15 minutes with: Becky Noble

How would you describe your role to somebody who wasn’t familiar with consulting?

I help public sector organisations deliver complex digital change and make it work in the real world. A lot of it is about using AI in a practical way, not just for the sake of it, but to speed things up, improve decision-making, and modernise legacy systems so they meet citizens’ needs and expectations. I care about AI because it gives us a real chance to change how government works at the scale citizens actually experience it. This doesn’t mean marginal gains. It means fewer people falling through the cracks, faster help when it’s needed, and services that work around people’s lives rather than forcing them to navigate the system. My role is about ensuring the strong, responsible, and safe delivery of new technology.

My vision is a government that anticipates need, that learns and improves in real time, and that uses AI to free up human judgement for the moments that genuinely matter. Done well, citizens shouldn’t notice AI, they should just feel that government works better for them.

What makes PA stand out?

We start with the real problem and the people living with it, and use our 80‑plus years of hands‑on delivery experience to create real outcomes. We’re genuinely human‑centred, and we’re very clear that technology, especially AI, has to serve people, be used responsibly, and earn trust. So, we start by setting out the value we’re trying to create, not the tech, and then work out the right solution. What’s changed is that AI‑powered prototyping lets us do that much faster than before. And all of this is governed by strong ethical frameworks, which matters even more in the public sector.

Importantly, we practice what we preach. We’ve gone through constant innovation and evolved in line with new advancements and trends so we can provide the latest offerings and capabilities to our clients. This is particularly true of AI. While we have substantial track record with our clients, we started with ourselves. We’ve adopted agent-based solutions and measured the impact they had. Our experience isn’t just grounded in what’s worked for our clients, but what’s worked for ourselves. It’s a bit cliche, but the other thing that makes PA stand out is the people. I love working with my colleagues. We have compassionate experts, all focused on delivering lasting change. We strive to do the right thing, not the easy thing, and support and challenge each other to think outside the box.

What is the biggest challenge your clients face when putting this technology into practice?

I work with big central public sector departments, and pretty much all of them are struggling with the need to deliver safe, joined-up, digital services that meet rising citizen expectations. And they’re doing that on top of ageing systems and constrained budgets. There’s a huge interest in AI and new technology, but also an understandable concern about risk, ethics, and unintended consequences. Many organisations are still working out how to take advantage of the tools responsibly without over-buying or locking themselves into solutions they don’t really need.

Departments are getting stuck in ‘pilotitis’. They’ve picked lots of small use cases and are running prototypes, but don’t have the framework to bring these pilots properly into BAU from a process, technology, and data perspective. It’s costing them a huge amount to prove small, localised value, but they’re not achieving enterprise-level improvements. We want to switch to the thing that makes a really big difference. What will make them 10 times more efficient? What will take a problem they’ve never been able to solve, and find an outcome that truly drives different behaviours? How do we do that with the right checks and balances and the right risks associated? How will they measure return? And how will they make certain humans are in the loop?

There’s also a growing realisation that AI changes an organisation’s risk profile in subtle ways. One of my colleagues gave a really good example: you accidentally email a confidential digital document to a colleague who isn’t working on the project. If, in future, a risk manager asks an LLM to surface examples of where corporate risk policies haven’t been followed, the LLM will find it. So, this technology is changing policies, controls, and governance.

What’s different about how PA approaches AI?

We take a human-centred approach, understanding user needs, applying the technology on top, and ensuring that aligns with the AI governance framework, AI blueprint, and the right ethical and responsible practices. There are a lot of complications in the public sector, but the problem area that AI can help with in the public sector specifically is constrained budgets. How do we deliver more for less? And, critically, how do we make sure we get return on investment?

The way we tackle this at PA is by starting with the problem and the people, not the technology.

We focus on the total cost of service over time, not just build costs. We set clear baselines and metrics so we can measure benefits not just at go-live, but months and years after the service becomes operational so we can adjust as we learn. This means we can combine fast, pragmatic delivery with strong governance, surfacing value quickly while managing risk – essential in the public sector. We create sandboxes where we trial wonderful innovative ideas in a safe space, and make sure governance is in place before anything is scaled. This leads nicely to our big focus on responsible AI. We’re not just building solutions, we’re helping clients embed the right guardrails, governance, and ethical frameworks from day one, so the public really trust what’s being built.

What advice you would give to somebody who wanted to get into your line of work?

I have three tips. First, relearn how to learn. The pace of change in tools and technologies is unlike anything we’ve seen before. Staying current means continually updating your skills. Being comfortable with learning, unlearning, and relearning is critical at all layers.

The second tip would be to become AI fluent. AI fluency is critical now to work effectively and drive value from your ideas early.

Linked to that, the third tip is to focus on outcomes, not slideware. Don’t just talk about ideas, build them. Even a small proof of concept can bring an idea to life. With the tools we now have, it’s amazing how quickly people will get on board when they can actually see it.

What are your future goals, professionally but also personally?

I’ve got two young kids, a four-year-old and a two-year-old, and they really shape how I think about my work. I want to help be part of building the future of digital services for the UK government in a way that genuinely opens up access to world-class education and opportunities for them. At the same time, I’m very conscious of the risks that come with these technologies, especially for children. The world they will grow up in will look very different to the one we know now, and I want to help make the most of the opportunities while balancing risk, particularly with our kids in mind.