Accelerating the adoption and impact of AI in the public sector

Dr Janak Gunatilleke - Director, Head of Healthcare Data Analytics, KPMG UK

Alan Lewis - General Manager, Public Sector, Microsoft UK

There has never been a better time to harness the productivity and resource saving advantages that AI offers. It is the only way in which we will be able to drive the continual and needed improvements in citizen services with reduced budgets. But leaders need to recognise that it is more than simply a technology shift, but rather a transformation that impacts critical processes and decision making. It needs to be carefully planned, managed, and implemented if it is to unlock the expected operational benefits, improvement in job satisfaction for employees whilst also delivering value for citizens.

Here we talk about what we are seeing in the public sector and some lessons learnt from those that are doing it well. Before we delve into those lessons it is important just to remind us all about the art of the possible.

Although AI has been around for a number of years the specific branch of AI based on large language models – or generative AI as it has become known, has delivered a step change. It’s a fast-evolving area, and the easy-to-understand aspects focus on providing support and assistance internally to staff (writing documents, templates, meeting summaries etc) but in time this is expected to have a much larger impact. Copilot for Microsoft 365 can revolutionise how tasks are handled and seamlessly integrates with existing platforms to prevent duplication of efforts and ensure a smooth experience for the employee. It can help users remove mundane work from job tasks or complete writing tasks faster so reducing the time spent on manual and repetitive work while freeing up their time to focus on work that employees enjoy more because they feel able to deliver the outcomes that they entered public service for.

For the citizen, the opportunities for personalisation, in particular, could transform the way in which externally facing public bodies provide support. We have seen an increase in public sector organisations using generative AI as a gateway to organisational knowledge; its ability to discover patterns over disparate systems and retrieve information from multiple sources quickly allows the user to get answers to the questions they have. Taking this a step further it can bring together multiple sources, both organisation and citizen data to provide specific responses on a case by case basis.

Even more significantly it could lead to completely new operating models which substantially remove steps in the processes and ways of working that we see today. This will help organisations to optimise their operations and processes, reduce costs, and alleviate digital debt. By exploring the true art of the possible leaders will put themselves in a really good position to begin to transform the public sector bodies that they are working in – as long as they know where to start!

As we know the potential uses of AI are myriad. As you begin to innovate, engage and experiment you will want to provide assurance to internal stakeholders and citizens that you are acting responsibly and ethically. For example, Microsoft has established the Responsible AI standard which provides a framework for developing and using AI systems responsibly.

As we embark on this new way of working, public sector organisations should develop and be governed by responsible AI strategies, which should incorporate principles, practices and tools to enable those across the organisation to assess, adopt, and manage AI. Having this detail in place can also provide confidence to your employees and allow them to take the first step on seeing what AI could mean for them personally.

In such a large field it is hard to know where to start. Public sector bodies can help themselves by narrowing down their focus to specific areas where AI can bring value. Finding examples, use cases, early on that solve a problem or generate an improvement is critical.

When identifying use cases what you are looking for is stories and evidence that prove the value to the organisation. You might find it useful, therefore, to consider a number of dimensions such as the level of impact on the citizen journey or the areas of operation (front, middle, back-office functions). If all your use cases are back office, with limited impact on the citizen journey, you may find it hard to create energy for the ongoing transformation. Involving your teams to create use cases (and subsequently proof of concepts) that cover a range of aspects of what AI can deliver can be a critical part of overall change management.

Don’t establish the framework in a vacuum – do it at the same time as you start to experiment with use cases and you can make the governance more practical and relevant to the day-to-day experiences of your stakeholders and employees.

This is not a technology change, it is a transformation, and therefore all the usual aspects of change management will need to be considered from stakeholders to funding, from risk management to creating energy and enthusiasm. When we have seen this work well the program is able to easily answer the following questions;

• Does it tie back to a problem or used case that is real and important so that it gets the attention it deserves?
• Is the roadmap clear and ideally self-funding?
• Are there quick wins in the plan that not only improve services but also release savings?
• Is the work building confidence so that more complex change that is transformational is being de-risked?

Any AI plan will consist of many steps and phases. Wherever you are on the journey, creating a structured plan will be critical to allow for effective implementation, adoption and evaluation. 

We know AI in the public sector can improve the citizen experience. Chatbots offer access to high quality information, enabling them to address the bulk of citizens’ more routine questions. Content generation is facilitating digital interactions which are customised for the end user, through chatbots and online services, meaning quicker response times and less user navigation. Citizens will no longer have to read through long lists of FAQs to find what they need.

AI will also improve the working experiences of employees who have joined the public sector. Call centre agents can capture information from calls with AI automation reducing the need for the citizen to repeat their circumstances, can use Azure OpenAI services to translate documents into multiple languages creating cost and time savings and even make suggestions about how to simplify documents to make them more accessible.

All of this is possible now, but it will be important to manage the expectations of employees who will want the more mundane aspects removed from their role with AI and also of citizens who have heard the media narrative about the magic AI can deliver.

By continuing to remember that there is a human at the heart of all this work we will be able to see transformational change that impacts large parts of our society.