Standardize the learning cycle, not the venture
Cohorts often contain teams working in very different industries, company-size segments, regions and buying situations. Giving everyone the same persona or checklist can make discovery look consistent while making the learning less relevant.
Standardize a smaller set of things instead:
- begin with a decision the team needs to make
- state one assumption that could change that decision
- define the customer context that can challenge it
- separate simulated signals from existing evidence
- identify what remains uncertain
- commit to a real-world next test
The questions can be shared. The customer context and conclusions should remain venture-specific.
Give every team the same discovery loop
A simple cohort workflow can use six steps.
Step · Team task
- 1. Decision: What decision is the venture preparing to make?
- 2. Critical assumption: What belief would materially change that decision if it were wrong?
- 3. Customer context: Which role, industry, region and company-size context is relevant to that assumption?
- 4. Simulated exploration: What questions, objections, alternatives or missing perspectives emerge?
- 5. Evidence gap: What is still unknown, and what cannot be resolved through simulation?
- 6. Real-world test: Which customers, observations, experiments or other evidence should the team pursue next?
This creates a common learning rhythm without turning discovery into a form-filling exercise.
Make coach reviews about evidence, not confidence
Founders can become more persuasive without becoming more correct. A useful coach review therefore should not ask, “How confident are you now?”
Ask instead:
What did you believe before the exploration? What did the simulated conversation challenge or add? Which part is still only an assumption? What evidence would change your mind? What will you do before the next review?
The coach can also challenge the customer setup. A venture may be talking to a user when the uncertainty is really about an economic buyer, or using a broad “enterprise customer” when the relevant context is much narrower.
That is coaching the quality of the learning process, not grading the venture.
Example: different ventures, one learning rhythm
Illustrative example, not program or customer results.
Imagine an accelerator cohort with ventures working on very different B2B problems.
One team is developing maintenance-planning software for industrial companies. Its critical assumption is that maintenance managers experience enough coordination friction to change their current process.
Another team is exploring a workforce-planning service. Its uncertainty is whether operations, HR or finance owns the outcome strongly enough to sponsor change.
Each team configures a different modeled customer and runs a focused Virtual Customer conversation. They do not compare simulated enthusiasm between ventures.
At the next coach review, both teams use the same format: assumption, customer context, what the simulation surfaced, what remains uncertain, and the next real customer test.
The process is comparable. The evidence and decisions are not.
Keep simulated learning out of venture scoring
Virtual Customer can help teams articulate a problem, expose assumptions, compare customer perspectives and prepare real validation. It cannot determine which startup is strongest, predict which venture will succeed or provide a defensible investment ranking.
A fluent simulated conversation can make a weak assumption sound plausible. A difficult simulated conversation can expose a useful research question. Neither result should become a cohort leaderboard.
Investment, funding and continuation decisions need evidence appropriate to those decisions, including actual customer behavior, commercial progress, technical evidence and program judgment where relevant.
Use one learning record across the cohort
A coach does not need every team’s entire conversation. A short team-compiled record can make reviews more consistent:
Decision → critical assumption → customer context → simulated signal → evidence already held → unresolved question → real-world test → next decision
Over successive reviews, this creates a visible learning trail. Coaches can see whether a team is resolving uncertainty or simply rewriting the story around the same unsupported belief.
It also preserves an important distinction: Virtual Customer can help generate and challenge hypotheses, while the team remains responsible for gathering evidence and interpreting what real customers show.
Start with one critical assumption per team
Ask every venture to bring one decision and one assumption that could change it. Have each team explore that assumption with an appropriate modeled customer, record what the conversation challenged, and identify one real-world test that could produce stronger evidence before the next coach review.