Coordinate an internal AI-adoption research program

Why use this

1. An organization studying AI use may hear compelling anecdotes without knowing whether they represent different employee roles or only enthusiastic early adopters. Interview logistics and transcript analysis can become disconnected from the original research question.

2. This job could organize a bounded study and preserve the link between cohort choices, interview evidence and findings. Making participation and evidence gaps visible helps reviewers avoid treating uneven feedback as a company-wide conclusion.

Why an agent / simpler alternative

Useful judgment

Define meaningful role/context comparisons and synthesize practices while separating participant evidence from broad claims.

Simpler alternative

Use a small manual interview study when the research question and participant group are narrow.

Unsuitable when

The research would monitor employees secretly or turn adoption signals into individual performance judgments.

Proposed inputs

1. Name

   Research brief

   Owner

   Program owner

   Minimum scope

   Questions, voluntary participation scope and reporting audience

   Quality check

   Define acceptable use and confidentiality

2. Name

   Approved research records

   Owner

   Research team

   Minimum scope

   Consented cohort metadata and transcript excerpts

   Quality check

   Deidentify where possible and track sample limitations

Access and setup — proposed or unknown

Proposed minimum

1. Begin with fictional cohort descriptions and excerpts; no employee-source connection, interview invitation or transcript upload.

Compatibility unknowns

1. Current workspace access, cohort definitions, transcript tooling and runtime compatibility are unknown.

External actions requiring approval

1. Access grants, participant outreach, recording, transcript disclosure and shared findings require appropriate authorization.

Expected deliverable and review checks

Proposed artifact

Research plan and thematic findings table with evidence pointers, cohort coverage and unresolved questions.

Critical checks

1. Keep participation voluntary; distinguish frequency from significance; protect identities in small cohorts; avoid causal adoption claims.

Missing output evidence

This overlay adds no new output inspection, execution or validation; consult the referenced canonical source outcome limits.

Failure signals and proposed responses

1. Signal

   Only enthusiastic users are interviewed

   Why it matters

   The findings can overstate adoption

   Proposed response

   Report the sampling limitation and propose broader voluntary coverage.

2. Signal

   A recognizable quote reveals a participant

   Why it matters

   Deidentification fails despite removing names

   Proposed response

   Paraphrase or omit it within the approved reporting scope.

Effort and cost — unknown

Setup work

1. Define questions, consent process and reporting boundaries.

Ongoing review

1. Review themes, representation and confidentiality before sharing.

Unknowns

1. Participation, transcript quality, research effort and service costs are unknown.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

Invent three employee-role cohorts and four anonymized interview excerpts, including a nonuser and a uniquely identifiable anecdote.

Acceptance checks

1. Preserve nonuse evidence; flag disclosure risk; label sample limits; schedule no interviews.

Stop conditions

1. Stop if participation consent or the permitted audience is unclear.

Limits

1. Interview synthesis cannot establish organization-wide adoption, productivity gains or individual employee performance.