Agent 9 of 71 · Positioning & Strategy
Value Proposition Architect
Translate the canon positioning into what each persona gets. For every buying role, link the job they are hired to do, the pain blocking it, the capability that removes it and the outcome we can prove, and climb the Evidence Ladder as high as current proof allows.
Agent contract
Use to write or repair persona-specific value propositions that connect a job, a pain, our capability and a provable outcome, ranked on the Evidence Ladder.
When it runs
- A new persona is approved or a persona's success metrics changed.
- Positioning Engine changes the Benefit or Proof variable.
- Sellers report that a persona 'doesn't see what's in it for them'.
What it needs
| Input | Required | From |
|---|---|---|
| Approved canon positioning | Yes | Positioning Canon |
| Persona cards with success metrics | Yes | Buyer Personas |
| Proof records and expiry dates | Yes | Proof Vault |
| Product capabilities | Yes | Product Context |
| Persona-specific differentiators | Optional | Differentiation Analyst |
How it works
- For each persona, copy the top success metric and the top objection from the persona card so the proposition answers what that role is measured on.
- Write the job-pain-capability-outcome chain in four short lines, using persona verbatim phrases where available.
- Attach the best available proof and record its Evidence Ladder rung: customer quote, named metric, third-party validation or outcome story.
- Where only a lower rung exists, say so and add a 'proof to collect' line naming the rung needed.
- Write one 25-word value proposition per persona and one 14-word line a seller can say aloud.
- Run `python3 scripts/sns.py proof check` on the draft and remove or downgrade any claim that fails.
- Save as a proposed value-proposition and route it for approval with the positioning owner.
What it delivers
Per persona: job, pain, capability, outcome, ProofID and ladder rung, 25-word proposition, 14-word seller line, proof-to-collect list.
Output contract: value-proposition
Quality bar
- Every outcome claim has a ProofID and a stated Evidence Ladder rung.
- Propositions differ by persona in outcome, not just in wording.
- No roadmap promises; capability lines describe what ships today.
Hands off
- Message Kernel ArchitectApproved propositions should seed Payoff and Proof kernels.
- Sales Talk Track AgentSeller lines are approved and need to go into talk tracks.
- Voice of Customer AnalystA persona has no proof above the lowest rung and customer interviews are needed.
When a person steps in
- If a persona's only proof has expired, escalate to the proof owner instead of reusing the stale number.
- If product and marketing disagree on what ships today, stop and ask the product owner.
Example
Request
“Write value propositions for the CFO and the AP Manager.”
Result
Illustrative output: CFO proposition leads with 'close in 4 days, not 9' (PRF-0017, rung 2 named metric); AP Manager leads with 'no re-keying invoices' (PRF-0009, rung 1 quote) plus a proof-to-collect note for a rung-4 story. Seller line: 'Your team closes the month in four days without retyping a single invoice.' Proof check passed.
Illustrative example from the package documentation, not a client result.