NHS England - Transformation Directorate

An invitation for data-rich AI problems that could benefit from a skunkworks approach

Giuseppe Sollazzo, Head of Skunkworks at the NHS AI Lab, opens round 3 of the Lab’s problem-finding programme and reveals the selected artificial intelligence projects from the last round.

At the end of January, The AI Lab launched a new collaborative opportunity: a series of Dragon's Den-style opportunities for colleagues in the UK public healthcare sector to pitch their data-rich problems to us in return for a skunkworks project. The idea is for us to judge the entries, select several of the best options and work with them for 12 weeks to get the potential AI-driven idea to a proof of concept.

Photos of the experts involved: Giuseppe Sollazzo at NHSX, Simon Snowden - NHS England and Improvement Senior Systems & Solutions Development Manager Emma Doyle - NHSX Head of Strategy Niamh McKenna - CIO, NHS Resolution Francois Lemarchand - NHS AI Lab Senior Data Scientist
The panel of experts involved in choosing the final round 2 projects

Make your AI project application

Our second round had over 30 fantastic applications - more about these projects later on - and we have now opened applications for round 3.

Our ‘pitching event’, when we'll choose the new projects, is planned for Thursday 1 July. Staying true to our iterative approach, we're making some changes to the process to make it easier for a wider group of health and care organisations to apply, and to give extra support during the application phase.

To apply for round 3, please submit your expression of interest by emailing us at aiskunkworks@nhsx.nhs.uk to ask for an application form.


The deadline for returning completed applications will be 11.59pm on 31 May 2021.

Selected applicants will be invited to pitch their problem to us on 1 July 2021.

You can get advice about what makes a good project and find out more information about the process on the AI Virtual Hub.

A good application should give a detailed understanding of the problem you face, and what you need to do to solve the problem - but not what the solution looks like or how it works. By taking your problem to the Accelerated Capability Environment (ACE) supplier community, we can choose from a number of solutions put forward by the experts. Keeping an open mind and a focus on the problem will make your application stronger.

Become a member of our AI Virtual Hub to hear about dates in May for drop-in clinic sessions. These webinars are for anyone with questions about the process or their applications - so do get in touch to join us there!

Successful applicants from the last round

More than 30 applicants applied for round 2. From hospital trusts and NHS organisations, asking for help with everything from A&E admissions to preventing patients from falling, and from appointment booking systems to optimising nursing placements.

We promised to pick and resource the ones that made the best match for an AI-driven approach and had the most potential for wide-spread use.

In the end, we chose 7 problems and asked them to pitch to our board of ‘dragons’. Many thanks to our experts from NHSX, the NHS AI Lab, NHS England & NHS Improvement and NHS Resolution for their time and insight. Between us, we selected 3 incredibly varied bids as our next cohort of projects.

Round 2 projects

Over the next 12 weeks, we'll work in partnership with the organisations that came to us, by providing the AI Lab’s knowledge and expertise together with the technical capabilities provided by the Accelerated Capability Environment (ACE) and their community of suppliers. The following will run as 12-week agile projects:

1) Long stayers AI risk stratification modelling at Gloucestershire Hospitals NHS Foundation Trust

The problem:

This project focuses on an important issue in the healthcare system: the impact of long hospital stays on both the patient and the hospital. 4% of all admissions at this trust stay for 21 days or longer, but this comprises 34% of all bed days. Long stayers have higher mortality rates, and shorter lengths of stay are associated with better outcomes.

Potential benefits:

Better outcomes for patients means trusts can focus on other patients and free up precious time. We want to be able to calculate the risk of long stays more accurately and in a customised way. Actioning this useful information will help trusts to reduce the impact on the patient of a long stay, and free up important capacity for the hospital.

2) Comparative CT algorithms at George Eliot Hospital NHS Trust

The problem:

Comparing computed tomography (CT) scans is a labour-intensive task and there are no satisfactory automation tools that can speed up the process. The major difficulty involves aligning scans 100% accurately when overlaying one on top of another, making it difficult for radiologists to accurately measure changes. Additionally, a large increase in volume of lesions may not be detected when viewed in a single plane. It is estimated that radiological misses vary from 3% to 30%.

In this highly experimental project, we will explore how AI might be able to support the radiologist comparing scans by identifying organs and lesions (tissue growth), reporting the change in size, and highlighting areas of concern to the radiologist. The project will also look at how we might develop algorithms that attempt perfect alignment of images to assess the volume of a lesion in 3 dimensions.

Potential benefits:

The benefit of a project like this is the contribution to earlier detection of potentially life-threatening tumours, along with a gain in time saving for the radiologist.

3) AI recruitment shortlisting tool at the London Talent Team of NHS England and NHS Improvement

The problem:

This project will look at the issue of bias in using AI to help compare and review job descriptions and applications. The problem of using AI for recruitment is not new and is often under scrutiny because the algorithms and systems used must be carefully modelled to avoid any unconscious bias. This project will test how algorithms perform while placing careful scrutiny on the issues of ethics, equality and inclusiveness.

Potential benefits:

Recruitment for an organisation as large as the NHS is a time-consuming and expensive operation. If artificial intelligence can successfully improve the speed and efficiency of selection processes, this could lead to fairer opportunities, greater inclusivity and reductions in time and cost.

What we will achieve

As you can see, these projects cover a range of clinical and non-clinical applications. They do have one feature in common: that of being problems seen first-hand by our colleagues working in the NHS.

By working together on these problems, we aim to “test, learn and share”.

We will “test” whether AI can provide solutions - and we are open to the possibility of failure. It is likely that some of our “tests” will not be a success, and we will share this because we think that by failing we can learn something about the safe adoption of AI.

We will “learn” using a co-production model so that we can expand our knowledge of NHS problems and educate the wider NHS about the potential of AI for health and care. This is particularly important, as one of the aims of the AI Lab is education - making the adoption of AI easier and based on solid evidence.

We will “share” through our open approach - by blogging, running webinars, and sharing open source code derived from these projects, as we’ve recently done, for example, for our pilot project Data Lens.

What about those who were not selected?

There are no winners and losers. We received an overwhelming amount of very high-quality problems and it was really hard to make a decision. Our resources only allow us to select a few for each round, but we do hope that by taking part in this selection, we’ll be starting a conversation with all the teams who pitched. We are trying to make the application process as simple as possible, and we hope to be able to work together with many of the applicants, one way or another.

Our experts

With thanks to:

  • Simon Snowden - NHS England and NHS Improvement Senior Systems & Solutions Development Manager
  • Emma Doyle - NHSX Head of Strategy
  • Niamh McKenna - CIO, NHS Resolution
  • Francois Lemarchand - NHS AI Lab Senior Data Scientist

Accelerated Capability Environment (ACE)

The NHS AI Lab is working with the Home Office programme: Accelerated Capability Environment (ACE) to develop some of its skunkworks projects, providing access to a large pool of talented and experienced suppliers who pitch their own vision for the project.

Accelerated Capability Environment (ACE) is a Home Office unit providing access to more than 250 organisations from across industry, academia and the third sector who collaborate to bring the right blend of capabilities to a given challenge. Most of these are small and medium-sized enterprises (SMEs) offering cutting-edge specialist expertise.