Every senior leader in the public sector is now thinking about AI. Most of the conversations Ive had can be either overly focused on the tech itself or potential cost savings. Which tools should we adopt? How do we procure these safely? Which procurement route is easiest? What governance should we put in place? What are the cost benefits and upfront spend required? When will we see improvements? Which roles will go?
All critical questions but these arent the ones that will determine whether AI delivers value for organisations.
The public sector is excited about the productivity wins and improved citizen services resulting from AI, whilst mindful of the impact on existing roles and cost reduction possibilities as well as the safety and ethical challenges surrounding AI tools.
The private sector is using AI to augment roles knowing the value that can be generated and scaled. This can hopefully lead to that elusive 'growth' we all keep hearing about.
Augmentation of roles not replacement. Growth not industrial decline, though, as with other sectors that transformed following technological change, there will be challenges along the way as businesses redefine how they operate. Witness the changes in the Financial and Insurance sectors now underway prompted by AI.
The questions that will determine value for the public sector need to be asked upfront by any organisation thinking about how to introduce AI:
Where is AI actually needed - what problem are you actually trying to solve?
This was core to the AI approach I learnt about on Oxford University's AI programme, focus on identifying the real needs first. Not simply bringing in AI for AI's sake. Use it to free up time, service citizens more cohesively, make employees' daily experience better or reduce operational cost and backlogs.
Having a clear view of the organisation's problems can be transformational and pivotal. Technology however cant fix broken processes (or relationships!) it only digitises them. If caseflow processes are slow because accountability and sign off are slow or unclear, an AI layer on top will simply create faster, poor decisions rather than slow ones.
If people across functions arent working collaboratively, AI can provide the tool but wont be as effective without some good team-building that essential OD and leadership legwork upfront to build understanding and rapport, tackling the blocks and opening up dialogue.
How will your people actually experience AI in your organisation?
Public sector people are weary of change programmes and sceptical of something being done to them in the name of efficiency. AI, attracting huge amounts of publicity about job replacement, is now here and amplifying that message.
If we fail to engage people, unions, staff representatives early on, and honestly, around what it means for their jobs, there will be resistance borne out of fear.
Even when roles remain, many will see AI as an excuse to load more work upon them, before they've had that personal, direct experience that AI really can free up time to take on more work programmes, or clinics or get involved in new policy development. Take it slow and build awareness through real-time, 'lived' examples. Enable people to play with the tech and see what it can do for them first.
Who is in the room when decisions on AI choices are made?
AI systems reflect the data they are trained on (and the decisions made) by the people who build and deploy them. Dont leave it just to your techies, great though they are! The best AI projects ive seen foster a cross-functional approach with digital, HR, tech, operations, customer contact points and strategy all in the room.
In a public sector context, where people using services are often those with the least power in the system, this matters enormously. Are the communities, internal and external, most affected by automated decisions, involved in designing the necessary safeguards and access?
Common sense co-design is the key as ever with any technology, process or business change introduced.
If AI is carrying out the junior tasks that grads and apprentices currently undertake, what do they do now?
Ive been thinking a great deal about the impact of AI on jobs for young people. AI promises to take on the tasks we've traditionally reserved to train up our junior staff. What will AI do to the career pathways and 'feeder' grades to leadership roles in the future? How can we 'bring in and bring on' that talent we will need in 10-20 years if AI has snaffled all the interesting work?
Keeping the person in the loop is critical to AI success and governance. Our junior staff can perform this role, acting as the eyes and ears for senior leaders, tracking performance, managing the teams of AI agents we will be 'employing', prompting their work questions and providing critical oversight.
The question now may be how can we make that work fulfilling and a chance to learn the tools of the trade, not just how to oversee the AI tools?
How can we design jobs that dont require someone to stare at a laptop all day, 5 days a week? These are aspects of all roles we must think through carefully as we move into the new Age of AI.
However, none of the above means being so fearful of AI we stall and experience another 'AI Winter', as happened decades ago when AI wasnt up to the task or had the computing power available to operate. There are of course major global security challenges with AI and its good to hear that the UK is set to lead the debate at the next G20 meeting, developing a Global AI Code. Great news and a chance to bring a divided World together.
On a personal level, we should be playing and experimenting with AI tools now to find out what works for us and which tools suit our industries.
I use the MS domain boundaried Co-Pilot to support my work, both report writing and workshop design, and have built a little team of AI agents, who are great, though prone to arranging the odd meeting without me, usually in the early hours of the morning!
I am a fan of ambient scribes having used one for a couple of years, cutting out rather a lot of transcription following my consulting stakeholder interviews.
I know from first-hand experience that AI is making me more productive ( and freeing me from the hell that is MS Excel ... IYKYK!)
As ive said before though, AI is a co-pilot not an autopilot, im still in the driving seat, planning the route, putting the brake on when needed.
How can you approach AI in your organisation?
I think the organisations that will get the most from AI will treat it as another digital tool in the box to support them. They will focus in on, and think carefully about, those niggling issues that AI could help resolve.
Above all though, they will invest in supporting their people and the citizens they serve through good engagement, co-design approaches and training. There must be clear routes to fix issues and ways to suggest improvements that are communicated and understood. Exciting times ahead for us all.
Contact me at Alison@walkerfairway.co.uk if you want explore some of the AI approaches that might work for you or simply want to hear how to develop your own team of AI agents, but dont know where to start.
