This website uses cookies to ensure you get the best experience. This enables us to understand how you use the site and track any patterns with regards how you are using our website.

By continuing to use our website you consent to all cookies in accordance with our Cookie Policy.

Skip to content

Efficiency and productivity are words that are being used in relation to healthcare systems across the world – many are struggling with a shortage of clinical staff combined with an ageing population, and so are tasked with the challenge of doing more for less.

Artificial intelligence (AI) is set to be a crucial part of the solution for this productivity challenge. Its impact is perhaps more visible in our everyday lives, with AI already being used to perform routine tasks at work, to make virtual assistants like Siri and Alexa even more helpful and for getting accurate recommendations when we’re shopping online. (Full disclosure, I had ChatGPT critique this blog for me, I found it harsh but fair.)

The rate of change is remarkable. OpenAI released their latest model called GPT-4o which is rumoured to have a future in powering Apple’s Siri assistant. The vision of machines that we can talk to in order to carry out our wishes is edging ever closer. The concepts we have previously seen in films and on TV may become reality. Will we continue to use our keyboards and mouse as the input tools or will we use voice to command a fleet of AIs to do our bidding?

But, as with all advances in technology, AI is only as strong as its application. At this very moment, thousands of companies around the world are frantically scrambling to understand how they can bring AI to their products. The airwaves are awash with companies advertising this. Potentially millions more are dreaming up ways to bring new products into the world. Many of these are simply leveraging the power of the Large Language Models that OpenAI, Google and the other large technology companies are making available – at a price. 

However there is another approach which is to leverage the power of these new models with existing datasets. This is where ImproveWell sees the path forward to delivering impactful AIs for our customers. Our aggregated dataset has grown in partnership with more than 70 NHS organisations and 10 global health organisations. One part of this dataset comprises suggestions from frontline healthcare staff on what can change to improve patient care. We currently present this thematically to review the changes are being suggested. We believe this is an area where AI can significantly enhance our users experience and insights. By using our dataset to train our models, all of our customers can benefit while maintaining data privacy and security. 

An Innovate UK grant has enabled us to team up with Bournemouth University to enter into a Knowledge Transfer Partnership (KTP) to develop these bespoke AI models for our platform. This KTP will allow the platform to provide AI-powered data insights, enabling teams to quickly identify the most promising ideas and frequently reported challenges, and analyse characteristics that make ideas successful. This will help ensure valuable healthcare resources are used effectively.

The project began in November last year and we are making steady progress prototyping our models. We intend to introduce AI-powered features as the models come on-stream. For example, our new Impact Analysis Tool will harness this technology to show predictions against ideas of implementation, potential cost savings and even which ideas could contribute to retention – building on the work of our White Paper from last year.

Developing products in AI is a new experience for many technical teams, these models by definition produce output that can be different each time. Building confidence in this is a well-known issue in AI, models can famously hallucinate which is why we will release AI-powered features incrementally and through Beta programmes. We will talk to our AI about the data and have it learn from our feedback. Over time, the models will become more accurate which will allow us to further scale up ImproveWell and shape our offering to customers. 

Listening to ideas and ensuring staff feel their voices are heard in itself has a positive impact, so it is important to do this across different levels of healthcare organisations and to not come with preconceived ideas of what a good idea is. AI can help teams prioritise where to invest resources in implementing ideas, by categorising them in real-time and highlighting the ones that are most likely to have an impact. 

While many apps already use AI to process language from big datasets to make meaningful analyses, we’re proud to be the first platform to develop bespoke AI technology geared towards turning suggestions from frontline workers into improvements to healthcare systems, ultimately saving them time and money through highly effective prioritisation.

By John Masterson, Chief Technology Officer at ImproveWell

Sign up for the Engagers & Improvers monthly newsletter

Insight, news, and opinion. Direct to your inbox.

* indicates required

No spam – just valuable content. See our Privacy Centre.