demo-framing-and-expectation-setting

Frames a demo/prototype/PoC for the customer before presenting it with the right term (PoC vs. Pilot vs. MVP) and the right promise — what this demo PROVES, what it does NOT prove, and what happens ne

SKILL.md

GitHub

Demo Framing & Expectation Setting

Purpose

Prevent the most common, expensive-to-fix mistake in demo delivery: the wrong frame before the first slide. If a customer walks out of a demo believing a production-ready solution is three weeks away, and reality is six months of development work, the problem isn't the quality of the demo — the problem is that no one framed the demo correctly before it started. This skill produces that frame: what term correctly describes what's being shown today, what this demo proves and what it does NOT prove, and what concretely happens next if the demo succeeds.

This is a different question than ../../../ai-strategy-and-governance/skills/ai-use-case-feasibility-and-poc-scoping/SKILL.md, which defines a PoC's TECHNICAL boundaries (what data, what success criteria, what scope limits). This skill answers the customer communication question: how is that same PoC framed in conversation so that no false expectations are created. Use both together — technical scoping first, then this communication frame.

Anchored in research

  • The PoC / Pilot / MVP distinction (synthesis of multiple 2026 sources, see References): three different stages that answer three different types of uncertainty (technical feasibility / operational fit / product development direction).
  • "Pilot purgatory" research (McKinsey, BCG, IDC, MIT syntheses, see References): a large share of enterprise AI pilots never reach production, and the bottleneck is typically operational (management commitment, workflow redesign, scale-up investment) — not the technical success or failure of the demo/PoC stage.

Method

  1. Name precisely what's being shown today, with the right term, and don't use the terms as synonyms (see the pack's ../../CLAUDE.md):
    • PoC, if the question is "does this work technically with representative data" — no real users yet, no production load.
    • Pilot, if the question is "does this work with real people and real operational conditions" — technical feasibility has already been demonstrated.
    • MVP, if the question is "what should be built next, based on real user feedback" — a product-development stance, not a proof stage. If you're not sure which, ask yourself: "which ONE uncertainty does this answer today?" If there's more than one answer, you're probably merging stages — separate them.
  2. Write a one-sentence "proves/doesn't prove" pair before the demo:
    • "This demo proves that ___ [a precise, narrow claim, e.g. 'the model extracted the supplier business ID from 20/20 test invoices']."
    • "This demo does NOT prove that ___ [anything outside the test scope, e.g. 'it works with all invoice formats', 'it's secure for production use', 'it scales to 10,000 invoices a month']." Present both to the customer before the demo, not only if someone asks.
  3. Tie the frame back to the technical scoping done in ../../../ai-strategy-and-governance/skills/ai-use-case-feasibility-and-poc-scoping/SKILL.md if that's already been done — use the same success criteria, don't invent new ones at demo time.
  4. State up front what concretely happens NEXT if the demo succeeds — who decides, on what timeline, what resources the pilot/production stage would require. This is a direct countermeasure against "pilot purgatory" risk: if no one has agreed in advance what a successful demo leads to, it leads to nothing, regardless of the demo's quality.
  5. State out loud what the demo does NOT yet resolve organizationally — workflow change, user training, management commitment, budget for full scale. Technical success in the demo doesn't mean these are solved.
  6. Choose the tone of the frame based on the audience: for a technical audience you can emphasize accuracy figures and limitations directly; for an executive audience, route the frame through ../../../change-and-communication/skills/executive-narrative-and-storyline/SKILL.md before the demo, so the technical "proves/doesn't prove" pair translates into business language.
  7. Document the frame in writing before the demo (one paragraph is enough) and share it with participants — this reduces the risk that the post-demo memory drifts (people forget most of a demo's content quickly, but a written frame stays).

What this skill does NOT do

Refinement notes

Areas to keep deepening with real practice:

  • your own examples of how you've framed a demo successfully (or unsuccessfully) with a specific customer
  • your own standard-phrase/slide template for the "proves/doesn't prove" pair (into ../../references/)
  • rules of thumb for when a customer typically over-interprets a demo — which signals predict this

This is an internal working note, not a claim about the skill's current usability. Track depth privately via the maturity field in skills_index.json (see ../../../meta/maturity_levels.md). Don't add new fields to the frontmatter — name and description are the only ones allowed (see ../../../meta/frontmatter_schema.md).

Continue from here

References

  • The PoC vs. Pilot vs. MVP distinction — synthesis of multiple 2026 sources on the staging of enterprise AI projects
  • "Pilot purgatory" research — McKinsey/BCG/IDC/MIT syntheses on why a large share of AI pilots never reach production
  • ../../references/ — the pack's shared background material
  • ../../CLAUDE.md — the pack's shared guardrails