Promise versus reality – what AI can (and can’t) do for LGR

In the second part of this three-part series, Ross Murray explores what AI can realistically do to support local government reorganisation (LGR), and where it currently falls short.

As argued in the first article in this series, the greatest risk facing council leaders is not failing to get to Vesting Day. It's successfully reaching it while recreating the issues of yesterday at twice the size. If you ask people whether AI can help with a merger, you get completely opposite answers. Tech sales teams promise overnight savings and effortless integration. Sceptics dismiss it as an expensive distraction. The truth is in the middle. Getting it right depends on understanding exactly where AI adds value and where it fails. 

Previously, we looked closely at this balance in our research paper, AI Adoption in Public Services: What’s Working, What Isn’t, and Why It Matters. This research highlighted that one of the biggest challenges around AI adoption in the public sector is separating genuine opportunities from exaggerated claims.

What leaders are being promised

Some of the marketing claims around AI are not realistic. You can’t plug in an AI tool and expect it to automatically merge different databases, cut headcount, and balance the budget overnight.

Reorganisation is fundamentally about people and culture. AI cannot resolve political disagreements, fix broken processes, or choose how a service should be designed. Most importantly, it cannot replace senior leadership judgment. As I’ve emphasised in previous articles, AI is only as good as the information you give it. If your merging councils enter LGR with poor data and weak governance, a new piece of software will not fix that. 

Where AI is likely to help

That said, dismissing AI would be a mistake. There are practical areas where AI is already making a difference during transitions, and there are councils making use of it now:  

1. Customer contact. During reorganisation, residents often find themselves navigating multiple websites, policies and separate phone lines inherited from the previous councils. AI-powered customer contact tools can create a consistent “front door” for residents while integration continues behind the scenes. When boundaries change, residents have hundreds of basic questions about council tax and bin collections. AI assistants can handle these routine FAQs 24/7, pulling answers from a single, approved knowledge base. Many councils including Westminster, Peterborough, and South Cambridgeshire are already using chatbots to handle general queries.

   
   

2. Data unification and insight: Newly formed unitary councils must knit together dozens or even hundreds of legacy IT systems, from finance ledgers to service databases. AI can assist by intelligently matching and cleaning data across systems to create a “single version of the truth” about residents, properties, and finances. For instance, machine learning can quickly spot where separate entries refer to the same person or address, a job that can take months to do by hand. It still requires human oversight, but it can cut the time needed to get a unified data platform from three years down to one. 

   
   

3. Back office automation. Merging councils creates a mountain of paperwork, from HR transfers to processing supplier invoices. Robotic process automation (RPA) can handle these repetitive, rule-based tasks. For example, automating the migration of thousands of staff records into a single HR platform saves significant time. Generative AI can also take on basic admin tasks, like drafting meeting summaries or reviewing long reports. However, there is a catch: automation only works if the underlying process is clean. As some new unitaries have found, automating a broken process just creates faster chaos.

   
   

4. Predictive analytics and service planning. By applying machine learning to combined datasets (from multiple former councils, plus health and partner data), new unitaries could spot patterns and predict needs better. For example, Somerset Council are planning to use machine learning to predict which households are at risk of becoming homeless, allowing them to step in early and cut emergency housing placements. But it is important to keep in mind the fact that these tools need high quality and integrated data, and the necessary platforms cab take years to build. It may be better to see advanced analytics as a medium-term goal of LGR, rather than a quick win.

The reality on the ground

The big question is whether councils are in a position to capitalise on AI while they’re busy reorganising. For most, the honest answer is no. In 2025, the LGA found that 52% of councils were at the beginning of their AI journey, and only 7% claimed to be at an advanced stage. The same survey found that the biggest barriers to AI adoption are lack of funding, tight staff capacity and a shortage of digital skills. Our own research suggests that these challenges still remain.

The Public Accounts Committee (2025) has said that AI could “radically change public services” by freeing staff from drudgery and improving data use, urging the government to help councils overcome hurdles. But at the moment, new unitary councils are more likely to be grappling with basic digital tasks, like merging email systems or consolidating websites, than rolling out full-blown AI solutions.

This gap between ambition and reality means leaders must make their own practical assessment of what is possible. Lessons can be learnt from Gartner’s hype cycle (see below), which shows that AI in public services is going through the “peak of inflated expectations” and many are starting to enter the “trough of disillusionment” as they realise that AI can’t solve every challenge overnight.

Where does this leave council leaders? 

AI’s promise for LGR is real in specific areas, but the conditions have to be in place for that promise to materialise. Disappointment happens when we expect the technology to fix deep, structural problems that it isn't designed for.

The lesson is to focus on practical, problem-led uses that fit your immediate merger goals. Invest in your data quality, train your staff, and get your governance sorted.

In the final part of this series, Ross looks at the specific steps leaders need to take to manage this process safely and outlines principles for a smart approach to navigating AI while managing LGR.