Prove the need, size the market, price the value — then decide. Every figure traceable, with AI as your coworker.
Two lines in. The AI researches all three questions and tells you where the idea leans.
Electric delivery van, 3.5 tonnes → UK parcel and courier fleets
Built from the head start.
The market looks favorable. Winning it looks mixed.
1Is the need real?
favorable1Is the need real?
favorable
Rests on 8 drafted needs · published evidence on 5 ·
2Is the market worth it?
favorable2Is the market worth it?
favorable
Rests on 2 sizing routes · growth rates · the largest players · 5 segments ·
3Can we win it?
mixed3Can we win it?
mixed
Rests on the competition search · substitutes · 4 alternatives ·
Read from everything the head start found. Each rating is yours to make, in the factor's own tool.
The platform’s own worked example: an electric van into UK courier fleets. Figures illustrative.
Simplified views. Every tool has more behind it.
1 · The need
Every need rated for how much it matters and how well it's already met. The ones worth building on stand apart from the table stakes.
✦AI drafts the candidate needs and finds the references. You rate them, and the tool ranks.
2 · The market
One pool of spend, cut as many ways as you'd like. The tool guides you to measurable or arguable cuts. You select the targets and the sequencing.
✦AI works out how the market can be sized and proposes the figures. The engine does the arithmetic.
3 · The edge
What they need on the left, what you'll claim on the right. A line is a fit. A need with no line, and a claim with no line, are both findings.
✦AI reviews the fit against your rating and says what it would need to see. The rating stays yours.
All three on the platform’s own worked example — an electric van into UK courier fleets. Figures illustrative.
Built to be usable and defendable — around the ways AI goes wrong. Every press fills a prompt written for one situation with your own market and figures.
Hallucination
On your own
It invents figures, fluently.
Here
It never touches the arithmetic.
Poor sources
On your own
It cites a report mill like a statistics office.
Here
Every source graded. Report mills named, not dropped.
Unclear rationale
On your own
Easy to use. Hard to use well.
Here
How to ask is worked out. Every answer keeps its why.
Inconsistency
On your own
Three people, three answers.
Here
Three people, one prompt.
Each is built for its job from the pieces below — the head start runs ten in one press.
✦Every AI button in the platform, one square each
and we keep adding more →
Run by the head start · 10Ready when you want themAbout 67 in the platform today
Kept current — each job runs on the model that suits it best, and moves when a new one does better.
The needs you ranked are what the value proposition is matched against. The segments you targeted are what the growth plan is built from. Every tool reads what the others recorded.
Most major business initiatives — mergers and acquisitions, capital investments, market entries — fail to ever pay off.
Whichever way in, two kinds of reader
One part of one tool per film. Press play.
One market, three cuts, and what you walk away holding.
£182.5M is the real TAM the Market Model reports; the cuts and shares are illustrative.
The other method on the same card — four named terms, and a band summed from the two that stated one.
Real World Bank series for the UK, 2015–2024; the structural shifts and closing figure are illustrative.
What a node is, what a route is, and where the material ends up.
The chain's own chemicals example. Tonnages and routes are illustrative; the bodies named are real, their figures are not.
Illustrative — invented figures; several tools are still being built.
31 fleet managers interviewed. Not range — predictable charging downtime.
£180M a year of courier vans replaced — sized top-down and bottom-up, agreeing within 4%.
⚠ Looks done. Isn’t. Reads Worth it, and the work behind one of the answers under it has not started.
£4,100/yr cheaper to run, £9,000 more to buy than the diesel van — a 2.2-year payback.
Three routes costed. Leasing turns the £9,000 sticker into a monthly line — the objection that decides it. Year-two volume of 1,400 units fits the line's 2,000 of spare capacity. No capex.
A decision, not a rating — Launch · Enter small · Reshape · Wait · Don’t enter.
Illustrative — invented figures; several tools are still being built.
31 interviews across 14 teams. The top unmet need belongs to the engineer, not the designer.
$738M counted by designer seats, $1.16BN counted by teams — 44% apart, and the gap is the finding.
Two entrenched rivals, and a day per feature of spec-writing that appears on nobody's budget.
Free for anyone who only looks or comments. Paid per editor — the pricing structure IS the route in. Rendering pro-grade vectors in a browser is the bet. Funded and signed off — a cost, not a blocker.
A decision, not a rating — Launch · Enter small · Reshape · Wait · Don’t enter.
Illustrative — invented figures; several tools are still being built.
Seven needs, four OEMs. The top one was on no specification — it was in the warranty returns.
$1.1BN a year, sized two ways 7% apart — and eleven programmes reach most of it.
No qualified equivalent at this thickness — and the warranty saving lands in a budget the buyer does not hold.
①②③ said go. The build said not at this volume. The plan goes back — the market does not. The chemistry is qualified. The line that makes it at volume is not funded.
A decision, not a rating — Launch · Enter small · Reshape · Wait · Don’t enter.
Example factor: Unmet Needs
NoneWeakModerateClear needAcute & validated
Ten ratings like this answer the three questions.
See why some evidence for this factor counts more than others.
The worked examples are open to anyone. The tools are on trial while in development.