13th November 2026

The people behind AI: How Hannah brought Minerva to life

Hannah Cameron Sustainability Engagement Specialist, Sustainable Business team

Hi I’m Hannah! I work with teams across our markets to help tell the story of all the amazing sustainability work happening across Vodafone, through campaigns, events and communications. My focus is on engaging our employees, but I also help share those stories with wider audiences.  I joined Vodafone two years ago on the Group Business graduate scheme. My first placement was in product management for Microsoft 365, where I started getting to know Copilot and how it could be useful in everyday work. I then moved into customer experience in Group Commercial, before joining Sustainable Business after finishing the scheme. I'd always dreamt of working in sustainability, so this was a really exciting move for me.

The inspiration behind Minerva

The idea came from a team brainstorming session about how we could use AI to work more efficiently. One idea that came out of the discussion was using AI to bring together sustainability case studies and evidence from across our markets, which were spread across websites, annual reports, ESG snapshots, and thought leadership reports. When you’re writing communications or preparing speaker notes on a topic like digital inclusion or climate resilience, you want to know what Vodafone is doing and which examples will be relevant to your audience. But without one place to look, you have to spend time searching through different sources and piecing the information together yourself. That was the problem I wanted Minerva to help solve: making it easier for colleagues to find relevant stories and evidence from across Vodafone, so they can spend less time searching and more time working out what they want to say.

Creation process and solving a problem

Building Minerva was much easier than I’d expected. Having worked with Copilot before helped, but creating an agent was new to me. Tools like Agent Builder made it feel much more manageable. The biggest lesson was how important testing is. Getting an agent up and running is one thing; checking whether its answers are accurate and useful is another. I learnt that choosing its knowledge sources carefully could help with that. Minerva uses the documents and websites I give it, rather than searching the whole internet, which helps keep its answers focused on Vodafone’s own information.  I also learnt how important it is to be clear about its limitations. I put together a guidance note explaining its sources, guardrails, and what it can and can't do, so I have something clear to share with colleagues when questions come up.

Minerva's purpose and potential next steps

Minerva was originally created to help us work more efficiently, but ultimately I want it to help more of our markets’ work get noticed and talked about. From digital healthcare services in Egypt to connected sensors helping manage tree watering in Germany, there are some brilliant examples that colleagues might otherwise miss.  I hope that also helps teams learn from what’s happening elsewhere in Vodafone. An example from one market could give colleagues in another an idea to explore. That’s where I think Minerva could support our sustainability agenda beyond communications. I plan to keep evolving it too. One idea is to introduce different audience personas, such as a policymaker, customer or employee, so colleagues could explore how their communications might come across to each and refine their messages to help them land.

Key takeaways and advice to anyone wanting to experiment with AI, but doesn’t know where to start!

It’s made me more confident about trying ideas myself, rather than assuming building something with AI would be too technical. I work in communications and engagement, not technology, but I was still able to build something useful for my team. One of the nicest parts has been hearing from colleagues who’ve enjoyed using Minerva and now want to try something similar for their own business area. I hadn’t expected that response, and it’s been exciting to see people thinking about what they could create. My advice would be to just go for it! Think of a specific task you’d like to make easier, open Agent Builder and describe what you want your agent to help with. You might find it’s much easier to get started than you thought.