Intake · Forcing-question triage

An intake agent that triages raw ideas in 10–15 minutes — one hard question at a time.

One forcing question at a time, light research between question two and three, explicit gaps marked before handoff. No more two-hour strategy sessions that end in "maybe."

Book a 30-min scoping call →See how it works

10–15

Minutes per idea, end to end

Forcing questions in startup mode

Verification dimensions before handoff

The Challenge

Ideas don't die from bad research. They die because triage is slow, vague, and inconsistent.

A founder captures an idea in Telegram. A week later, somebody starts a two-hour Zoom to "expand this." The conversation wanders. Nobody asks the one question that would have killed the idea in three minutes. The idea gets filed as "promising" and quietly forgotten.

The failure mode is always the same. Asking "tell me more" generates more content, not more signal. Asking "what's the market?" triggers a research rabbit hole. Ideas get promoted without interrogation or buried without a hearing. Either way, you're optimizing for the wrong thing.

The root cause isn't the team's IQ. It's that there's no forced structure between "raw idea" and "handoff to planner" — no one-question-at-a-time discipline, no research gate, no scoring, no explicit gap-marking. Triage becomes a vibes exercise, and vibes don't scale past five ideas a week.

This agent puts structure in the middle: six forcing questions asked one at a time, 90 seconds of research between Q2 and Q3, and a 16-point score before anything goes to planner — with unknowns marked as unknowns, not papered over.

How the agent handles it

Capture. Detect mode. Ask one hard question. Research the gap. Score. Handoff.

Every idea gets classified before it gets interrogated.

The agent summarizes the idea back to confirm, then detects mode — startup (revenue-focused) or builder (learning-focused). Different modes get different question sets. There's no neutral triage; the questions you'd ask about a side project aren't the ones you'd ask about a launch.

One forcing question at a time. Vague answers get pushed back.

Not "tell me more". Real questions: "What would make you open Slack and say holy shit?", "Who must this disappoint?". Answers that dodge get challenged before the next question fires. The whole point is to surface hidden assumptions, not collect more wordcount.

Ninety seconds of research sits between question two and three.

Enough to check competitors, market signals, and existing solutions — not a two-week market study. Findings get shared with the user, and unknowns get marked as unknowns: "I don't know X. Do you want me to dig, or is this for planner?" No false confidence on handoff.

Nothing gets promoted without a 16-point score.

The verify-idea skill scores on 16 dimensions — market, team, technical, timeline, wedge, kill criteria. Threshold: 10/16 to promote. Below that, it's back to the questions or into the archive. Planner receives a brief, the score, and the explicit gap list — ready to decide, not ready to re-validate.

What you get

Three things change once triage is structured.

10–15min

End-to-end time per idea

Down from the typical 1–2 hour ad-hoc strategy session. Ideas don't linger waiting for a free afternoon.

0unreviewed

Ideas dying silently in the inbox

A Sunday-morning digest surfaces every pending idea with a recommended action. No six-month-old ideas rediscovered during cleanup.

16points

Scored before every handoff

Planner receives a brief with the score, the forcing-question answers, and explicit unknowns. Nothing to re-validate.

Numbers observed in Brilworks' internal reference deployment. Actual figures on your stack will depend on idea volume, team composition, and how strict you want the promotion threshold.

Is this right for you?

Honest fit criteria. We'd rather say no than oversell.

✓Strong fit if

You capture 10+ ideas per week across Slack, Telegram, voice notes, and email
Validation today takes 1–2 hours per idea and feels like a vibes exercise
Ideas regularly die in the inbox or get promoted without real research
Your team runs both startup-mode and builder-mode projects and needs different questions for each

✗Not a fit if

You generate fewer than five ideas per week — the machinery isn't earning its keep
Your triage is already fast and structured today
You want the agent to make yes/no calls for you (this surfaces judgment, not replaces it)
Your team isn't on Telegram or any async tool that supports one-at-a-time Q&A

Book a 30-minute scoping call.

We'll walk through your current intake, map it against the six-question pattern, and tell you honestly whether structured triage is the right next step.

Book a scoping call →See other case studies