Agents in a Day: Turning AI Curiosity into Capability

Learning by building

Delivered at Microsoft’s Paddington offices, attendees spent the day actively building agents using Microsoft Copilot Studio. Through guided labs, participants designed conversations, connected knowledge sources, and explored how agents can automate and augment real business processes.

By the end of the day, every attendee had built a working agent and gained a clearer understanding of what agentic AI looks like in practice.

Real-world impact: Defra and the Environment Agency

A highlight of the day was the conversation with Defra and the Environment Agency, which brought agentic AI firmly into a real public‑sector context. Rather than hypothetical use cases, attendees heard how AI-enabled agents can support regulatory services, respond to ministerial priorities, and improve access to complex guidance and data.

This session demonstrated how agents can help teams navigate large volumes of policy, legislation, and operational information, safely, securely, and with the right governance in place.

From policy complexity to practical assistance

The Defra use case showed how agents can act as assistive colleagues, helping users find answers faster, interpret guidance more consistently, and reduce manual effort across regulatory workflows. Importantly, the discussion didn’t shy away from the realities of public‑sector delivery, including data sensitivity, compliance, and trust.

It reinforced a key message: AI adoption in government isn’t about replacing expertise, but about amplifying it.

Built with governance in mind

Throughout the workshop, and particularly in the Defra discussion, there was a strong emphasis on security, governance, and responsible AI. Attendees explored how Copilot Studio aligns with existing Microsoft 365 controls, ensuring that agents surface only information users are entitled to see.

This was especially relevant for public‑sector organisations, where confidence in data handling and auditability is critical to adoption.

Designed for technical decision‑makers

The workshop was aimed at solution architects, IT leads, and data and AI professionals who need to make informed decisions about AI adoption. The content was deliberately technical and practical, while remaining accessible to those earlier in their agent journey.

This balance meant the day worked equally well for organisations just getting started and those already experimenting with Copilot and AI tooling.

Strengthening the Microsoft–Hitachi partnership

Running Agents in a Day, in collaboration with Microsoft, and featuring real customer perspectives like DEFRA, highlighted the strong Microsoft–Hitachi partnership. By combining Microsoft’s platform strengths with Hitachi Solutions’ delivery expertise, the session stayed grounded in what works in real environments.

It also highlighted how collaboration accelerates confidence, not just innovation.

What attendees took away

By the end of the day, participants left with:

• A working agent built in Copilot Studio
• A clearer understanding of agent maturity, from conversational to autonomous
• Practical insight into public‑sector and regulatory use cases
• Confidence to continue experimenting back in their own environments
• Clear next steps for discovery, governance, and scale

AI enablement starts with experience

The success of Agents in a Day reinforced a simple truth: the fastest way to understand AI is to build with it. Hands-on experience cuts through uncertainty and turns abstract concepts into practical capability.

With real examples like Defra demonstrating what’s possible today, organisations can move forward with confidence, turning AI ambition into real outcomes.