AI is changing what’s possible in digital delivery. But not everything that’s possible is worth doing. It’s easy to fall into the trap of thinking “we’ll just add AI to our service and it’ll fix it”. Or that an entirely novel AI solution is the right thing. Getting the most from AI is often actually about identifying what it can achieve better than any other option. And if it can’t, then not using it for the sake of it.
At Marvell, we're increasingly helping organisations identify where AI can make existing services more efficient, more sustainable and better for users. Without losing sight of the human expertise that underpins good design and delivery. Whatever the context, our approach is the same: start with the problem to solve, understand the users and then apply AI where it adds measurable value.
"Public sector organisations should only use AI technologies when relevant, appropriate and proportionate."
AI isn’t a standalone product. It needs context, infrastructure, clear user journeys. There’s not even a single definition of what AI is. It’s a multitudinous range of different technologies, with wildly differing strengths, weaknesses and appropriate applications.
The government’s AI playbook advises that public sector organisations should only use AI technologies when relevant, appropriate and proportionate. And that’s something we’re committed to when we deliver public services.
We help organisations identify and use the most suitable and sustainable option for their needs. There are important environmental considerations when it comes to using AI technologies, as they can be energy and water intensive. So we make this clear to organisations, and help them identify how they can offset these effects. We also aim to provide robust testing and continuous monitoring and improvement. Both vital for trustworthy and ethical AI adoption.
So what might all this actually look like? It might be using AI to improve data quality or analyse user feedback at scale. It might be exploring where pattern recognition could make services faster and more accessible. Or it might be using natural language processing (NLP) to improve automated interactions with users. Whatever it is, we’ll use AI technologies as part of a wider suite of tools to enhance outcomes, reduce manual effort and unlock new insights.
"We've adopted a range of AI-led tools ... but only ever use these methods alongside our own knowledge and expertise."
We’ve also adopted a range of AI-led tools into our everyday practice. Enhancing and facilitating the great, thoughtful work of our multi-disciplinary experts. But we only ever use these methods alongside our own knowledge and expertise.
Whether that’s our UX and content designers using generative AI to help speed up high-fidelity prototyping or running readability checks to verify the reading age of content. Or our user researchers using AI to help with desk research or transcribing and identifying trends in large volumes of primary research, such as user interviews.
Or our software developers incorporating AI tools into their build and test workflows. Using large language models (LLMs) to help with first-pass code reviews or conducting unit tests (testing individual, isolated components of software). As with all our work, it’s always about using the best, most efficient tool for the job.
We’re part of many government procurement frameworks, including the AI dynamic purchasing system. The system allows the public sector to understand and see the benefits of any relevant AI technologies. That could include helping with AI discovery, building AI-facilitated software applications, using machine learning to help with data analytics or developing augmented decision-making tools.
If you're a public sector organisation, we can deliver your digital service. Future-proof, easy to use and efficient to run. And making the most of AI, when it’s the right solution. Find out more about how to buy our services.
Whether you’re ready to start your project now or you just want to talk things through, we’d love to hear from you.