Hey Kaleb & Co. · Implementation Pack

The Research Repository Decision Kit

Buy or build it with AI—those aren't two different answers. They're different invoices for the same purchase. This kit starts with the questions underneath, so you know what you're paying for before you sign.

Take the readiness assessment

If you haven't already, read the companion essay: Buy or Build with AI: You Pay Either Way.

A note on who this is for...

This kit addresses the research or ops leader making the recommendation. But most repository decisions also need engineering capacity you don't control, budget approval from above, legal and technology assessments, and possibly a procurement process. If you're building the case, not signing the check, the kit helps you structure the argument — but bring the people with the authority into the conversation before you commit to a path.

Interactive assessment

Nine questions. One path. Your checklist.

Three pass-or-fail gates, then six readiness dimensions. Answer honestly—the quiz scores your responses and recommends a path with a concrete checklist at the end.

Red Pen · A compliance caution

Your consent form likely doesn't cover AI...

Consent for storage is not consent for AI processing.

If your participant consent says "we'll store findings for future reference," that doesn't cover feeding data into an AI system. Since August 2026, GDPR and the EU AI Act stack—two regulatory frameworks, two sets of fines—have been applied simultaneously to any research repository with AI capabilities.

The evidence review flags this directly: "AI broad-retrieval raises the stakes — it can surface PII or out-of-consent-scope material widely and instantly." Governance load is a real, ongoing cost line for every path.

What to do.

Before choosing Build, Buy, or Partner: audit your existing consent forms against AI processing. If they don't explicitly cover AI-assisted retrieval, summarization, or synthesis of participant data, update them to include a simple statement before you put anything into a system that does. This applies to every path on this page that touches AI...which, in 2026, is all of them.

Don't know what to write?

Check out my free disclosure drafting tool to help you draft these statements for participants and partners. Check it out >

Take it with you

The printable decision kit

Everything above, packaged as a three-page PDF you can paste into a planning doc or hand to whoever's about to sign the invoice. No rankings, no "winner" — just the decision logic and the readiness dimensions drawn from the research.

Download the kit (.pdf)

There is no free research repository.
Know what you're paying for.

Work with me

Make the build-or-buy decision with your eyes open.

I help UX research teams put AI into their workflows with intention and integrity — and that includes knowing when to build, when to buy, and when to stop and fix the human system first. If that's the problem you're sitting with, let's talk.

Book a call

Quick reference

Terms used on this page

Research repository: A system (tool, database, wiki, or shared drive) where a team stores and retrieves research findings so they can be reused across projects and teams.
Atomic/nugget model: Storing insights as small, standalone units (one finding per entry) rather than as full reports. Improves reuse but adds tagging and maintenance work.
Toil: Repetitive, manual operational work that scales with usage but doesn't produce lasting value — tagging entries, fixing broken links, deduplicating records. Borrowed from Google's SRE practice (Rau, 2016), which caps toil at 50% of an engineer's time.
RAG (Retrieval-Augmented Generation): An AI technique that searches a knowledge base for relevant documents and feeds them to a language model before it generates an answer. Only as good as what it retrieves — and validation is only possible after the system is running.
MCP (Model Context Protocol): An open standard that lets AI agents connect to external tools and data sources. In this context, it allows a custom AI agent to query a vendor repository and surface insights inside the tools a team already uses.

References & Resources

Adel, A., & Alani, N. (2025). Can generative AI reliably synthesise literature? Exploring hallucination issues in ChatGPT. AI & Society, 40(8), 6799–6812. https://doi.org/10.1007/s00146-025-02406-7

Bainbridge, L. (1983). Ironies of automation. Automatica, 19(6), 775–779. https://doi.org/10.1016/0005-1098(83)90046-8

Barnett, S., Kurniawan, S., Thudumu, S., Brannelly, Z., & Abdelrazek, M. (2024). Seven failure points when engineering a retrieval augmented generation system. In Proceedings of the IEEE/ACM 3rd International Conference on AI Engineering — Software Engineering for AI (CAIN '24) (pp. 194–199). ACM. https://doi.org/10.1145/3644815.3644945

Bastius, M. (2026, June 1). Double fines in 2026: When GDPR and the EU AI Act strike together. heyData. https://heydata.eu/en/magazine/double-fines-2026-gdpr-eu-ai-act

Burghardt, J. (2025). Stop wasting research: Maximize the product impact of your organization's customer insights. Rosenfeld Media. https://rosenfeldmedia.com/books/stop-wasting-research/

Chelli, M., Descamps, J., Lavoue, V., Trojani, C., Azar, M., Deckert, M., Raynier, J.-L., Clowez, G., Boileau, P., & Ruetsch-Chelli, C. (2024). Hallucination rates and reference accuracy of ChatGPT and Bard for systematic reviews: Comparative analysis. Journal of Medical Internet Research, 26, e53164. https://doi.org/10.2196/53164

European Parliament and Council of the European Union. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), Article 50. Official Journal of the European Union, L, 2024/1689. https://artificialintelligenceact.eu/article/50/

Glass, R. L. (2001). Frequently forgotten fundamental facts about software engineering. IEEE Software, 18(3), 112–113. https://doi.org/10.1109/MS.2001.922739

Grove, L. E. (2025, September 17). Pragmatic knowledge management: From scattered insights to serendipitous intelligence. The ResearchOps Review. https://www.theresearchopsreview.com/p/pragmatic-research-knowledge-management

Hill, C., Dahil, A., Simpson, G., Hardisty, D., Keast, J., Pinn, C. K., & Dambha-Miller, H. (2026). Large language models for thematic analysis in healthcare research: A blinded mixed-methods comparison with human analysts. PLOS Digital Health, 5(4), e0001189. https://doi.org/10.1371/journal.pdig.0001189

Parrott, K. (2026, August 21). The healthcare company that built the AI tool it couldn't buy. Every. https://every.to/p/the-healthcare-company-that-built-the-ai-tool-it-couldn-t-buy

Rau, V. (2016). Eliminating toil. In B. Beyer, C. Jones, J. Petoff, & N. R. Murphy (Eds.), Site reliability engineering: How Google runs production systems. O'Reilly Media. https://sre.google/sre-book/eliminating-toil/

Rosala, M. (2024, July 26). Why research repositories fail and how to get them right. Nielsen Norman Group. https://www.nngroup.com/articles/why-repositories-fail/

Sharon, T. (2016, April 8). The atomic unit of a research insight. Medium. https://tsharon.medium.com/the-atomic-unit-of-a-research-insight-7bf13ec8fabe

Towsey, K. (2024). Research that scales: The research operations handbook. Rosenfeld Media. https://rosenfeldmedia.com/books/research-that-scales/