Mind the gap: Why AI projects need operating model
With the massive hype around AI, public service leaders could be forgiven for seeing AI as a quick solution to service delivery challenges. However, simply introducing new tools does not guarantee a smooth path to better outcomes. In our rush to capture value from AI, my fear is that we forget the fundamentals of how to do technology adoption well. AI sits in a class of its own too and I believe that the way in which it will fundamentally reshape our world, demands that we learn the lessons of past failures. There are many.
I might sound like I’m stating the obvious when I say new technology requires alignment with existing structures, strategies, and cultures. My fear is that if we succumb to the hype we will stall or, worse, create complex issues that undermine trust in both the technology and public service leaders.
In my advisory work, I have seen tech implementation fail when not embedded in a broader perspective that considers the DNA of an organisation. Leaders must take time to clarify how AI fits into broader strategy, who will own it, and how systems and data will interconnect. Misalignment, wasted resource, employee resistance and citizen opposition are often the result, in my experience.
I advocate for operating models as a lens through which to consider how AI can be safely and effectively implemented across an organisation or even complex system. Operating models help ensure consistent data standards, transparent decision-making processes, and clear accountability. That infrastructure minimises duplication and inefficiency, letting AI capability develop in a stable environment. The culture piece is vital. If teams do not understand the purpose of AI or see potential benefits in their day-to-day work, they may avoid adoption or even sabotage transformation efforts – inadvertently or otherwise.
In my view, a structured, organisation-wide approach offers the best chance of realising an implementation’s AI’s value. By seeking to anticipate how new technology interacts with governance, skills, leadership, and culture, organisations are more likely to see tangible benefits and position themselves for future transformation. Adopting AI without preparing the organisation is incredibly risky – but I’m beginning to see it more and more.
There are plenty of examples of failed public sector technology implementations including the now infamous NHS National Programme for IT (NPfIT). Launched in 2002, the programme sought to digitise healthcare across England. It failed. Many local NHS trusts found the imposed systems incompatible with existing clinical workflows and organisational cultures (Greenhalgh et al, 2010). The programme was ultimately dismantled, with huge financial write-offs. In my view, we stand on the cusp of making similar mistakes again.
NPfIT amongst others, is a salutary reminder that technology alone is not enough to drive sustainable benefits. I believe that the lesson remains the same: building the right operating model is essential for technology to deliver tangible value and in our new age of AI, we ignore this at our peril.
Read More Workforce & AI Skills
