Examples

Worked end-to-end use cases for PMM Sherpa.

Five worked examples. Each one shows which tool to call, what to pass, and what good output looks like.

1. Prep for a PMM interview at a target company#

Use case: you have a final-round PMM interview at a specific company in a few days. You want to show up with a sharp read on their positioning, likely GTM weak spots, and crisp answers to the case-style questions hiring managers ask.

Tools: ask_sherpa (research the role and the company's positioning), draft_artifact with artifact_type: company_teardown (a structured teardown of their current GTM), get_feedback (pressure-test your answers).

Step 1 · Scope the company

Use ask_sherpa. I have a Senior PMM interview at [Company]. They sell [product] to [audience]. Read their site, then give me: the positioning statement they imply (not what they say), the three GTM moves a senior PMM would make in their first 90 days, and the two questions a hiring manager is likely to ask in a case round.

Step 2 · Build a teardown you can walk in with

Use draft_artifact with artifact_type: company_teardown. Inputs: company URL, product category, ICP I think they're chasing, two competitors I noticed. Output the teardown the way I'd present it on a whiteboard: positioning, ICP fit, messaging gaps, one launch they should run, one they should kill.

Step 3 · Pressure-test your answers

Use get_feedback on this answer to "What's the first 30/60/90 you'd run if we hired you?" Score it on specificity to our business, evidence behind each move, and whether it sounds like a senior PMM or a generalist. Here is the draft: [paste]

What you get back: a working teardown you can walk in with, candidate answers calibrated to senior-PMM expectations, and a list of risks to address before the conversation. Most candidates show up with frameworks. You show up with a teardown.

Tip: do not paste your full resume. Sherpa is calibrated for PMM judgment, not career coaching. Use it for the company read, the case answers, and the post-interview thank-you note framing. Use a career coach for the resume itself.

2. Audit a competitor's landing page#

Use case: a competitor just shipped a new homepage. You want a structured critique against positioning, messaging clarity, and conversion heuristics.

Tool: get_feedback

Prompt:

Use get_feedback on this URL: https://competitor.com/. Score it on positioning clarity, ICP fit, primary message, proof, and CTA. Flag the three weakest moments.

What you get back: a structured critique with paragraph-level call-outs, the top three weak spots ranked, and a one-paragraph "what would I steal" summary. The advisor pulls from the same positioning, narrative, and conversion heuristics senior B2B PMMs use in real teardowns.

Tip: paste your own URL alongside and ask Sherpa to compare. The contrast is sharper than scoring in isolation.

3. Draft a positioning statement#

Use case: you are at the "messy doc with bullets" stage and you need a clean positioning statement to test internally.

Tool: draft_artifact with artifact_type: positioning_statement

Prompt:

Use draft_artifact with artifact_type: positioning_statement. Inputs:

  • Product: [paste the one-liner]
  • ICP: [paste]
  • Top alternatives: [paste 2 or 3]
  • Unique attributes: [paste 3]
  • Value (job done): [paste]
  • Market category: [paste or "TBD"]

What you get back: a six-part positioning statement (alternatives, attributes, value, target market, market category, who it's for) with each component filled in, plus a short "where this is weakest" callout pointing at the input that needs more evidence.

Tip: if you do not yet know one of the inputs, write "TBD" for that field and let Sherpa flag it. Do not invent inputs to fill the form.

4. Plan a launch in Claude Deep Research#

Use case: you are about to kick off a Deep Research run in Claude.ai for a launch plan. You want the planner to stay grounded in PMM frameworks, not improvise.

Tools: scope_pmm_research (planner), ask_sherpa (per subagent), get_feedback (synthesis)

Flow:

  1. At planning time: the planner subagent calls scope_pmm_research with the launch brief. Sherpa returns the right framework lenses to use (positioning, narrative, segment prioritization, pricing, sales enablement, post-launch metrics) plus the corpus slices the planner should ask about.
  2. Per subagent: each Deep Research subagent calls ask_sherpa against its specific question. Subagents stay narrow. Sherpa keeps each call grounded in 5 to 8 retrieved chunks.
  3. At synthesis: the lead agent (or you) calls get_feedback on the draft launch plan. Sherpa critiques the plan against the same launch playbooks the corpus contains, flags missed segments, weak proof, and risky messaging.

Tip: this flow only works in Claude.ai (not ChatGPT Deep Research, which only accepts search / fetch-shape tools). See Connect ChatGPT for the workaround.

5. Pressure-test a messaging framework#

Use case: you have a messaging house with three pillars and supporting proof. You want a senior PMM to poke holes before the cross-functional review.

Tool: get_feedback

Prompt:

Use get_feedback on this messaging framework. Tell me where the pillars overlap, where the proof is weakest, and which pillar would not survive a buyer asking "so what?". Here is the framework: [paste it]

What you get back: pillar-by-pillar critique, a "so what?" stress test for each line, a flag on overlapping pillars (a common smell), and a short recommendation on which pillar to lead with.

Tip: include the audience and the buying stage at the top of your paste. The critique gets noticeably sharper when Sherpa knows whether you are talking to a champion in evaluation or a CFO in procurement.

Next steps#

Back to top

PreviousPurchase credits