Skip to content

Decisions, not dashboards.

Kevin T. Holland

Product leader and data product builder, hands-on executive

Question

Which decision will this change, and who makes it?

Evidence

Selected projects

Selected products and open-source work on decisions, data quality, evaluation, and human review.

  • InstaDoc.health

    Co-founded / healthcare

    Symptom assessment and care routing, with clinician review when care is requested.

    Open product

  • Data Team OS

    Open source

    Skills and a file-based workspace for data-product decisions, evidence, quality checks, and human review.

    Open source

  • CensusChat

    Solo build / open source

    A natural-language interface over US Census data, with a validation gate between generated queries and the database.

    Open source

  • Hitch

    Reference build / open source

    Governed agent writeback: reads from live records, writes back only through an approval step the server enforces.

    Open source

Healthcare product work

  • $15M platform ARR

    Led product work across NeuroBlu’s $15M ARR platform business, including $3M in standalone subscriptions.

  • 49M patient lives

    Scaled one clinical dataset from 0.5M to 49M lives, with 59M+ clinical notes.

  • 3 clinical AI models

    Helped ship them to pharma customers, with an evaluation suite the team used.

Decision

Ask which decision it will change, before anyone builds it.

Limits

  • Employer work appears as outcomes, without links, because it is not mine to demo.

  • Hitch is a reference build, not a product in use.

Owner

Kevin T. Holland

I lead healthcare data and AI products from hard problem to operating system: products people trust, buy, and use to make a decision.

For questions about the writing or public projects, email Kevin.

Contact Kevin View selected projects

What the work taught me

All writing
  1. Data Products

    What Makes a Product Worth Building?

  2. Essay

    From BI to AI: How Decision Support Is Changing

  3. Data Products

    AI Made Building Cheap. It Made Judgment Expensive.

One dashboard, counted

Is Your Data Team a Dashboard Factory? DPOS, Part 0
  1. Part 0Is Your Data Team a Dashboard Factory?
  2. Part 1DPOS: The Operating System for Data Products in the AI Era
  3. Part 2The DPOS Kernel: Principles for the AI Era
  4. Part 3Role Operating Systems
  5. Part 4The Integration Layer (soon)
  6. Part 5Implementation (soon)

Observed

4

Opens, ever

One dashboard. Eleven days to build. Four opens in the access logs.

Excluded1

Someone in finance clicked by accident. Not a use.

Screened

3

Real opens

The requester opened it twice. Her deputy opened it once.

Limit

Measured / inferred

Decision impact is not measured in this account.

Decided

Decision

Ask which decision it will change, before anyone builds it.

4 opens, access logs, DPOS Part 0.

Turn evidence into action

Read the DPOS series

The hard part comes after the chart loads. I build the structure that carries evidence into an owned decision and a repeatable workflow.

  1. ClarifyName the decision and the owner.
  2. GroundTest the claim against real evidence.
  3. BuildShip the smallest system that can change the call.
  4. OperateMeasure the outcome and close the loop.

Read the operating record