Product

The discipline this file enforces: a hard line between what exists and what is planned. Founders blur them without meaning to — the demo becomes "our platform," the roadmap becomes "our capabilities" — and then get caught out by one specific question in a diligence call.

---

What exists today

[TBD]

What a real customer could use this week. Be concrete and unflattering.

The vocabulary, used correctly

Getting these wrong in a pitch is a small thing that reads as a large one:

What is still a slide

[TBD]

Everything you describe in the future tense. Naming it here protects you: a founder who volunteers "this part does not exist yet" is trusted on everything else they say.

Product lines / SKUs

Cost and price feed 07-money. If you cannot fill the cost column, that is the most urgent gap in your stack.

Which of these actually makes money? [TBD]

Multi-product early companies almost always have one line quietly funding the rest. Naming it changes how you allocate your time — and it is the honest answer to "what is the business?"

Roadmap

Milestones with dates and a definition of done, not a wish list.

Explicitly not doing: [TBD]

The "not doing" list is the more useful half. Everything you have consciously deferred belongs here, so that a good idea in month four gets weighed against a decision rather than slipped in.

How it is built

"Do you own it?" is not rhetorical. If an agency built your app, an intern designed your logo, or a friend wrote the code, the intellectual property may not be yours by default. Assignment has to be written down. This surfaces during diligence, at the worst possible moment, and it is cheap to fix now.

Vendor dependencies

A verbal maintenance promise is not a maintenance agreement. If a vendor has committed to supporting what they built, get it in writing with a response time, an uptime expectation, source access, and a handover clause. "He said he'd fix anything for two years" is a conversation, not a contract.

Intellectual property

Decision on patents: [TBD — filing, not filing, or deciding later — and the reasoning]

Record the reasoning, not just the decision. "We are not filing because the design changes every quarter and the money is better spent on customers, and we accept the copying risk" is a defensible position. Having no position is not.

Trade secrets — what are you deliberately not disclosing, and how is that actually protected? A formula everyone in the workshop knows and nobody has signed an agreement about is not a trade secret.

Safety, ethics and claims

Fill this in if your product touches health, finance, safety, education, children, or personal data. Skip it if genuinely not applicable — and be sure before you skip.

Human in the loop

If any part of your product makes or recommends a decision automatically — especially using AI — answer this explicitly:

Treat any automated system as capable of being confidently wrong, regardless of how well it has been trained or guarded. If an output could harm someone — a health suggestion, a financial recommendation, a safety instruction — there must be a named point where a qualified person validates it before it reaches the user. Buyers, regulators and serious investors all ask this question, and "the model is very accurate" is not an answer to it.

Sustainability and impact claims

If you make environmental, ethical or social claims, they must be precise, current, and survivable under scrutiny.

Greenwashing is when a term is used to make something sound sustainable that mostly is not — "vegan leather" that is predominantly plastic being the canonical example. The route through is not to overclaim or to go quiet, but to be transparent about the stage the product is actually at. A proprietary formulation or process does not have to be disclosed. What the product is, what is still conventional or synthetic, and what the actual goal is, do. Customers forgive a company that is honestly part of the way there. They do not forgive one that claimed to have arrived and got checked.

Feedback loop