This site is intended for health professionals only

Gen-AI at scale: A roadmap for transformation

Gen-AI at scale: A roadmap for transformation
By Carl Prest Data and AI specialist at Microsoft UK
10 June 2024

Over the past year, the Microsoft data and AI team have been at the forefront of working with health and life sciences organisations that are pursuing Generative AI (Gen-AI). In this first of a three-part series, Carl Prest, data and AI specialist at Microsoft UK, shares lessons learned on how healthcare organisations can make a start on their Gen-AI journeys.  

Generative artificial intelligence (Gen-AI) has broken down the barriers to entry for AI at large. You no longer need a team of data scientists or data engineers to produce innovative use cases; the technology has been democratised and is in the hands of the people.

Yet, for many healthcare and life science organisations, AI implementation is still a daunting and arduous process. A common question we get from working with early adopters is: What concrete steps must we take to kickstart our implementation journey?

Taking the first step towards generative AI-powered innovation is simple – it starts with coming up with great ideas.

Ideas – the art of possible

When it comes to ideation and brainstorming, engaging domain and technical experts is paramount. Domain experts have unparalleled knowledge of the business. They understand the challenges and have insights into where the opportunities lie, where there is room for improvement, and which avenues are worth pursuing.

Therefore, technical experts need to be able to put themselves in the shoes of the domain experts to ensure that everyone can go on the journey together.

In healthcare, the best ideation work often comes from bringing back-office staff – HR, finance, and administration – together with clinical staff. This helps set the right tone from the offset. It allows technical experts to hear a diverse range of views and ultimately puts them in better stead to turn ideas into reality down the line.

One example of how organisations can approach this is via ‘art of the possible’ workshops with design thinking at their core. This is a concept we use here at Microsoft, and it’s about creating a supportive and open brainstorming environment. The goal of these workshops is to get all the ideas out of everyone’s heads – no idea is stupid, and nothing is off the table. Here, all ideas find a voice, fostering an environment where imagination knows no bounds.

Choosing priorities

Following ideation, prioritisation emerges as the next critical step. This is where the team needs to decide which ideas are worth pursuing and which are not.

Distinguishing impactful ideas from the rest necessitates a systematic approach. If you have many ideas from a brainstorming session, grouping them by theme and then dot voting can be a useful mechanism for deciding which will have the most impact. Dot voting is an activity where each member of a group votes with limited dots on the ideas.

Mapping use case ideas out onto two axes of business value and complexity will then help you get closer to understanding which use case you need to begin with. The ideal first use case will be high impact but low complexity. It’s enough to grab the attention of people around your business and help you build momentum, but also simple enough that it can be achieved quickly – in weeks rather than months.

Doing a back-of-the-napkin business case at this point will also be helpful as ideas start to gain momentum. Stop and think at this point – could you measure the impact of completing this work?

Finally, focusing on risk will help complete the picture, particularly in highly regulated sectors like healthcare. It is important to commence with low-risk use cases that will gradually expand the level of risk your team is able to tolerate as confidence in the technology and processes grows.


When it comes to generative AI, making sure you have the right level of organisational buy-in to your project is critical. However, many have yet to realise that while the significance of sponsorship and organisational support isn’t a novel concept in the realm of Gen-AI, its crucial role in organisational transformation projects remains steadfast.

What has truly surprised me, however, is the heightened level of executive engagement witnessed in Gen-AI initiatives. This increased involvement could be attributed to its undeniable potential to revolutionise organisations, the remarkably low barriers to entry, and the rapid pace at which prototypes and solutions can be developed. Whatever the driving force may be, leveraging this executive interest is imperative.

In my experience, Gen-AI projects that have advanced swiftly have been sponsored by senior leadership within the collaborating organisation. Conversely, projects lacking adequate sponsorship often stall, are relegated to the realm of pet projects or are relegated to secondary tasks. So, disregarding the importance of sponsorship poses substantial risks.

Should you encounter challenges in securing the appropriate level of sponsorship, it’s prudent to revisit your prioritisation strategy and high-level business case. Assess whether the Gen-AI solution holds the potential to truly transform your organisation. If it does, obtaining the necessary executive sponsorship should be attainable, ensuring the project’s progression and ultimate success.

Common Hurdles

To make any Gen-AI project work well, it is important to identify and solve typical problems. We can grow from our mistakes if we treat them as opportunities to improve.

Here are some key challenges that the team here has observed from working with health and life science groups:

  1. Lack of sponsorship: As discussed, the primary reason promising Gen-AI ideas remain dormant is inadequate sponsorship within large organisations. Without strong backing from leadership, even the most brilliant concepts struggle to materialise.
  2. Lack of time: Implementing Gen-AI requires dedicated effort, often amidst the demands of daily operations. Establishing a robust business case and securing executive sponsorship can alleviate this challenge, enabling teams to prioritise and allocate sufficient time for Gen-AI initiatives.
  3. Lack of clarity on potential: Uncertainty regarding Gen-AI’s capabilities and potential applications can hinder progress. Understanding Gen-AI’s strengths—such as content generation, summarisation, code generation, and knowledge mining—can help focus efforts effectively. Additionally, exploring existing publicly available resources and tool kits can help provide valuable resources for kickstarting projects.

Remember that embarking on your Gen-AI journey doesn’t have to feel daunting. It’s a case of cultivating transformative ideas tailored to your organisation’s needs and prioritising a use case with a compelling business case, ensuring alignment with strategic objectives. Then, secure executive sponsorship to drive momentum and overcome obstacles.

If you keep these tips front of mind, you’re well primed to start turning your Gen-AI aspirations into tangible reality. And that’s what the next article in this series will cover – how to make it real. 

Carl Prest is a data and AI specialist at Healthcare and Life Sciences, Microsoft UK

Want news like this straight to your inbox?

Related articles