Transformation Directorate

Crowdsourcing Our NHS AI Lab Skunkworks Project

Kassandra Karpathakis and Lauren Harkins

A Skunkworks is an unfamiliar concept to most people. It is really about creating an environment to stimulate and grow innovative ideas. We wanted to test how to get this environment right before opening our doors formally. So, we decided to ‘soft launch’ the NHS AI Lab’s Skunkworks by crowdsourcing one of our first projects from within NHSX, the Department of Health and Social Care (DHSC) and NHS England and Improvement (NHS E/I). Here is how we did it, kept it lighthearted and succeeded in selecting our AI Skunkworks project.

The NHS AI Lab Skunkworks complements the funding, guidance and focused project work in the wider NHS AI Lab by supporting the health and care community to take some of the most innovative artificial intelligence (AI) ideas from scratch to a proof of concept. We are a team dedicated to finding ideas driven by clear needs derived from challenges people face in the health and care system. Once we find these ideas and understand the need, we work with teams and experts to quickly validate if AI is actually able to solve the problem. 

“AI can be defined as the use of digital technology to create systems capable of performing tasks commonly thought to require intelligence.” 

- Office for AI, A guide to using artificial intelligence in the public sector

Leaning on the Office for Artificial Intelligence’s (2019) definition of AI technology we encourage ideas from across machine learning, predictive analytics, image recognition, robotic process automation, and natural language processing.

In the end, the most valuable proof of concepts may be adapted into a minimum viable product for further development. Other proof of concepts will help us and teams in the health and care community learn and up-skill in how AI can be used to meet current and future challenges. 

The Dragon’s Den

To get the ball rolling, we crowdsourced some of our first AI Skunkworks projects from teams within NHSX, NHS E/I and DHSC who have first hand experience of issues facing health tech delivery. To keep it fun, stimulate creative thinking and cross-organisational collaboration we went with a “Dragon’s Den” style approach to selecting our projects. 

Over the summer, teams were invited to explore their AI ideas with subject matter experts from NHSX. These experts provided support and guidance to interested teams - helping scope their ideas, make connections and finesse their ideas for the Dragon’s Den. 

By the end of summer, five teams were ready to pitch to the Dragons! 

For those who aren’t familiar with the TV show, a Dragon’s Den is a chance for budding entrepreneurs to pitch ideas to a panel of experts who will decide if it’s worth investing in. Our four Dragons, named the AI Dragon, Data Dragon, Commercial Dragon, and Design Dragon carefully listened and assessed whether or not these five big, innovative ideas met the scope and criteria for being awarded AI Skunkworks support. 

The Dragons played a critical role in the due diligence process prior to support being awarded by the NHS AI Lab Board.  

Introducing Data Lens

We selected Data Lens to be one of our first AI Skunkworks projects. The Data Lens focuses on bringing together disparately published health and social care metadata into one coherent view. Metadata summarises data about data, to make it easier for people to find and work with it and to describe the data. 

The team intends to use AI techniques and tools, including natural language processing in different languages, to make this metadata easier to locate, search and use for data analysts across the health system. We hope the Data Lens will enable analysts within the NHS who are researching initiatives across health and social care to unearth more data sources, connect diverse data sets and improve the overall use of data in their work. 

It has been incredibly rewarding for us to see how the idea pitched by Paul Ross (Data Engineer) and Craig Shenton (Data Developer) evolved from an initial conversation, to surviving a grilling by the Dragons, to successfully presenting their vision to the NHS AI Lab Board. 

We will publish more on Data Lens once the proof of concept is completed.

What did we learn? 

The ‘soft launch’ of the NHS AI Lab Skunkworks helped us quickly test an approach that balanced creativity, support and collaborative working with commitment, due diligence and rigour. We also learnt more about the appetite for and knowledge of AI within NHSX, DHSC and NHSE/I. 

We were able to create a non-hierarchical environment to encourage open discussion, testing and exploration of ideas. Our messaging and communications focused on teams coming to talk to us at drop in sessions, even if they weren’t sure if AI could solve their problem or how their idea could work in practice. 

We offered support to teams every step of the way so that people could access the expertise they needed to finesse their pitches for the Dragons. By doing so we also up-skilled teams in what AI is, how it could be used to solve problems in health and care, and how to scope AI-based health tech projects. 

We highlighted the importance of project ownership and about this being a great learning opportunity. A mandatory requirement was that successful teams dedicate a minimum of two days a week to the project throughout its lifecycle, working closely with the AI delivery team. This isn’t just about project ownership for us, it is about harnessing the learning, up-skilling and being able to share that back to NHSX and wider.

The learning from our ‘soft launch’ forms the foundation of how the NHS AI Lab will take the AI Skunkworks forward - with or without the Dragons. Watch this space!