London Tech Execs #17 - Bringing AI to the Edge Without the Operational Overhead

As AI moves into production, where workloads run is becoming just as important as how they're built.

For many organisations, running everything centrally in the cloud is no longer practical. Real-time performance, data sovereignty, cost and availability requirements are pushing AI workloads closer to where data is generated and used, from retail stores and hospitality environments to manufacturing sites.

But deploying AI at the edge introduces a different set of challenges. How do you manage infrastructure across hundreds or thousands of locations? How do you deal with different hardware, unreliable connectivity and the operational complexity of keeping everything running?

This roundtable brings together technology leaders to explore the practical challenges of running AI at the edge, the architectural decisions involved, and how organisations can scale without creating an operational burden.

Natasha Wright, Senior Director Cloud Alliances at Spectro Cloud, will lead the discussion, drawing on her experience in cloud-native infrastructure and real-world examples of AI being deployed across distributed environments.

We'll cover:

1. When does running AI at the edge actually make sense? What are the real drivers behind moving workloads away from centralised cloud environments?

2. What are the biggest challenges of deploying AI across hundreds or thousands of locations? How do you handle different hardware, connectivity limitations and infrastructure requirements?

3. How do you manage edge AI infrastructure at scale without increasing operational overhead? Where can Kubernetes, automation and decentralised architectures make a difference?

4. How do you balance performance, cost, security and data sovereignty? What trade-offs are organisations making when deciding where AI workloads should run?

5. What does successful edge AI look like in practice? What can we learn from real-world deployments in retail, hospitality and manufacturing, and what still needs to improve?

Agenda:

• 5:30pm to 6:00pm – Arrival and networking

• 6:00pm to 6:15pm – Introductions

• 6:15pm to 7:00pm – Session part one

• 7:00pm to 7:10pm – Break

• 7:10pm to 7:45pm – Session part two and discussion

• 7:45pm to 8:30pm – Networking and wrap-up