Built on My Own: Do I Need AI and wecalledit
Two live products I designed, built and shipped end to end, on my own.
- Makes ideas tangible
- Tests assumptions early
- AI and data
- Ships
- Where
- Independent
- When
- 2026
- My role
- Designer, builder, solo
- Themes
- Solo buildAI-nativeRestraintLive product
2 live
products shipped solo, AI-natively
-
01Situation
Alongside the big programmes, I build. These are 2 live products, and each one says something about how I work.
-
02The move
doineedai.co.uk walks people through whether they need AI at all, and it is designed to sometimes say no. wecalledit.app was framed, researched, designed, built and deployed by me alone, using AI tools to speed up the making without handing over the judgement.
-
03Result
Both are live. They are my evidence that I can take something from nothing to shipped on my own, and that I stay close to real delivery rather than drifting into pure strategy.
Two live products, side by side
Both products, live, side by side. One is a diagnostic you can take right here.
“A tool designed to sometimes tell you no.”
Take the diagnostic in the frame. It is the real thing.
- Framed
- Designed
- Built
- Shipped
1 person. AI-native. Live.
The discipline
One tells people when AI is not the answer.
The other proves I can ship alone.
Together they are how I keep both muscles working.
The situation
Alongside the big programmes, I build. Two live products that both say something about how I work.
Do I Need AI: the discipline of saying no
doineedai.co.uk came out of something that genuinely bothered me. I kept seeing people, and organisations, reach for AI as the answer to problems that did not need it, where something simpler would have worked better, cost less, and broken less often. AI had become the reflex rather than the considered choice.
So I built a tool that walks a person through the honest question: do you actually need AI for this, or would a simpler, non-AI approach fit the problem better? It is deliberately not a sales funnel for AI. It is the opposite. It is a tool designed to sometimes tell you no.
I care about this enough to have spent my own time building it. Knowing when AI is the wrong answer matters just as much to me as knowing how to build with it. Restraint, in a field this hyped, is a credential.
wecalledit: the discipline of shipping alone
wecalledit.app is a product I designed, built and shipped entirely on my own, from the first rough idea all the way to a live, working web application. There was no team to hand pieces to. I framed the problem, did the research, designed the experience, built the front end and deployed it.
I built it AI-natively, using AI tools to accelerate the implementation and the iteration rather than to replace the judgement. That distinction is the whole point for me: the AI speeds up the drafting and the boilerplate, but I am reviewing and improving what it produces rather than accepting the first thing it gives me. Building it solo meant I owned the full loop (discover, build, ship, iterate) as a single person, which is a very different and very clarifying experience from owning one slice of a larger team’s work.
What they prove together
More than anything, these two products are evidence that I can take something from nothing to live on my own, and that I stay close to real delivery rather than drifting into pure strategy. They keep my hands in the actual making of things, which I think makes me a better designer for teams, not a worse one.
The lesson
Design strategy without delivery becomes a slide deck. Delivery without judgement becomes noise.
Both of these products exist because I want to keep both muscles working.
How I found out .
Each method, and why I chose it.
WhyDo I Need AI started as a question I kept asking in meetings. Writing it down as a tool forced the logic to be honest.
The people it had to work for.
Composite and anonymised. Needs and frictions kept, names removed.
Persona 1 of 3
The small business owner
Has been told 3 times this month that she needs AI.
- Needs
- A straight answer about whether a simpler tool would do.
- Friction
- Every advisor sells the thing they build.
Written on the
inside cover
- Most "AI problems" people brought were data problems or process problems wearing a costume.
- A tool that sometimes says no earns more trust than one that always says yes.
- Shipping alone is the fastest way to find out which parts of your process were actually someone else's.
What we made to make the decision.
Artefacts first, arguments second.
Most of this is under NDA, so each piece is described rather than shown.
-
The diagnostic
10 questions, 2 minutes, 4 verdicts. 2 of the 4 tell you that you do not need AI.
-
Build log
What the AI drafted, what I changed, and why. Kept for both products.
What changed, and what I keep.
- 2
- live products, designed, built and shipped by 1 person
- 10 questions
- in the Do I Need AI diagnostic, 2 minutes to complete
- 2 of 4
- verdicts that tell you that you do not need AI
- 0
- trackers, cookies or analytics on either product
- A free, public tool that pushes against the reflex to reach for AI, at a time when most companies investing in it see no measurable return.
- Proof that I stay close to real delivery. Design strategy from someone who still ships reads differently to a team.
- An AI-native build loop, with human review of every line that ships, that I now bring to client work.
- 01 Design strategy without delivery becomes a slide deck. Delivery without judgement becomes noise.
- 02 Restraint, in a field this hyped, is a credential.
- 03 Shipping alone shows you which parts of your process were really someone else's.
I would have put a way to reply on both products from day 1. Without analytics, a reader's email is the only feedback loop, and I added it late.