Research Note
Responsible Internal Experiment Research Note
Internal entrepreneurship is a governed experiment, not permission to bypass the company. A responsible pilot defines one job, one decision owner, one data boundary, a re
In this article
Responsible Internal Experiment Research Note
Editorial conclusion
Internal entrepreneurship is a governed experiment, not permission to bypass the company. A responsible pilot defines one job, one decision owner, one data boundary, a reversible test, baseline evidence, review criteria, and a stop, iterate, or scale decision.
The method in the public guide is a Venture Step synthesis. External sources support risk management, prototype discipline, and learning methods. They do not establish one universal corporate process.
Evidence map
| Source | Supported use | Boundary |
|---|---|---|
| NIST AI RMF Core | Govern, map, measure, and manage as continuous risk-management functions | Voluntary cross-sector framework; AI RMF 1.0 is under revision |
| NIST Generative AI Profile | Cross-sector risks and actions for generative AI design, use, evaluation, and governance | A profile, not organization-specific authorization |
| GOV.UK alpha guidance | Test the riskiest assumption without prototyping the entire service | Government digital-service context |
| GOV.UK prototype guidance | Prototypes help test before commitment and should not be copied directly into production | Digital-service design context |
| Test and Learn annex | Small-scale prototype tests can expose hidden assumptions with proportionate qualitative and quantitative evidence | UK government evaluation guidance, not an employer policy |
| ABA Formal Opinion 512 | Concrete example of competence, confidentiality, verification, supervision, and responsibility remaining with professionals | Legal ethics guidance rather than a general business rule |
Pilot design boundary
The guide must not invite readers to place client, employee, confidential, regulated, personal, licensed, or production data into an unapproved tool.
Approval should cover the problem, data, system, participants, reviewer, evaluation, retention, procurement, records, security, privacy, legal, and release path as applicable.
Measurement model
Speed, quality, rework, control performance, user experience, and total cost should be evaluated separately.
A faster first draft can create more downstream correction, monitoring, exception handling, and review. The unit of comparison is the completed and accepted workflow, not the model response.
Worked-example boundary
Until a cleared real case exists, the public guide should use an explicitly hypothetical document-intake pilot. It should not invent a customer, employer, result, or performance number.
Sources
Follow the evidence.
- NIST AI RMF Measure guidanceairc.nist.gov
- Test and Learn annexassets.publishing.service.gov.uk
- Commercial Observer on Henrycommercialobserver.com
- Experimental evidence on professional writingdoi.org
- OECD AI adoption by SMEsdoi.org
- Eurostat enterprise AI reportec.europa.eu
- UK Companies House recordfind-and-update.company-information.service.gov.uk
- Navigating the Jagged Technological Frontierpubsonline.informs.org
- Avison Young company overviewretail.avisonyoung.com
- ALTA/NSPS Land Title Survey Standardsalta.org
- ABA Formal Opinion 512americanbar.org
- About Avison Youngavisonyoung.co.uk
- US Census 2022 SUSB tablescensus.gov
- NAIOP I.CON West 2024 attendeescredaglobal.org
- GOV.UK alpha guidancegov.uk
- GOV.UK prototyping guidancegov.uk
- Henryhenry.ai
- How Compass Commercial Scales with Henryhenry.ai
- Luke Tatman's LinkedIn profilelinkedin.com
- From Static to Strategic: AI's Role in Next-Generation Industrial Real Estatenaiop.org
- nber.org: w31161nber.org
- NIST: Artificial Intelligence Risk Management Framework, Generative Artificial Intelligence Profilenist.gov
- OECD 2026 adoption releaseoecd.org
- RICS Real estate agency and brokeragerics.org
- RICS Responsible use of AI standardrics.org
- RICS Property Agency and Management Principlesrics.org
- RICS AI in real estate valuationrics.org
- Henry company profileycombinator.com
- California DRE corporation recordwww2.dre.ca.gov
- California DRE salesperson listwww2.dre.ca.gov