In short
A checklist of what should be in place before an AI system goes live, in eight areas including staff readiness and human factors. It was built from one hospital’s experience. Use it as a prompt list, not as a standard.
For whoever runs the assessment: an HR or operations lead, an internal project team, or an outside consultant.
What it measures
Whether the conditions for a safe AI deployment are in place before go-live. The checklist has eight areas. Two of them are staff readiness and human factors, which many technical checklists leave out.
Each item is written so that it can be audited: either the thing is in place or it is not.
How to use it
- Read it as a list of questions to ask before go-live.
- Strike the items that are specific to healthcare and add your own for your sector’s rules.
- Give each item an owner and a date.
- For the staff-readiness items, use a proper measure, such as a knowledge test or a readiness survey.
What it does not measure
- Settings outside a hospital. Items about clinical safety will not fit a business. Others will be missing.
- Attitudes, skills or culture in any depth. A tick in the staff-readiness area is not a measurement.
How it was tested
The authors built the checklist from safety and implementation gaps they met when deploying AI at a hospital in the United Kingdom’s National Health Service. Experts in that setting judged whether the items looked right and covered the ground.
That is the weakest kind of test.
Licence and access
The article is open access under a CC BY licence, so the checklist may be adapted with credit.
Alternatives
Other assessments with a similar job, and when to prefer each.
- What makes this technology project hard? NASSS-CAT: complexity assessment tools for technology projects Prefer it when you want materials with wider use behind them, and an interview guide as well as a checklist. The project team, and staff in interviews · Two checklists, a 35-question form and an interview guide Free to use, ask before changing
- What will help or block the rollout? Consolidated Framework for Implementation Research (CFIR) Prefer it when you want to interview people about what will help or block the rollout. People at several levels, in interviews · Interviews on the factors you choose, out of 48 Free to use, ask before changing
The details
- Testing
- Built from real AI deployments at one UK hospital and reviewed by experts there. Nobody has tested it elsewhere, and nobody has shown that passing the checklist leads to a better rollout.
- Looks at
- Organization
- How the answers are collected
- A checklist completed by the team responsible for the deployment.
- Length
- 8 areas of requirements that can be audited
- Cost
- Free
- Licence
- CC BYYou may translate and adapt the checklist if you credit the authors.
- Languages
- English
- By
- Adesuyi, Singh, Patel and Saravanan, 2026
Sources
- Adesuyi, Singh, Patel and Saravanan (2026). Translating Real-World Safety and Implementation Gaps Into a Deployment-Derived AI Readiness Preimplementation Checklist for NHS Health Care Providers: Checklist Development Study. JMIR AI. doi:10.2196/93900
Related studies on this site
- Which barriers hold back AI maturity? A survey of 108 specialists van Veen, Zhu and Ravesteyn (2026). How AI Implementation Barriers Shape AI Maturity? The Interacting Roles of Skilled Employees, Technological Infrastructure and Data Quality. CONF-IRM 2026 Proceedings, 17.