Over the past year, I have worked with teams who are excited by artificial intelligence but uncertain about what it really means for them. Some worry they need to be technical experts to make sense of it. Others see the potential but are unsure where to start.
This has led the team at Hippo to explore how capability can be built in ways that feel practical and relevant to the work people are already doing. Rather than focusing on creating experts, the emphasis has been on giving teams space to experiment, ask questions and see how AI could support their priorities.
That approach shaped our recent session at DigiGov Expo, Upskill, Uplift, Empower: Building AI Capability in Government at Scale, where we shared how hands-on learning programmes such as AI Week and Policy Impact Workshops have helped public servants build confidence and curiosity around AI.
From this work, a framework has started to take shape — one designed to guide learning in a way that is accessible, responsible and grounded in real outcomes.
Shaping a gradual path into AI
When we first began working with government teams on AI, the same barriers came up again and again. People were curious but hesitant. They wanted to understand the technology but often felt it was out of reach, too technical or too fast-moving to grasp. At the same time there was an appetite to see how it could help with real priorities, provided it was done responsibly.
That combination of curiosity and caution shaped the way the framework developed. It needed to start with the basics, then move into live contexts and finally create room for safe experimentation. Rather than one-off training or theory, it had to be a gradual process that built confidence step by step.
This is how our AI Week upskilling programme came about, a week of hands-on learning designed in partnership with a government department, with the goal of demystifying the fundamentals, bringing people together and showing the potential for AI tools to sit alongside existing ways of working. Through talks, demonstrations and case studies, teams were able to see the technology in action rather than only in theory. For many, it was the first time they could connect what they had read or heard about AI to the realities of their own work.
From there, Policy Impact Workshops allowed teams to take the next step. These sessions looked at live priorities and explored how AI might influence them in practice. By working with real policies and delivery challenges, the conversations shifted from abstract potential to concrete application. It helped people see not only what AI could do but what it should do in the context of their responsibilities.
The next phase, Design in a Day, is currently in development. It will build on the foundations of AI Week and the Policy Impact Workshops by using rapid prototyping to help teams test ideas quickly and explore how AI could shape future services or policies. While this stage has not yet been delivered as part of the programme, it draws on an approach we have used successfully in other contexts. That experience gives us confidence that it will create the same environment for experimentation, a space where early concepts can be explored and start to turn into practice.
Why this way of learning mattered
Two elements have been central to how the framework has taken shape.
The first was placing responsibility at the centre. Issues like fairness, bias, job impact and environmental cost are not footnotes to innovation, they are central to it. By bringing those considerations into the conversation from the start, teams were able to see responsibility as a practical foundation rather than an abstract principle.
The second was keeping the approach hands-on. Real-world case studies, prototyping sessions and open discussion created an environment where people felt able to explore without fear of failure. It was not about producing polished solutions but about building confidence through experience and collaboration.
Together, these elements, responsibility at the centre and space to experiment, are what made the framework effective and what allow it to adapt across different contexts.
A framework that travels across sectors
What has been striking is how transferable this style of exploration can be. Although our early work was with government, the same questions are being asked in almost every sector. In financial services, there is curiosity about how AI might support decision-making but caution about fairness, bias and regulatory risk. In healthcare, teams are exploring whether automation could relieve pressure on staff, but only if patient trust and accountability are protected. Across the wider private sector, leaders want to understand where efficiency gains can be made without losing sight of the impact on jobs, skills and sustainability.
In all of these examples, the barriers are familiar. People worry that they do not have the technical expertise to engage. They are unsure how to connect abstract ideas to the work in front of them. They want to move forward but only if it feels responsible and aligned to their organisation’s values.
The framework is effective across these contexts because it is not tied to a single technology or policy area. It is built around people, their confidence, their ability to question and their need for space to experiment. By beginning with accessible fundamentals, moving into live application and then opening the door to prototyping, the same approach helps teams in any organisation move from uncertainty to practical insights.
What makes it powerful is that it does not promise complete solutions overnight. Instead, it gives organisations a way to explore, test and learn with care. That balance of curiosity and responsibility is as important in a boardroom as it is in a policy team. Wherever AI is being considered, the challenge is the same: how do we build capability that lasts and trust that endures?
Lessons that will last
Working through this framework has underlined a few lessons that feel just as relevant outside of government as within it. The first is that confidence comes before capability. People need to feel able to engage with the basics before they can apply them responsibly. Without that foundation, the risk is that AI either feels inaccessible or is adopted without enough questioning.
The second lesson is that ethics cannot be an afterthought. Whether it is bias in data, the impact on jobs or the environmental cost of large-scale computing, responsibility has to sit at the centre from the start. Building capability is as much about strengthening the ability to challenge and question as it is about learning how to use new tools.
Capability grows best when it is hands-on and collaborative. Short bursts of theory are useful but what makes the difference is when people are given the time and space to apply ideas to their own work. When they can share experiences, try things out and learn together, the results are far more lasting.
These lessons have shaped how we think about capability-building more broadly. They remind us that technology itself is not the differentiator, it is how people understand it, use it and embed it responsibly. That is what keeps innovation human and what ensures it delivers outcomes that endure.
Look ahead to Think Digital Government
In November, I’ll be speaking more about this model and other capability building projects on the Think Digital Government ‘Digital Skills and Capability Panel.’ More than ever, it’s important for suppliers to embrace their responsibility to knowledge share, invest in long-term capability building, and embed digital skills directly into delivery teams.
