Rebuild the Positioning: Choosing the Right AI Stance for a Regulated Buyer
Objective: Given Coinbase's institutional AI-powered transaction monitoring feature and a compliance-focused buyer, choose the correct positioning stance and rewrite the statement to fit it.
You're a product marketer at Coinbase, positioning a new AI-powered transaction monitoring feature to institutional compliance teams at banks and exchanges evaluating Coinbase for custody. The current draft leads with model architecture, and it's not landing with buyers.
Diagnose which of the three AI positioning stances fits a compliance buyer, then rebuild the statement around the Jobs-to-Be-Done frame with specific, methodology-backed proof.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, easy to share with compliance and legal for review before it ships
The process
2 steps
Step 01 of 02
The lesson names three stances: lead with outcome (hide the AI), lead with magic (show the AI), or lead with trust (explain the AI), with trust reserved for buyers facing reputational or regulatory risk from the AI's decision.
The buyer is a bank's compliance team deciding whether to trust Coinbase's AI to flag suspicious institutional transactions. Which of the three stances fits, and why do the other two fail here?
Procedure
- List the buyer's actual fear (approving a system that misses a suspicious transaction, or can't explain a flagged one to a regulator)
- Rule out lead-with-outcome (compliance can't just trust a black-box result with regulatory exposure on the line)
- Rule out lead-with-magic (model novelty doesn't reduce the buyer's regulatory risk, it can even increase suspicion)
- Confirm lead-with-trust: the positioning must explain what the model was trained on, how it flags, and where a human reviews before any action is taken
Stance justification: Lead with trust. The buyer's job isn't 'get faster flags', it's 'defend this decision to a regulator.' Outcome-only positioning ('40% faster flagging') doesn't address that fear. Magic-first positioning actively raises suspicion in a compliance context. Trust-first, explaining training data, flag logic, and human review, is the only stance that removes the actual blocker to adoption.Healthy
The stance decision is made explicitly, in writing, before any copy is drafted, and it's justified by the buyer's specific fear, not a general preference for one stance.
Unhealthy
The team picks a stance by instinct or copies whatever a competitor did, without naming the buyer's actual fear first.
What this means
Stance follows buyer risk, not company preference. A compliance buyer's fear (explaining a decision to a regulator) determines the stance more than what the AI is technically capable of.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Sales says the AI pitch isn't landing with compliance teams | Check whether the deck leads with model architecture instead of the buyer's regulatory fear | 30 min |
Step 02 of 02
The lesson argues a feature is valuable for the job it does, not the technology behind it, and that trust-first positioning for regulated buyers needs specific, methodology-backed proof, not vague claims.
The current draft opens with 'Built on a proprietary transformer model fine-tuned on blockchain transaction data.' What's the job-first, trust-backed rewrite?
Procedure
- Name the job first: helping a compliance team clear or escalate a flagged transaction faster and defensibly
- State the outcome with a specific, methodology-backed number, not a superlative
- Add one sentence on human oversight (a compliance analyst reviews every high-risk flag before action, the model never acts alone)
- Cut every remaining mention of model architecture from the customer-facing statement, keep it in the technical appendix instead
Rewrite: 'Coinbase's transaction monitoring flags suspicious institutional activity with 91% precision, validated on 500,000 transactions reviewed by our compliance team in Q1 2026, so your analysts spend less time chasing false positives and more time on real risk. Every high-risk flag is reviewed by a human compliance analyst before any account action is taken. Full model documentation is available for your audit team on request.'
Healthy
The rewrite leads with the job and a specific number, then explains human oversight, with model architecture detail moved to a technical appendix for buyers who ask.
Unhealthy
The rewrite still opens with 'proprietary transformer model' or a vague 'highly accurate' claim, which fails the same skepticism test the original draft did.
What this means
For a regulated buyer, the job is defensibility, so the proof point (sample size, review process) matters more than the phrase 'AI-powered' ever could.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Compliance buyers ask for more detail after reading the positioning | Point them to the technical appendix instead of front-loading architecture into the main pitch | 5 min |
Final deliverable
A one-paragraph stance justification plus a rewritten positioning statement that leads with the job, a specific methodology-backed proof point, and explicit human oversight language.
See a reference example
Wise AI Sanctions Screening — Rebuilt Positioning (excerpt) 'Wise's sanctions screening flags high-risk cross-border transfers with 94% precision, validated on 1.2M transfers reviewed in Q4 2025, so your compliance team spends less time on false positives and more time on genuine risk. Every flag is reviewed by a licensed compliance analyst before a transfer is held. Full model documentation available on request.'
Success criteria
You're done when you can:
- Explicitly justifies the trust-first stance based on the buyer's regulatory fear, not a generic preference
- Rewritten statement leads with the job, not the model architecture
- Includes one specific, methodology-backed proof point and one human-oversight sentence