Skip to content

Webinar: AI Maturity in the Public Sector

On October 26, the Alternative People Survey 2026 from Re:State, Public First and Civil Service World gave the clearest picture yet of AI inside Whitehall. 82% of civil servants now use AI at work, up from 30% a year earlier. Yet the share who say their department has the capability to deliver its priorities fell from 35% to 33%. At the Ministry of Defence, 71% disagreed (Re:State, 2026).

Four days later, the chair of the Public Accounts Committee wrote to say the committee was extremely disappointed that ministers had refused to reinstate a government chief digital officer. The post has been empty since September 2025 (PublicTechnology, 2026).

Use is spreading fast. Maturity is not keeping up.

Adoption is not maturity

Two years ago the problem looked like slow uptake. In March 2024 the National Audit Office surveyed 87 government bodies. Only 37% had deployed AI, usually one or two use cases each, and just 21% had an AI strategy. Limited access to good quality data and ageing legacy systems were named as core barriers (NAO, 2024).

Uptake is no longer the issue. The same 2026 survey shows what people actually use AI for. 74% use it to draft documents and 59% to search for information. Only 16% use it to automate a process.

That is the maturity gap. Individuals have adopted AI faster than their organisations have learned to use it. Drafting help makes one person quicker. It does not change how a service runs, what it costs or how it treats citizens. Those gains need clean data, modern infrastructure, tested models and clear governance, and most of that sits at organisational level.

An organisation that cannot see this gap will keep buying tools and running pilots. It will report rising usage and wonder why the savings never show up.

Five things a mature organisation gets right

Elsewhen's research with public bodies points to five dimensions that separate organisations that scale AI from those that stall.

Opportunities. There is no single right use of AI in the public sector. A mature organisation has a repeatable way to find, assess and rank opportunities, instead of chasing whichever tool arrived last.

Data readiness. Agents and models act on the data you give them. That means clean and current datasets, strong privacy and security controls, and pipelines and APIs that let systems talk to each other.

Technical infrastructure. Legacy systems that worked for one pilot will not carry a hundred agents. Scalable infrastructure decides whether a proof of concept can become a service.

Model development and evaluation. You need a defined way to choose models for each task, clear performance measures, robust testing and the ability to switch models as better ones arrive.

Ethics, compliance and risk. Citizens and auditors need to trust the output. That takes bias checks, explainable decisions and a clear plan for when a model gets something wrong.

The Department for Work and Pensions shows what maturity buys you. A machine learning model in the Child Maintenance Service predicts which payment arrangements are likely to fail, so caseworkers can act early. DWP estimates it keeps an extra 715 cases a month out of arrears and saves the time of 49 full time staff (National Audit Office, 2026). That result rests on good data, a measured baseline and a process redesigned around the model.

Three priorities follow. Give departments real ownership of their AI programme, with a named owner for outcomes. Fix data and infrastructure before scaling, because every later use case depends on them. Then measure results in money and social value, against a human baseline, so savings are evidenced and not assumed.

Where to start

Begin with an honest score. Rate your organisation on each of the five dimensions and look for the weakest one, because that is what will stop your next pilot from scaling. For most public bodies it is data readiness or infrastructure, not ideas.

Then pick one use case that your current maturity can support and prove it properly. Elsewhen's AI maturity assessment scores you across the five dimensions and gives you a report with practical recommendations, from quick wins to longer term moves. From there, a four week sprint takes one use case from idea to evidence, tested with real users against a human baseline. Elsewhen is a Crown Commercial Service supplier on the Artificial Intelligence DPS, G-Cloud 15 and Digital Outcomes and Specialists 7.

The savings are already in the Budget. Maturity is what will turn them into results.

Find out where your organisation stands. Get in touch.

READY TO GET STARTED?

Transform how your most important work is delivered

Meet the team, discuss your ideas with our experts and receive a proposal for your project.