The evidence
The Institute's claims are bounded by what can be shown. This page holds the field's failure pattern, cases from the work that shaped the discipline, and the standards under which future evidence is published.
Evidence I · From the wider field
95%
of enterprise generative AI pilots deliver no measurable return
MIT NANDA · 2025
42%
of companies abandoned most of their AI initiatives in 2025, up from 17%
S&P Global
80%+
of AI projects fail, roughly twice the rate of comparable non-AI projects
RAND
52%
of workers are worried about AI in their workplace
Pew Research · 2025
The failures are not technology stories. They are stories about people who were never brought along, and the successes are stories about people who were.
These figures establish the problem, not this method’s success. Each will be linked to its original source with scope notes as the library is compiled.
Evidence II · Formative cases, before the founding
Three cases from clinical and delivery work led by the Institute's founder before its founding, told from the person outward, because that is where translation succeeds or fails.
Case Study I · Discovery
A hospital division asked for a chatbot. The discipline's first questions revealed that the real problem was how work moved between people; the answer became a staffing redesign, and the chatbot was never built. The most valuable outcome of a translation is sometimes the tool it prevents.
Case Study II · Identity
An automation promised to return ten hours a week to a clinical team. Discovery showed the ten hours it saved were the hours of human contact the team loved most, the reason they had entered the profession. The project was redirected before it dismantled the meaning it was meant to serve.
Case Study III · Elevation
A cancer care platform succeeded precisely because it was translated, not deployed: fitted to how clinicians actually judged, communicated, and cared, it made them more themselves, not less, and the capability remained with the teams who ran it.
Across every case, the variable that decided the outcome was never model quality. It was whether discovery, identity, and trust were examined before anything was built, the precise work the four questions exist to do.
Formal, fully documented versions of these cases are being prepared for publication under the standards of the Academic Council.
Evidence III & IV · Current demonstrations and verified publications
Current demonstrations: the first countrywide demonstration is under way; milestones and findings are published as they are reached. Verified publications: none yet. The bar exists before the claims do, and that order is deliberate.
Every engagement is documented truthfully, including what does not work; negative results are recorded and published with the same care as success. The first countrywide demonstration is under way and is stewarded quietly to protect the focus of local partners; complete, open-source case studies are published as milestones are reached. Nothing is presented as evidence that has not met the standard held by the Academic Council.