Market Intelligence
Preamble
Market Intelligence turns external research into constraints. It scans startups, products, repos, papers, platform shifts, pricing pages, complaints, and category winners only when the result can change a decision.
Research Can Become A Shelf
The Factory can collect more context than it can use. A competitor teardown can be interesting. A repo can be impressive. A platform shift can feel urgent. A pricing page can look like strategy. None of it matters until it changes how the Factory chooses a source, validates a wedge, writes a PRD, routes a runtime, rejects a tactic, or allocates attention. Market Intelligence exists to stop research from becoming a shelf.
Constraints Beat Links
A useful scan should produce a constraint update. That update can change a decision, hypothesis, source priority, domain chip, PRD/spec gate, roadmap choice, rejected path, or runtime requirement. It can also reinforce a current constraint when the evidence says the existing path still holds. The output should preserve what was observed, why it matters, what it changes, what remains uncertain, and which surface should absorb the lesson.
Competitors As Primitives
Factory research should extract primitives that change the system. Polsia contributes company instances, heartbeats, specialist agents, dashboards, and the portfolio question. Cofounder contributes memory-centric company context, task state, and agent workflows. Agent coding platforms contribute isolated worktrees, review agents, CI gates, sandboxes, and auditable traces.
The test is whether those primitives change Factory design. A competitor note that does not alter a harness, branch contract, runtime object, or rejected path is archive material.
The First Real Proof
Market Intelligence lands when a scan changes or reinforces a live Factory constraint. The branch succeeds when research enters a decision surface. It fails when the archive gets richer and the roadmap stays unchanged for reasons nobody can inspect.