In our experience, there are three key elements that can help organizations identify, develop, and deploy AI to maximize benefits and improve productivity:
1. Integration into your workflow
Even in non-clinical areas, it's important to consider existing workflows and processes: AI solutions need to integrate with other systems and ways of working (including collecting and re-entering data into other “main” systems), otherwise it will just create more work and take longer overall.
2. Identify and track benefits
Given the importance of increased efficiency and productivity, it is essential to have mechanisms to set performance benchmarks and monitor the benefits gained. It is wise to pilot solutions, evaluate them, and depending on the results, decide whether and how to scale up.
The roadmap should be self-funded wherever possible, starting with phases that improve services with quick wins to free up cash to fund subsequent phases of increasingly complex and ambitious opportunities. This iterative approach ensures buy-in and sustains momentum throughout the transformation.
3. Take the right approach to assessment and risk
Evaluating non-clinical AI solutions requires a different approach than clinical AI solutions – for example, randomized controlled trials are unlikely to be required – and a pragmatic evaluation methodology that balances rigor with execution timelines is needed to keep pace with the rapidly changing AI technology landscape.
Risk management remains a critical requirement for non-clinical AI solutions. The KPMG Trusted AI approach is our methodology for designing, building, deploying and using AI strategies and solutions responsibly and ethically. It consists of 10 key elements:[viii]:
