AutoMarkAI

Methodology · Rubric AIS-1.1

The AutoMark Idea Score

One number from 0 to 100 that answers the question an investor asks first: is this worth building? It is built from seven pillars, each drawn from a framework founders and investors already use, scored by two independent AI analysts against fixed anchors.

How the score is made

  1. 1

    Two analysts, independently

    Anthropic and OpenAI models each assess the idea on their own, without seeing the other's answer.

  2. 2

    Seven pillars, fixed anchors

    Each pillar is scored 0–10 against the published descriptions below, with one line of evidence.

  3. 3

    Weighted and scaled in code

    Pillar scores are averaged across both analysts, weighted, and placed on the published AutoMark scale. The headline score is arithmetic, not a model's opinion.

  4. 4

    Agreement reported

    We publish how closely the two analysts agreed. A high-agreement score is more reliable than a split one.

Weighted sum = Σ (pillar weight × pillar score ÷ 10), pillar score = mean of both analysts. Idea Score = AutoMark scale (weighted sum).

The AutoMark scale

At the idea stage, even excellent concepts rarely earn more than 7 out of 10 on moat or distribution, because neither exists yet. The weighted sum of the strongest ideas therefore tops out in the 60s. The scale maps that real spread onto 0–100, so the strongest ideas reach the high 80s and weak ones stay low. Between the points below the score is interpolated linearly.

0 → 020 → 2232 → 4045 → 6055 → 7462 → 8568 → 9280 → 97100 → 100

The seven pillars

Problem & demand

Weight 20%

Jobs-to-be-Done · pain frequency × intensity

Is there a frequent, painful job that people already pay or hack around to get done?

0–2
nice-to-have, rare pain, nobody pays today
3–4
real but occasional pain, weak workarounds
5–6
frequent pain, people hack around it
7–8
frequent, costly pain with money already spent on alternatives
9–10
hair-on-fire, budget exists, buyers actively searching

Market size

Weight 15%

TAM / SAM / SOM, bottom-up

Is the reachable market large enough to build a venture-scale business?

0–2
SAM < $10M
3–4
SAM $10-100M
5–6
SAM $100M-1B
7–8
SAM $1-10B
9–10
SAM > $10B and growing

Competition & moat

Weight 15%

Porter's Five Forces · Helmer's 7 Powers

Can it win against incumbents and defend the position once it does?

0–2
commodity against funded incumbents, no power available
3–4
crowded, differentiation is features only
5–6
a credible wedge, one power reachable later
7–8
clear wedge plus a specific power the plan reaches within 2-3 years (proprietary data loop, workflow switching costs, network effects)
9–10
structural power incumbents cannot copy

Business model

Weight 15%

Unit economics · LTV:CAC ≥ 3, CAC payback < 12 months

Does each customer earn back what it costs to acquire and serve them?

0–2
LTV:CAC < 1 or negative gross margin
3–4
LTV:CAC 1-2, payback > 24 months
5–6
LTV:CAC 2-3, payback 12-24 months
7–8
LTV:CAC 3-5, payback 6-12 months, gross margin > 60%
9–10
LTV:CAC > 5, payback < 6 months, gross margin > 75%

Distribution

Weight 15%

Traction channels (Bullseye) · Thiel's power-law of distribution

Is there one repeatable, affordable channel to the first 1,000 customers?

0–2
no identifiable channel
3–4
only paid ads at uncertain CAC
5–6
one plausible owned or community channel
7–8
a specific, proven channel for this exact buyer (named communities, partners, embedded or B2B2C distribution, marketplaces) or built-in virality
9–10
viral loop or embedded distribution (marketplace liquidity, product-led sharing)

Timing

Weight 10%

“Why now?” · Idealab timing factor

Has something changed recently that makes this possible or necessary now?

0–2
too early or too late
3–4
no clear trigger
5–6
helpful tailwind
7–8
a recent shift (technology cost, regulation, behaviour) makes it newly possible
9–10
a window that is opening now and will close

Execution feasibility

Weight 10%

Build complexity · time to MVP · regulatory exposure

Can a small team ship a credible v1 quickly without a regulatory wall?

0–2
regulatory wall or deep-tech R&D
3–4
heavy integrations, licences or hardware
5–6
standard SaaS or app, 3-6 months to MVP
7–8
MVP in 4-8 weeks with modern AI-assisted tooling
9–10
MVP in days

Grades

A85–100Venture-grade
B+75–84Strong
B65–74Promising
C55–64Needs work
D40–54High risk
E0–39Not yet viable

The score judges an idea's potential if a capable team executes it well, so an early idea is not penalised for traction it cannot have yet. It is strict about real weaknesses: small markets, crowded categories with no wedge, weak willingness to pay and no route to customers. A typical idea lands in C or D; only large, painful, well-timed markets with a clear wedge reach A. Every report includes the moves that would raise the score.

Certificates and verification

Every report issues a certificate with a unique number and a verification code: a cryptographic signature of the certificate's contents. Anyone can open the verification page printed on the certificate and check the score, grade and code against our records. Certificates are kept for five years. Rankings, where shown, compare the idea with every idea assessed before it.

Limits

The score assesses the idea as described, using public knowledge available to the models. It does not see the founder, the team, private traction or customer data, and it cannot predict results. It is not investment, legal or financial advice. Treat it as a structured, independent second opinion and a checklist of what investors will probe.

Score your idea