Shaping Safer AI Adoption: Examining the AI Readiness Project

The AI Readiness Project

AI technologies are transformative and have the potential to revolutionise the health and care sectors. In healthcare, AI is already widely used in some clinical pathways, such as skin cancer diagnosis or stroke detection, and in social care workers frequently report significant time savings due to the increasing prevalence of AI scribing tools. There is also a distinct political drive in the UK to make more use of AI technology to support health and care delivery. Yet the benefits of AI tools are not guaranteed, and variable organisational uptake, poorly implemented systems, or organisational failure to adopt AI technologies which could reduce delays and inefficiencies are some potential harms associated with AI adoption.

The AI Readiness Project, commissioned by NHS England and The Health Foundation, aims to support organisations in assessing and developing their readiness to safely adopt specific AI technologies while improving outcomes, upholding public safety, and enabling responsible innovation. This case study describes the project lifecycle and outputs.

Mapping the Project Lifecycle

The AI Readiness Project followed three distinct project phases:

Inductive phase: The project began with an exploratory phase designed to understand what really shapes an organisation’s readiness for AI, and what kinds of support might be most helpful. This started with a focused review of existing research, building on earlier evidence in the field. The team also carried out semi-structured interviews with expert contributors to gather a wide range of real-world insights and suggestions to build on this academic knowledge.

Deductive phase: After gathering these initial insights, the team focused on making sense of the patterns that emerged. The aim was to develop a framework which clarified the distinct causes of harm introduced by AI tools within provider organisations, and to determine how these risks can be effectively controlled to support safe and effective AI adoption.

Integration and validation phase – Next, the team strived to ensure the framework stood up to scrutiny and resonated with people in real-world health and care settings. At the integration stage, the team refined the framework by testing it with different groups via separate public and professional workshops. Following this, two initial outputs were developed – The AI Readiness Assurance Framework and the AI Readiness Checklist – and validation and refinement of these outputs is being achieved through a national survey and semi-structured interviews with a diverse group of health and social care providers.

Examining the Project Outputs

AI readiness is conceptualised as having two distinct components – decision making readiness and implementation readiness. Decision-making readiness refers to an organisation’s ability to make an informed choice about whether to adopt a particular AI technology, based on a realistic understanding of its potential benefits and harms in their local context and the extent to which risks have been mitigated. Implementation readiness refers to the organisation’s ability to safely put a chosen AI technology into practice, with a clear grasp of the associated risks and appropriate safeguards in place. These insights informed the development of two initial resources:

The AI Readiness Assurance Framework supports organisations to strengthen their decision-making readiness for AI. It sets out distinct categories of potential harm (external, financial, operational, service outcomes, and workforce) alongside risk controls (organisational structure, human capital, problem formulation, adoption integration, and evaluation and monitoring), expressed at a level to ensure applicability across diverse settings and AI technologies. By understanding these causes and controls and embedding them into local processes, organisations can enhance the quality of their decision-making and more effectively prioritise investments across innovations which meet the needs of their services and service users.

The AI Readiness Checklist guides organisations to assess their implementation readiness for a specific AI technology, by identifying potential harms from a specific AI tool, mapping existing risk controls, and offering practical guidance to address any gaps. In doing so, the checklist helps providers demonstrate good practice in AI governance, create a common language for stakeholders, attract responsible suppliers, and reduce legal risk. At a system level, it supports consistent, transparent and safe AI adoption across health and care services.

Get Involved

We welcome feedback from across the health and care system, including frontline, digital teams, governance professionals and policy partners, to inform the final publications. If you have any reflections, detailed comments or suggestions, please complete the survey which will be open until the end of May 2026.