AI Builder

Build things with AI.
Ship them. Do it again.

You don't need a technical background — AI does the technical work, you direct it. By the end you have a tool at a real URL, an automation running without you, and a product tested by a real user. The constraint isn't skill. It's deciding to build.

~14 hours · 4 sprints · your pace

Start now — £499 → or 3 × £183 · includes AI Ready (worth £129)

Built on iO-Sphere's simulation-based training format — the same approach that has trained 700+ professionals at some of the UK's most demanding organisations.

You don't need to be a developer. But you will be building real things — a working automation, an internal tool with a UI, and a deployed product for someone else. Claude Code writes the code. You direct it, scope it, and ship it.

This is for you if

✓ You want to build tools and automations for your work and your team — and you're tired of waiting for a developer or paying for software that doesn't quite do what you need.
✓ You've seen what's possible with AI and you want to be the person who builds the solution, not just the person who requests it.
✓ You face different types of problems across your work. This programme gives you confidence across all of them — not just one type of build.

This is not for you if

— You want to build a product someone else will pay for and use repeatedly. That's AI for Founders (£799) — a depth programme, one product, real users, real revenue.
— You're looking for a no-code drag-and-drop tool. This uses a real stack (Claude Code, GitHub, Netlify, Supabase) — real software that can handle real problems.

The programme

Four sprints. Four live deployments.

Three types of build — automation, internal tool, deployment for someone else. Your pace, no deadline.

Before Sprint 1 unlocks: you install GitHub, Claude Code, Netlify, Supabase, and Resend. Takes about 30 minutes. Sprint 1 unlocks only when all five are confirmed.

Sprint 1

The Fundamental Pattern

You bring the use case — what you want to collect, and who from. The pattern is prescribed: form, database, confirmation email. By end of Sprint 1 you have a live deployed URL and the fundamental pattern burned in: trigger → process → output → confirmation.

→ Build: Live deployed form tool — URL working, Supabase row on submit, confirmation email sending.

Sprint 2

Your Automation

Your own automation. Scope it first — one trigger, one specific process, observable output (Claude evaluates, one resubmission if needed). Build it. Then deliberately break it: identify the failure modes and add one guard. Stress-tested before it's called done.

→ Build: Live automation for a real task in your work — scoped, built, stress-tested. Evidence: screenshot + what it does and what happens when it fails.

Sprint 3

Your Internal Tool

Build something with a UI — something another person can use. Scope-check first. Build it. Then a real-person test: someone who isn't you goes through it while you observe without guiding. Submit what you saw.

→ Build: Live internal tool with a UI — scope-checked, built, tested with a real person. Evidence: URL + screenshot + observation notes.

Sprint 4 — Capstone

Build for Someone Else

Identify a real person with a real problem — not yourself. Scope it, build it, deploy it. Three-point test: (1) does it work? (2) did the person actually use it? (3) can you articulate what to build next — what worked, what they showed you, what changes. 2 of 3 to pass.

→ Build: Live deployment for a named person + two paragraphs on what you'd build next. This is your main portfolio piece.

What you leave with

✓Sprint 1 — live deployed form tool built for your use case (the pattern established)

✓Sprint 2 — live automation for a real recurring task, stress-tested

✓Sprint 3 — live internal tool with a UI, tested with a real person

✓Sprint 4 — live deployment built for someone else, 2/3 criteria passed ← the main portfolio piece

Inside the programme

This is what it looks like
when you're inside a sprint

A real brief. Real tools. A real thing to build. Not a tutorial — a job to do.

autonomy-platform.com · AI Builder · Sprint 2

📋 Sprint brief

You run a weekly status update manually. Slack message, copy from Notion, reformatted, sent. 45 minutes every Friday. Automate it.

Sprint 2 · Your work context · Role: Builder

Identify a recurring task in your work that has a consistent trigger, a consistent process, and a consistent output. Scope it in one sentence: 'When [trigger], [process], and produce [output].' Claude evaluates the scope — if it passes, you build. If not, one resubmission.

The test isn't 'does it run once'. It's 'does it run reliably when I'm not watching'. Build it, then break it — deliberately.

Your deliverable

Live automation + screenshot evidence + 2-sentence description: what it does and what happens when the trigger fails.

Automation Scope

Automation Scope — Sprint 2

When a new row is added to our Notion project tracker with status 'Ready for Review', extract the project name, owner, and deadline, format as a Slack message, and post to #projects-ready.

Claude evaluation: Single trigger ✓ | Specific process ✓ | Observable output ✓

Building...

Using:

Claude Code

GitHub

📋

Real brief

A real problem, a scope to pass, a thing to ship

🛠️

Real stack

Claude Code, GitHub, Netlify, Supabase

✦

AI assessment

Evaluated against pass criteria — not vibes

🚀

Live URL

Every sprint ends with something deployed

The track record

Built on a format that has already worked — at scale.

Autonomy is built on iO-Sphere's simulation-based training format — the same approach that has trained 700+ professionals at some of the UK's most demanding organisations.

BAE Systems PwC L'Oréal Bupa Uber NHS

“

"The simulation format is unlike anything I'd seen before — it actually feels like work, not study."

Michael

iO-Sphere alumni

How it compares

Real code. Real deployments.
Without becoming a developer.

	No-code tools	AI Builder	Hiring a developer
Format	Drag and drop — no real code	Real code directed with Claude Code	Hand it off and wait
You build automations	✓ some tools	✓ Sprint 2 — scoped and stress-tested	✓ if you can spec it
You build tools with a UI	limited	✓ Sprint 3 — tested with a real person	✓ expensive
You build for someone else	—	✓ Sprint 4 capstone — three-point test	✓ very expensive
What you walk away with	A workflow	4 live deployments	One thing (if it shipped)
Price	Free–£50/mo	£499 · includes AI Ready	£500–£5,000+ per project

Pricing

One price. Everything included.

£499

or 3 × £183

Lifetime access · no subscription

Included at no extra charge

AI Ready foundation

Normally £129 · yours free

✓Full AI Builder programme — 4 sprints, 10 nodes
✓AI Ready foundation included (£129 value)
✓4 live deployments — all real URLs
✓Claude Code + GitHub + Netlify + Supabase stack
✓AI-assessed feedback on every submission
✓Your pace — no deadline, progress saves automatically

Includes AI Ready foundation — normally £129, yours at no extra charge

Start AI Builder →

Secure payment by Stripe · Access within minutes · Questions? [email protected] — you'll hear back from a real person

Questions

The things people usually ask

No. Claude Code writes the code. You scope the problem, direct the build, and review the output. By Sprint 4 you'll understand what's happening — even if you can't write it from memory.

Founders builds one product someone else uses and pays for — that's a depth programme. Builder builds three different types of things (automation, internal tool, deployment for someone else) — that's a breadth programme. Different goals, different people. If you want to ship a startup idea, Founders. If you want to be the person who builds solutions across any type of problem, Builder.

GitHub (free), Claude Code (subscription required — Claude Pro or API access), Netlify (free tier), Supabase (free tier), Resend (free tier). The only paid requirement is Claude Code access (~£18/mo).

4 sprints, your pace. No deadline, no cohort. You advance when your build passes the assessment — which means it works, it's been tested with a real person (Sprint 3), and you pass the three-point test at Sprint 4 (2 of 3 criteria: evidence it works, evidence someone used it, articulation of what to build next).

4 live deployments, all at real URLs: a form tool built to your brief, a personal automation (stress-tested), an internal tool (tested with a real person), and a capstone build for someone else. The Sprint 4 capstone is the main portfolio piece.

Autonomy's foundation programme (normally £129, included free). Covers how to direct AI, validate outputs, and know when AI is wrong — directly relevant to directing Claude Code throughout the programme.

Build things with AI. Ship them. Do it again.