One decision.
Make the challenge useful.
A practical working session. Start with the choice, put alternatives on the table, then name the next test.
Start the working session →Founder decisions that consume runway, equity and focus.
- Pivot or persist
- First executive hires
- Funding offers
- Choosing and scoping advisers
US founders from pre-seed to Series A, and the people who back them.
- First-time pre-seed founders
- Runway-constrained seed CEOs
- Experienced but isolated founders
- Series A CEOs and portfolio support teams
Three routes to challenge, compared side by side.
- Tool, Build and Expert, each with its limits stated
- Readable samples before any email
- Evidence before endorsement
- You retain the decision
The useful question isn’t “Should we pivot?” It’s “Which evidence would justify either route?”
A seed-stage B2B workflow software company is weighing a pivot. Read the founder’s position, then reveal the countercase. Neither gets the last word.
“A broader launch will solve weak demand.”
- Five paid pilots suggest people will pay for the workflow at all.
- A broader launch increases the number of signals in the next quarter.
- Narrowing early could cap the story for the next round.
The opposing argument gets equal space. Revealing it doesn’t replace Case A.
“The strongest signal may belong to one narrow customer segment.”
- All three renewals come from one segment. Breadth may be diluting the evidence, not adding to it.
- A broad launch spends runway on segments that haven’t renewed.
- A narrow wedge can still be a credible fundraising story if retention is visibly stronger.
Observed · supplied
- 18 customer interviews completed
- Five paid pilots
- Three renewals, all in one segment
- Seven months’ runway
Interpreted · inference
- Renewals may reflect segment fit rather than product breadth
- Weak demand elsewhere may be a positioning problem, not a product problem
Missing · go and find out
- Why the two non-renewing pilots stopped
- Sales-cycle length by segment
- Whether the renewing segment is large enough for the next milestones
Nine items. One decision. You keep the call.
Work through the agenda in order, or open any item. Each one reveals its working material underneath, with nothing to sign up for.
-
Start with the decision, not the advisory category.
Write the commitment in one sentence, the date it becomes hard to reverse, and who holds final authority. Nothing here is sent anywhere. It stays in this browser tab.
Illustrative · syntheticIllustrative seed company · 12 people · pivot decision · six–nine months’ runway
- Product
- B2B workflow software
- Runway
- Seven months
- Evidence
- 18 interviews, five paid pilots, three renewals concentrated in one customer segment
- Authority
- Founder / CEO
You retain the decision. We never issue a verdict.
-
A seed-stage pivot is not a general strategy question.
Constraints change which kind of challenge is useful. They don’t decide the answer for you. Choose a decision type, then add only the context you’re comfortable sharing. Your choices update the rest of this agenda without moving the page.
Choose a decision type to see which constraints matter most.
-
The useful question is not “Should we pivot?” It is “Which evidence would justify either route?”
Synthetic example · not a real companySeed-stage B2B workflow software. 12 staff, seven months’ runway. 18 interviews, five paid pilots and three renewals, all concentrated in one customer segment.
Both cases, the evidence split and the reversible test are worked through in the sample challenge.Open the sample challenge- 01 · CalibrationConfirm stage, runway and the advice already received.
- 02 · Competing casesArgue each case at full strength. The founder speaks last.
- 03 · Evidence prioritiesSort observed, inferred and missing. Pick the cheapest test.
- 04 · Next stepsOwner, review date and reconsideration trigger.
Illustrative 40-minute session outline · not a promised service -
Do not buy an adviser for work a tool can do. Do not ask a tool to supply independence.
Each route is useful on its own and they can be combined. “Build” means an internal decision process, which doesn’t have to be software.
Three routes, same fields Route Suitable decision Founder effort Output Limitation 01Tool First pass on reversible decisions using verifiable, non-sensitive inputs 20–40 minutes Competing options, pre-mortem, organised context Cannot establish independence, resolve interpersonal dynamics or validate its own assumptions 02Build Repeated decisions where the team has context, time and psychological safety A few hours to set up, then per decision Dissent owner, decision brief, review triggers Shared incentives and authority differences can suppress disagreement 03Expert Irreversible commitments, conflicting interests or a weak internal countercase Brief, interviews, bounded sessions Bounded independent review with explicit access rights Costs money; independence and judgement need diligence -
Bring a sharper pre-read. Leave the confidential names out.
Every attachment has a readable sample with no email needed. We never ask for company names, cap tables or investor identities, and there’s no document upload. Delivery requests in this demo create a preview only. No email is sent.
-
Relevant disagreement beats an impressive biography.
Directory records are built from reviewed evidence, not paid placement. Evidence statuses describe the material we reviewed. They don’t measure competence.
Public directoryNo independently reviewed provider records supplied.
Until records have been reviewed, with sources and review dates, this directory lists nobody. Use these checks when you assess any adviser:
- Ask for a permissioned sample in which they disagreed with a founder, and what changed as a result.
- Request written disclosure of equity, referral fees and implementation work they could sell you.
- Confirm who sees the output, and whether founder material goes into AI tools.
- Compare the minimum cash commitment, not just the headline fee.
- Agree the exit terms before the first session.
Evidence statuses- Evidenced
Reviewed material directly supports the criterion. - Partly evidenced
Some support; gaps are listed. - Not evidenced in reviewed material
Absence of evidence, not evidence of incompetence.
Reboot, Business Talent Group and AdvisoryCloud come up in our research as adjacent suppliers and discovery channels. They are not reviewed or verified matches.
Fictional comparison · practice only Sample Engagement model Disclosed fee (fictional) Conflict disclosure AI and confidentiality Disagreement evidence No fictional samples match these filters.
-
Settle the boundaries before the first candid conversation.
Mark each term as settled, open or not applicable. Open each clause to see an illustrative wording. These are prompts for discussion, not legal drafting. Sensitive conversations are not automatically privileged.
0 of 10 boundaries settled
Engagement boundary checklist Term Question to settle Illustrative clause Status Markdown file, generated in your browser. -
Hard Question editorial desk. Edition 1, revised 27 October 2026. Status: internal editorial review; not yet externally reviewed. Any figures are labelled as examples or estimates.
-
Record the next test, not a verdict.
Saved in this browser tab only. Closing the tab clears it. Download the record or add a calendar reminder to keep it.
Match the challenge to the decision.
Don’t buy an adviser for work a tool can do. Don’t ask a tool to supply independence.
- 01
Tool Instant AI support
Competing options and a pre-mortem from verifiable, non-sensitive inputs. Limit: can’t establish independence or validate its own assumptions.
- 02
Build Internal process
A named dissent owner, a decision brief and review triggers. Not necessarily software. Limit: shared incentives can suppress disagreement.
- 03
Expert Independent human
A bounded review for irreversible commitments or conflicting interests. Limit: costs money; independence and judgement need diligence.
Each marker is an illustrative decision. Select one to see why it sits where it does.
Bring a sharper pre-read. Leave the confidential names out.
Four working templates to prepare the difficult conversation. We never ask for company names, cap tables or investor identities.
- 01Decision Diagnostic3 pp
- 02Tool vs Build vs Expert2 pp
- 03Adviser Selection4 pp
- 04Reconsideration Triggers1 p
Read the full evidence.
121 research pages across 12 sections, with sources, caveats and practical tools.
Browse all research →Relevant disagreement beats an impressive biography.
No independently reviewed provider records supplied.
The directory lists nobody until records have been reviewed, with sources and review dates. In the meantime, these are the checks to run on any adviser:
- Ask for a permissioned sample where they disagreed with a founder, and what changed.
- Get written disclosure of equity, referral fees and implementation work they could sell you.
- Confirm who sees the output, and whether founder material goes into AI tools.
- Compare the minimum cash commitment, not the headline fee.
- Agree exit terms before the first session.
Founder retains authority
We compare routes and surface evidence. We don’t issue verdicts or recommend one adviser.
Evidence before endorsement
No directory record without reviewed sources, a review date and a stated evidence status. No paid ranking.
No sale of founder submissions
In this demo your inputs stay in your browser. Suppliers never get access to founder submissions.
What will you test before you commit?
Start with one decision. There’s no account, nothing leaves your browser, and you keep the call.