Picture the scenario. You’re a taxpayer living in a European country, and you’re unsure about your property tax liabilities. So you call your national revenue agency for some guidance.  You expect to be holding on for a while to speak to a person. But instead you get straight through to an AI-enabled voicebot that not only understands your queries, but also answers them succinctly and speedily. That might sound like a vision of the future. But it isn’t. At least one system of exactly this type is running in Europe today. The problem? There are very few others like it.

Why are AI-enabled services not yet the norm in public services?

This recent Accenture paper, “Transforming Public Service with AI” based on our research among 300 Public Sector leaders in Europe – provides some interesting pointers.

A top-line finding is that governments can see the massive opportunities that AI presents, and are ramping up their spending on it. Some 86% of the respondents said they intend to increase their investments in AI in the coming year. And 90% percent expect a “medium-to-high” return on that spending.

But while investment is flowing, my day-to-day involvement with Public Service organisations highlights two disconnects that are limiting the value actually realised.

Disconnect 1: What’s the strategy?

The first is that the public sector is approaching AI on a case-by-case basis without first developing a coherent strategy. As a result, lots of proofs-of-concept (PoCs) are running to test out the value from AI, and sometimes going on to deploy it in specific use cases, as with the tax voicebot I mentioned above. And AI isn’t being scaled across the enterprise to transform an ever more diverse range of service activities.

This wider scaling-up is needed, because AI has the potential to redefine the entire value chain of a Public Service organisation. So while different parts of the value chain may still be tackled on a case-by-case basis, it’ll be vital to have an overarching strategy to keep everything coherent and connected throughout.

Disconnect 2: Underestimating the full cost of implementation

The second disconnect is that Public Service organisations often underestimate the full costs of implementing AI at scale. Strangely, those that have deployed AI use cases to date, tend to think it’s a relatively cheap technology.

Why? Well, it’s a bit like building a car. If you had all the parts ready-made and put them together using robots, you’d think assembling a vehicle costs just a few hundred dollars. But that doesn’t take into account the entire value chain for sourcing the metal, pressing it to create the chassis, building the engine, the design, the electronics and so on. Factor all that in, and the cost per car rockets from somewhere around US$200 to more like US$5,000.

If you’re going to scale AI across the enterprise, you’ll need to do several things: enable your legacy landscape, redesign your value chain, and assess, develop, test out and deploy literally hundreds of use cases. Far from being cheap, addressing all of these elements will be costly, even though the ultimate return on investment will be substantial. So, on top of having an AI strategy, Public Service organisations need a strategy that’s cost-effective. Otherwise the required investment could be prohibitive.

That said, get strategic scaling of AI right and the rewards are substantial. Accenture Strategy’s “AI: Built to Scale” research found that companies that are strategically scaling AI report nearly three times the return from AI investments compared those pursuing siloed proof of concepts. And further analysis validated a positive correlation between Strategic Scaling and a premium of 32%, on average, for three key financial valuation metrics.

Where to begin…?

To achieve benefits of this order, the starting point is to view AI not as just another technology, but as one whose disruptive impact is on a totally different and unique scale.

Saying AI is disruptive might sound like a cliché. But inevitably some clichés are true, and this is one. If you look at the evolution of IT over the years – from mainframe to client server to ERP and beyond – at root it’s all been about replacing paper processes. But AI does much more than replace what was there before. It lets you create processes you could never have imagined in a pre-AI world, unleashing a degree of disruption that’s unparalleled and unprecedented.

By way of example, take voicebots, which can help to transform the call centre and front office beyond recognition. Or natural language processing (NLP), enabling machines to interpret text or the spoken word with the same ease and fluidity as a human. Or computer vision, now being used to analyse a mass of aerial image data and pinpoint property extensions that haven’t been declared for tax purposes.

…With two key steps

None of this was possible before AI. And to leverage its disruptive potential at scale, I would invite Public Service organisations to start doing two things as a matter of priority:

1. Invest strongly in innovation to identify and open up new use cases that can be addressed with AI.

The benefits will range from delivering better outcomes in a cost-effective manner to paving the way for more dynamic economic development in local regions.

2. Accelerate this innovation by working proactively to build a “GovTech” Living Ecosystem.

A collaborative community encompassing Government at all levels, startups, established companies and civil society, all collaborating without boundaries to deliver better outcomes for all and realise the transformational power of AI in public services.

The fact is, the era of AI is here – offering Public Service organisations to effect radical and positive change across virtually everything they do. It’s a golden opportunity. And now is the time to start realising it.

Subscription Center
Subscribe to Voices of Accenture Public Service Subscribe to Voices of Accenture Public Service