pattern-and-analogy-connector

Connects loose observations into a meaningful opportunity by identifying analogies across industries and situations using the Capability Pattern Mapping method: superficially different cases are abstr

SKILL.md

GitHub

Pattern & Analogy Connector

Purpose

Connects loose observations into a meaningful opportunity by identifying analogies across industries and situations. The core method is Capability Pattern Mapping: instead of asking "has something similar been done in our industry" — which produces an endless, quickly-outdated list of cases — superficially different sets of cases are abstracted into one named, industry-agnostic capability pattern. The named pattern then becomes a diagnostic question that can be applied to any new industry or situation without first having to find a known example from that specific industry.

This is different from a case library. A case library answers "has an example of this been done" — it goes stale quickly and tempts people to copy the surface-level solution as-is. A pattern answers "what kind of structural situation is this" — it stays useful for years, and forces you to think through your own context instead of searching for a ready-made answer.

Anchored in research

  • Tang, J., Kacmar, K. M., & Busenitz, L. (2012), "Entrepreneurial Alertness in the Pursuit of New Opportunities," Journal of Business Venturing, 27(1), 77–94 — association and connection, one of the paper's three dimensions of entrepreneurial alertness (see also ../market-and-signal-scanning/SKILL.md for scanning-and-search and ../opportunity-evaluation-and-judgment/SKILL.md for evaluation-and-judgment).
  • Baron, R. A. (2006), "Opportunity Recognition as Pattern Recognition: How Entrepreneurs 'Connect the Dots' to Identify New Business Opportunities," Academy of Management Perspectives, 20(1), 104–119 — the argument that opportunity recognition draws on the same cognitive mechanism as pattern recognition generally: connections between previously unrelated events or trends are noticed because a person's existing knowledge structures make the connection visible in the first place.
  • A Capability Pattern Mapping method and worked example described by the user (an invoice/customs/CV document case, see point 3 below) — the owner's own abstraction technique, not an academic source.

A concrete, fully worked-out application of this to AI capabilities exists at ../../../ai-strategy-and-governance/references/ai-capability-pattern-library.md and its navigation skill ../../../ai-strategy-and-governance/skills/ai-capability-pattern-matching/SKILL.md. This skill here is the GENERAL method; the AI-strategy pack's pattern library is one concrete implementation of it, applied to one domain (AI solutions).

Method

  1. Collect 3+ superficially different observations/cases in which you suspect a similar underlying structure. They can come from different industries, different clients, or observations made at different times. Don't start by searching for "a comparable industry" — start by looking for structural similarity beneath the surface.
  2. Ask four abstraction questions of each case, questions that deliberately bypass industry-specific vocabulary:
    • What is the input? (e.g. "free-form text/PDF document," not "loan application")
    • What is the actor/role doing the work today? (e.g. "a highly paid specialist reads through the document," not "a loan processor")
    • What is the core cognitive function? (e.g. "searching for an anomaly/gap in a large volume of unstructured content," not "application review")
    • What is the outcome/decision the work supports? (e.g. "approve / reject / escalate," not "credit decision")
  3. Write a single sentence that describes every case in the set the same way, using only the answers from step 2. This sentence IS the pattern's name and definition. Example, from the user's own material: three superficially completely different cases — document review of loan applications in financial services, tariff-code verification of customs declarations in logistics, and CV screening in HR — all abstract to the same pattern: "Validation and anomaly detection in unstructured documents" (input: free-form document; actor: a specialist; cognitive core: searching for an anomaly in a large volume of text; outcome: approve/reject/escalate decision).
  4. Turn the pattern into a diagnostic question that can be asked of any new client without already knowing an example from their specific industry. Example: the pattern "Validation and anomaly detection in unstructured documents" produces the question: "Where in your process does a highly paid specialist have to search for anomalies in free-form text or a PDF document?" — the same question can be asked of a construction, insurance, or public-sector client without having first seen a known example from their specific industry.
  5. Test the pattern's coverage and sharpness before using it:
    • Coverage — can you find at least three genuinely different (across industry/context) examples for the pattern? If you can only find one, it isn't a pattern yet, just an isolated case — don't generalize too early.
    • Sharpness — is the pattern precise enough to distinguish it from its neighboring patterns? A pattern that's too broad ("AI helps with decision-making") doesn't point the diagnostic question anywhere useful; one that's too narrow ("PDF loan application review in financial services") doesn't generalize across industries.
  6. Use the diagnostic question in a new context (a client meeting, a workshop, your own observation) and record the answer as a structured hypothesis: the pattern's name, why this situation matches the pattern, and how this situation differs from the known examples.
  7. Validate the hypothesis with stakeholders or against your own experience-based checklist before carrying it forward (see ../opportunity-evaluation-and-judgment/SKILL.md).

What this skill does NOT do

  • Doesn't make the final decision for you — it produces a structured draft to support a human decision.
  • Doesn't confirm figures, market data, or competitor data from memory — it uses the inputs you provide, or clearly flags an assumption ([assumption — verify]).
  • Doesn't guarantee that a found analogy actually holds — it produces a hypothesis that still needs validation.
  • Doesn't maintain a ready-made, comprehensive pattern library within this general-purpose skill — that would make it unwieldy and go stale quickly. Industry- or solution-type-specific pattern libraries (e.g. the AI capability patterns) live in their own packs and point back to this method, not the other way around.
  • Doesn't replace in-depth industry research — the pattern is meant to speed up hypothesis generation, not to substitute for industry expertise.

Refinement notes

The method and the invoice/customs/CV example are the owner's own worked technique. Areas to keep expanding as more of it is put into practice:

  • your own validated patterns from opportunity-recognition work (not just from the AI context)
  • concrete examples of where abstraction went too far (a pattern turned out too broad to be useful) or too narrow
  • a template for assembling diagnostic questions (into ../../references/)

Once this section is filled in and validated in practice, update the maturity field in skills_index.json to draft, validated, or canonical (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

  • Next in this pack: ../opportunity-evaluation-and-judgment/SKILL.md — Structurally assesses the viability of the identified opportunity before resources are committed.
  • Concrete application to AI solutions: ../../../ai-strategy-and-governance/references/ai-capability-pattern-library.md (the pattern library) and ../../../ai-strategy-and-governance/skills/ai-capability-pattern-matching/SKILL.md (how the library is used in client work).
  • A ready-made skill chain for this situation: see ../../../playbooks/
  • This pack's shared guardrails: ../../CLAUDE.md

References

  • ../../references/ — the pack's shared background material
  • ../../CLAUDE.md — this pack's shared guardrails