Transformation

Lessons for government from AI rollouts

Written by Caitlin Barr & Maya Sgaravato-Grant | Jun 25, 2026 12:07:49 PM

For Alex Fortescue-Webb, Head of UK&I, as well as Global Head of Legal Engineering, at Legora, there are four key lessons that government can learn from the company’s major AI rollouts.

These insights were presented in a session entitled Designing a rollout for success: Practical tips for Government Delivery at Government Transformation Summit on 25 June.

The first and most essential takeaway is the need to recognise AI’s transformative potential, he stated. It is by enacting a fundamental shift in the way that AI systems are treated within an organisation, acknowledging them to be core to operational change rather than simple assistants, that their benefits can be maximised. 

The next three lessons related to the more strategic steps organisations can take to ensure the technology is used to its full advantage.

Fortescue-Webb highlighted activation as one - the higher the percentage of active users in the first six weeks, the more likely it is that rollout will be successful in the first nine months.

He said: “When you're rolling out AI to any client, you have an early window where you have a ton of interest and attention from the organisation.

“Making sure you take advantage of that is hypercritical.”

Leadership can play a key role in this, with the regular usage of the platform by C-suite users having a visible impact on uptake.

Senior figures who encourage others to use AI, share their learnings and support team members in picking up platform knowledge make a huge difference. The reverse is also true - leaders who are negative or neutral about AI similarly greatly affect outcomes.

Finally, users who champion the platform are far more successful at increasing engagement among their colleagues than outside trainers or even central teams who may be perceived as telling people what to do. If over 20% of a workplace are ‘power-users’, consulting with each other and learning from peers, overall usage rises. Setting up a ‘champions network’ could be a valuable way of ensuring AI platforms are used effectively in the long-term.