The Impact of Agentic AI: Transforming Local Government Efficiency

This article was originally published on https://futuria.ai on 31st October 2024, and was written by Rob Price, co-founder of Futuria with input from Craig Cheney from Cambridge Management Consulting . Its relevance is further amplified following yesterday's budget.

Accelerating Local Government Functions with AgenticAI

The Urgency for Efficiency in Local Government?

No-one can have missed the financial challenges for Local Government in the UK over the past few years. “In 2023, Birmingham City, Nottingham City and Woking Borough councils were reported ‘bankrupt’”, or more specifically that their forecast income is greater than their outgoings. Cost efficiency and reduction is hardly a new concept, but clearly the realities of growing and ageing populations, increasing poverty and squeezed funding are putting greater pressures than efficiencies to date have realised. Specifically this is challenging provision of both adult and children’s social care, in addition to housing and accommodation where increased need in combination with no new funding is causing significant concern.

At the recent “Future of Britain: Governing in the Age of AI” conference (July 2024), organised by the Tony Blair Institute for Global Change, speakers suggested that the only game changing opportunity presenting itself currently was the recent steps forward in AI, and specifically Generative AI and Large Language Models.? Needless to say, it will require more than a few amusing poems on ChatGPT or cool images in Midjourney to drive improvements in local services provision whilst decreasing the expenditure to levels within budget.

In the last year though we have seen some new AI developments that show promise, albeit that promise will need to translate into reliable operations within secure environments awash with personal data that must be protected. This new development is being referred to as AgenticAI, or multi-AI Agent teams. The question is, what does that actually mean?

Put yourself into the shoes of a council leader. Imagine that you had the power to bring the best people, with the best knowledge and information at hand, into a room to solve each and every problem statement that you are facing. Now imagine that we can quickly create AI Agents, with that same knowledge and information at hand, and then imagine we can get them to effectively collaborate to help you solve those problems.

It probably sounds far fetched, and yet there are already examples that are working effectively in secure organisations within the UK. Let’s explore what this can potentially enable a little more.

Budgeting and Spend Management: Enhancing Precision and Reducing Costs

What have you got planned over the next few years? What do you have to do vs. what do you want to do? What variables play into those decisions. Maybe it includes capital projects, provision of housing, technology products, services reform such as social care, operations, pensions and no doubt much more.

Imagine If you were able to do a budgetary cost estimate of everything in minutes, with multiple scenarios for each. Imagine if you could get a risk analysis for each scenario, with the degree of confidence in the execution of the project or service within the price given. Imagine if the AI Agent team could also provide you with proactive recommended interventions to help de-risk.

This is the reality of what can now be done1. We’ve seen AgenticAI deliver time savings in central government by a factor in excess of 100x, with massive cost decreases too. Fully worked up cost estimates, detailed in Word, with full referenceability in less than half an hour. For anything.

It doesn’t work entirely by magic. It learns over time. It can be pre-configured to apply your estimating methodologies and local policies, understand what has been done before (planned, estimated and actuals) and continue to learn, and verify, from other sources, including simply talking to your employees to understand more. But in weeks, you’d be amazed at just what is now possible.

Also ask yourself the question as how to find the most accurate estimate? Is it better to have teams of people follow a process to get 1 answer over time, or is it better to apply a distribution curve to a 100 or a 1000 (pick a number) automatically generated estimates for multiple scenarios to determine the statistically most likely?

Make the jump to execution. What if you had an AgenticAI team configured to support your project managers, service managers, operational leaders in everything that they do. There to help accelerate onboarding and gaining access to deep expertise, there to make informed recommendations, working in conjunction and in support of your teams. People have worried about AI replacing humans, but what if we can effectively harness it to help superpower humans?

In short, if you could quickly get to reliable budgetary estimates for everything you might want to execute, surely that would make a massive difference to budget planning, prioritisation, and start to improve against the risks of delay and ultimately therefore cost slippage avoidance.

Procurement: Accelerating Processes and Reducing Acquisition Costs?

It’s not just about cost estimating though. Apply AgenticAI to the entirely set of processes in the procurement cycle, with a focus on reducing risk and reducing elapsed time to next step outcomes. There are already now established Generative AI solutions that write bidding responses2, and no doubt soon will be in the generation of requirements documents such as ITTs, RFPs and even contracts. There are AI solutions that enable global search for any widget in any geography3, producing Gartner style reports in hours on recommended options, enabling you to accelerate the process of sourcing of likely suppliers.? And I’m sure that there will be solutions that assist with the review of responses to determine risks, likely actual costs and gaps with any supplier responding to any requirements.

It is important to recognise that there are two approaches that can work here to accelerate the procurement process. The first is traditional, mapping out the end-to-end process, determining the areas of delay or pain and focus on improving or automating those elements. The second is more novel, and perhaps new with AgenticAI. If we can identify the capabilities, tools and knowledge that is needed in that end-to-end process, ask the team of Agents assembled to identify novel approaches to accelerate the outcomes needed. In truth, there is a strong argument to try both where possible.

Delivery: Streamlining PMO Functions and Managing Risks?

Of course, it is all very well estimating costs faster, and with increased rigour, but the challenge is still to ensure that the service, project or operational needs being delivered are indeed delivered for the cost envisaged. We can apply AgenticAI here too, in essence acting as an enhanced PMO (project management office) function by taking progress input from a variety of sources, interpreting against all that is known, and making proactive intervention recommendations to the team to help keep all on track.

We could imagine an Agent Team that has specific agents focused on aggregating data – perhaps passed to it in existing excel reports, perhaps through tool interfaces to the financial systems, agents focused on determining and evaluating risks, and agents with a deep understanding of the contract terms, or operating model, resourcing or in fact anything that we can provide as a set of data or interface. There are of course numerous regulations (GDPR as a minimum), policies and ethical AI frameworks that must be adhered to, but we have seen solutions designed for highly secure environments already so we know that these are surmountable challenges. Do not compromise here though – it is critical that organisational data is protected from a security perspective, and a fully transparent, fully auditable solution is vital.

AgenticAI in HR and Finance: Driving Productivity Improvement

More widely, AgenticAI can play a role across the Operating Model of a local authority or council, to improve productivity and to enable existing teams to be able to achieve more, and deliver faster with the assistance of AI Team Members.

There are numerous use cases for its application across HR, campaign recruitment, performance appraisals, apprenticeships all examples that I’ve already seen. The technology is also beginning to ask questions of regulations – for example, for many years we have pushed job descriptions through tools that ensure gender neutrality, yet if we can easily create and promote a multiplicity of job descriptions and adverts that are targeted on niche, yet broadly diverse groups, perhaps there might be a more effective engagement across those diverse groups.?

In Finance too, we are seeing AgenticAI applied to finance functions – bringing a meld of machine learning tools with Generative AI to help automate process flows such as invoice processing, forecasting, accounting4, financial reporting and auditing.

Conclusion: Harnessing AgenticAI for Local Government Transformation

If your view on the power of Generative AI is driven by playing with ChatGPT, or creating images in Dall-E, and you’ve written it off as being not relevant to your job and your work in Local Government, then my plea is to look a little further. If you have worried about hallucination, or the security/privacy issues of using it in public sector, or the impact on jobs, then look at the emergence of AgenticAI as helping to resolve some of the features that are genuine concerns regarding Generative AI.

In regard to the impact on jobs, then it is undoubtedly true that the job landscape is constantly evolving. However, there are some wider incontrovertible megatrends that are increasingly making it difficult to recruit the people required to deliver the services needed – for example, ageing populations, shrinking populations (in some geographies). As a strong voice in the worlds CDR (Digital Responsibility) movement, then I’ve been talking about the necessity to think of these consequential impacts for nearly a decade, yet I’ve seen the reaction on the ground when public sector employees find themselves better able to do the things that need to be done to serve their departments or citizens without the need to rely on consultants in the supply chain.

Think of AgenticAI as enabling you to do far more with your existing teams, to onboard new employees faster and get them up to speed far quicker than before, to crash elapsed times to respond to public requests or deliver services. Think of it as helping to make your employees lives easier, by providing them with the information that helps them make decisions better, or complete activities faster. It is undoubtedly true that there are risks and dangers too of AI, but these can be understood, and mitigated in the context of specific use cases. Let the innovative power of its potential drive your engagement with it, rather than fear of the unknown.

In an environment where taxation is unlikely to suddenly significantly increase to provide greater funding, where the costs of delivering public services only continue to increase, we have to find some transformative ways of delivering game changing scenarios. AgenticAI presents that opportunity. We just need to figure out how to most effectively harness that power in harmony with the human teams that need that help today. AgenticAI can be a key tool for future-proofing your operations and delivering better public services for less cost than you do today.

Our product enables you to create high performing hybrid teams – humans and AI team members – working together to do some amazing things. If you’d like to know more about any of the points or organisations referred to in this article, please contact me at [email protected] .?