Know what the market may demand before the first workshop

1. Discover where demand becomes visible

The scan adapts to the company's commercial shape. It can follow named customers, channels, competitors, and suppliers; discover association, regulator, and trade-news sources for a segment; or combine both. It resolves public company, jobs, investor, news, and relevant sitemap pages where customer, regulatory, economic, technology, talent, supplier, and capital pressure may appear.

For B2C and B2B2C companies, aggregate consumer-review, complaint, and platform-policy sources replace named-customer assumptions. Individual reviewers are never treated as accounts or quoted as evidence.

2. Capture pressure with provenance

Every signal must retain its headline, exact source URL, observation date, affected segment or account, direction, and implied demand. Signals expire when they become stale. A source-less assertion cannot enter the demand view.

3. Review before anything counts

Discovered sources, resolved account websites, competitor candidates and claims, and extracted force signals enter review queues. Accepted, unexpired evidence can contribute downstream; rejected or expired evidence cannot. The public teaser auto-accepts only inside an isolated throwaway scope so it can produce a preview without changing a customer's real workspace.

4. Form likely-priority hypotheses

The system groups accepted signals into likely internal priorities and preserves the source chain behind each one. These remain hypotheses. They are compared with the known company model and classified as:

• Confirmed — the likely priority lands on a known offering, workflow, pain, asset, or constraint.
• Blind spot — public pressure suggests a priority, but the current company evidence does not show where it lands.
• Zombie — an internally important area has little current public-demand support and deserves review rather than automatic removal.

5. Turn uncertainty into discovery

Missing internal support becomes a focused interview agenda: who owns this pressure, what process handles it today, what evidence would confirm it, and what would refute it? Public evidence starts the conversation; employees, systems, and operating records establish what is true.

6. Influence ranking within strict bounds

An accepted priority can apply a visible, capped ranking adjustment to an AI opportunity already grounded in internal evidence. It never creates an ordinary candidate by itself. In exceptional cases, the workflow can open a bounded adjacent-offer review, demand-gap review, workflow-pressure review, or revalidation review.

7. Build account plays without inventing intent

For a reviewed named customer or channel, an account plan can connect accepted pressure, current and whitespace offerings, overlapping competitor claims, and proposed defend, deepen, expand, or land plays. Every play remains a reviewable commercial hypothesis tied to sourced evidence; it never asserts private customer intent.

8. Carry only validated opportunities into build gates

Selected opportunities still have to clear feasibility, safety, evaluation, pilot, and production-handoff gates. External urgency cannot override missing data, unclear ownership, unsafe actions, or weak evaluation coverage.

The result is a traceable chain from dated public evidence → likely business priority → internal validation → grounded opportunity → gated agent package.