What determines success or failure in AI initiatives?
Two weeks ago HBI hosted a webinar on the topic of ‘Value creation in healthcare organisations’, with lots of interesting discussion and insights from Ilmarin Schietzel, CEO of German physiotherapy chain United Therapy, and Klaus Boehncke, Global Digital Health Lead at consulting firm L.E.K. Consulting.
Much of this centred around digital and AI implementation in healthcare, and the factors that organisations need to prioritise to get this right.
The big takeaway is that success or failure isn’t usually determined by the actual technology itself, but by how it’s deployed. Too many organisations see AI purely as a labour-saving and cost-saving tool, and give too little thought both to the full range of possibilities the tools open up, and to their limitations.
“Where many digital initiatives go wrong is organisations are trying to automate the processes that already exist without thinking about how AI redefines the art of the possible,” Boehncke said.
“Instead, organisations should be redesigning processes from first principles; from their patients’ point of view, or their clinicians’ point of view. Redesign processes rather than try to automate what you already have!”
“You cannot — or you don’t want to — scale AI on inefficient or fragmented processes,” Schietzel declared.
Boehncke also highlighted the need for ‘change management’ in this context, and the importance of getting buy-in from employees and bringing them along in the transformation process.
Boehncke offered an example of a project that his own employer, L.E.K. Consulting, has been experimenting with: an AI ‘centre of excellence’ where people within the company can experiment with and share best practices for new tools: “When people joined this virtual community they could immediately see how these use cases were applicable to them. I think creating that kind of pull is really important in terms of change management.”
Then there’s the importance of having good quality data, as AI tools are only as good as the data they’re working with. “The data might not be ready, and the systems may not be as productive as people thought they would be,” Boehncke said. “You shouldn’t implement before actually doing some of your homework around where you need to integrate data points and how you bring all that information into the AI system.”
Schietzel predicted that the organisations that win over the next decade will not necessarily be the ones with the best AI models — “They will be the ones with the best data, the strongest processes and the highest adoption rates.”
If you missed the webinar you can watch it here!
We would welcome your thoughts on this story. Email your views to Martin De Benito Gellner or call 0207 183 3779.
