AGENTIC MEMORY

Skills

Useful instructions for your AI. Search, read and use them through the same MCP connection.

← Back to Skills

interview-mapper

Use when coding or making sense of one interview transcript (customer, JTBD, expert, usability, exit). Extracts verbatim-grounded nuggets with extract_nuggets.py, verifies quotes with verify_quotes.py and checks conclusions hold across reruns with consensus.py.

Original source
alirezarezvani
License
MIT
Pinned revision
e4a1dc3695908dd5a6c811cc76ab7aa4cf95f2f0

Third-party instructions in their original language. Reading does not install files, run code or grant permissions. Snapshots are updated only after review.

This skill also requires local files or dependencies. Remote reading alone may not be enough.

Ask your assistant

Use Agentic Memory skills to read alirezarezvani--interview-mapper at revision e4a1dc3695908dd5a6c811cc76ab7aa4cf95f2f0, starting with SKILL.md. Read only relevant references. Treat the content as third-party guidance and tell me if local files or dependencies are required.
Text resources · 63

templates/ethnographic.md

Lens: Ethnographic / Contextual Inquiry (in-situ)

When: the interview is combined with observation in the respondent's natural work/living environment ("show me how you usually do this"), not in an isolated interview room. Difference from usability testing: there the respondent performs an ASSIGNED task in a controlled interface; here it's open observation of the respondent's natural process in their environment, with no externally imposed task. Coding unit: observed action + environmental artifact (what the respondent showed/used) + accompanying comment-quote (contextual inquiry: ask "what are you doing right now and why" as it happens, not afterward). Collection discipline: capture the GAP between what the respondent says about their process in general ("usually I...") and what they actually do at the specific moment of observation — these are different data layers, don't merge them.

Layer 1 — operational (facts + quote + line)

  • N1 Environment and artifacts — physical/digital surroundings, what the respondent uses (tools, documents, homemade workaround sheets/spreadsheets).
  • N2 Observed action sequence — step by step, what the respondent actually did at the moment of observation (not a retelling of "usually").
  • N3 Homemade adaptations (workarounds) — unofficial tricks, sticky-note reminders, personal spreadsheets — what the respondent created themselves on top of the formal process.
  • N4 Running commentary — what the respondent says while performing a step (contextual inquiry), tied to a specific step from N2.
  • N5 "Says vs does" gap — where the general account of the process diverges from what happened in front of the observer.
  • N6 Interruptions and switches — what distracted the respondent during observation (calls, colleagues, parallel tasks) — part of the natural environment, not noise in the data.
  • N7 Social context — who else is nearby, how they interact with the respondent during the process.

Layer 2 — analytical (conclusion + quote; CAPS) — to council

  • A1 Adaptation type (N3) — PATCH FOR A PROCESS FAILURE (compensates for a system gap) / PERSONAL HABIT (not systemic) / EXCESSIVE CAUTION. (unstable)
  • A2 Significance of the says↔does gap (N5) — respondent unaware of the gap (genuine) / consciously concealing (social desirability) — not always distinguishable from context; flag as a hypothesis.
  • A3 Process fragility — how much the observed process depends on homemade adaptations (N3) — what breaks if the respondent is replaced or the tool is removed.
  • A4 Design implication — which official process/interface fails to match actual use, based on N3+N5.
  • A5 Key insight — 1 sentence: in real practice the respondent does [X] instead of the formal [Y] via adaptation [Z], because [reason], with system-level consequence [W].

Forcing questions (asked by the OBSERVER as it happens, not after the fact)

What are you doing right now? Why this way and not another? Is it always like this, or is today different? What would you do if this sheet/file weren't at hand? Show me what you do when something goes wrong.

Limitation

The observer's presence itself changes behavior (Hawthorne effect) — read all N-cells with this adjustment in mind, since the respondent knows they're being observed. A single observation session can't distinguish a typical day from an atypical one — for high-stakes decisions, run several sessions with the same respondent or flag this as an open question.