Author: Miika Leminen, Head of Innovation at CGI, SOTE’s AI solutions
Who gives innovation a push when it gets stuck?
Over the years, I have been involved in quite a few healthcare innovation projects, sitting on different sides of the table. They have aimed to make the world a better place through technology, drawing on top experts, research and data. Typically, the vision is clear: people genuinely want to make things better, and they put their hearts into doing so.
And yet, far too many good ideas never make it all the way.
One piece of feedback from a physician about a solution we had developed has particularly stuck with me. It went roughly like this: “This may not be the most useful thing in the world, but the problem itself was interesting.”
Technically, the solution may have been quite successful. But in healthcare, an “actually, that’s quite nice” innovation is not very far from a useless one.
A good pilot is only halfway there
When we talk about innovation, attention is often directed towards the technology itself. What I have noticed, however, is that promising innovations rarely fail simply because the technology is not good enough. Instead, there are easily a hundred other places along the way, particularly in the latter half of the journey.
These little hills often have to do with things such as turning a capability into a tool that fits naturally and cognitively into people’s work, and reorganizing workflows so that the expected benefits can actually be realized. Or with the rather fundamental question of who is ultimately willing and able to take on the responsibilities of being the manufacturer of the innovation, something that typically requires investment and a willingness to take risk.
An innovation partner for both uphill and downhill
This is one of the things I have learned both in the university hospital world and now at CGI.
Finland has no shortage of either technical expertise or healthcare domain knowledge. We run high-quality experiments, research projects and pilots. What is not always as easy to find is clear ownership and leadership for climbing that final hill, getting the solution all the way into production.
This is an area where we at CGI want to learn to take on a bigger role.
Our aim is not to be the partner that turns up for the technical implementation and disappears once the demo is ready. We want to combine the healthcare domain understanding and implementation expertise we have accumulated across our different software product solutions, to create something impactful for the everyday work of health and social care professionals.
From voxels to commercialization
The CleverHealth Network AI Head project is a particularly concrete example of this for me.
In fact, the most recent proper piece of software code I wrote with my own two little hands was related to that project, several years ago. We built tools for annotating the training data (in practice, haemorrhage voxels in head CT scans) and for validating those annotations. Since then, most of my hands-on work has shifted towards emails and presentation slides…
At the time, the world still looked very much like a research and development problem. It was only later, after moving to the vendor side, that I really began to understand how much more is required once the technology itself starts to work. Developing and bringing AI-enabled medical devices to market is quite a bit more complicated than you might initially imagine.
One lesson from the AI Head project has stayed with me in particular: along the innovation journey, the hardest challenges change shape as you move forward, and so do the ways of solving them.
Moving beyond “quite nice” innovations
RDI activities create promising solution concepts, and the best of them can be developed all the way into products. Change management and assessing real-world impact, in turn, determine whether those products ultimately generate any meaningful benefit.
Somewhere at the end of the chain, there should be a person whose care, working day or life improves enough to outweigh the costs of implementing and operating the solution.
Otherwise, we have mainly created interesting technology, without impact.
This is why we think the CleverHealth Network model works exceptionally well.
At its best, it turns the usual setup the right way round. Instead of simply looking for someone in healthcare willing to experiment with a clever technology developed by a company, the starting point is an acute, unresolved problem from everyday healthcare. Clinical, technical, research and commercial partners who are motivated to solve that problem then come together around it.
That significantly increases the chances of getting from “quite nice” to genuinely useful.
CGI has been part of CleverHealth Network for a long time, and I hope that in the future we will increasingly play exactly this kind of role: bringing new ideas to the table and connecting research and technology with practical implementation.
And above all, taking responsibility for helping the best ideas make the journey from experiments into real-world use.
