's only a question of when, and by whom.

Where this goes

Three years out, in concrete terms:

A project manager opens WisyPlan in the morning and already knows what changed in the last twenty-four hours, without opening a single email.

Beyond that: agents that don't just organise information but hold discipline expertise — structural, geotechnical, MEP — so the reasoning happens where the knowledge is.

And the real test: the day someone tries to remove WisyPlan and realises they'd be deleting the project's memory.

Not a tool projects use. The layer projects run on.

I read a line on New Year's Day that described the journey I was already on before I knew I was on it:

The reasonable man adapts himself to the world. The unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man.

— George Bernard Shaw

I'm not trying to adapt to an industry where being good at your job means missing your son's concert.