Back to the episode map

Research Note

Workplace Workflow Mapping Record

An AI product is evaluated against a work system. Without a current-state map, a team can automate the visible drafting step while leaving intake, evidence gathering, exc

Aug 4, 20262 min readBy Dalton Anderson

Workplace Workflow Mapping Record

Why the map precedes the product

An AI product is evaluated against a work system. Without a current-state map, a team can automate the visible drafting step while leaving intake, evidence gathering, exception handling, review, filing, correction, and accountability untouched.

NIST's AI RMF Core asks organizations to define the specific task, deployment context, expected benefits and costs, human oversight, third-party components, and affected people. Those questions begin with the work.

Observation boundary

The official procedure is not always the actual process. The map should follow a representative case from trigger to accepted record and include the people who perform and receive the work. It should distinguish observed practice, required policy, workaround, exception, and proposed future state.

Worker review is evidence, not ceremony. A manager's diagram can omit judgment, informal coordination, accessibility needs, exception work, and correction steps that make the process function.

Minimum map

The map should identify the trigger, inputs, data owners, source authority, decisions, transformations, tools, handoffs, queues, exceptions, controls, output, acceptance standard, system of record, rework, incident path, and accountable owner.

For each step, record elapsed time and active effort separately. Waiting for approval is different from work time. AI may shorten generation while leaving queue time unchanged.

Intervention decision

The map should expose the narrow constraint. The problem may be missing source authority, duplicate records, unclear ownership, bad permissions, scarce reviewer capacity, inconsistent intake, or a policy conflict. AI is not the default treatment.

The final record should state whether the constraint is suitable for AI assistance, ordinary automation, workflow repair, training, staffing, a policy decision, or no change.

Sources

Follow the evidence.

  1. NIST AI RMF Measure guidanceairc.nist.gov
  2. ftc.gov: ai companies uphold your privacy confidentiality commitmentsftc.gov
  3. youtu.be: 0cC1Ez33ryIyoutu.be
  4. daltonanderson.ghost.io: ai in the workplace a practical guide to get starteddaltonanderson.ghost.io
  5. NIST AI Risk Management Frameworknist.gov
  6. NIST AI Resource Centerairc.nist.gov
  7. eeoc.gov: prohibited employment policiespracticeseeoc.gov
  8. eeoc.gov: us eeoc and us department justice warn against disability discriminationeeoc.gov
  9. nber.org: w31161nber.org
  10. open.spotify.com: 7LIXDoSM2gG97vFGftskQsopen.spotify.com
  11. NIST Privacy Frameworknist.gov
  12. nber.org: w33795nber.org
  13. eeoc.gov: strategic enforcement plan fiscal years 2024 2028eeoc.gov
  14. NIST Generative AI Profilenvlpubs.nist.gov
  15. ftc.gov: start security guide businessftc.gov
  16. dol.gov: ten 07 25dol.gov
  17. hbs.edu: dell acqua et al 2026 navigating the jagged technological frontier 5c589c8c fbb5 458f b285 c944746cd717hbs.edu
  18. cisa.gov: cisa and uk ncsc unveil joint guidelines secure ai system developmentcisa.gov
Workplace Workflow Mapping Record