This autumn, the AI Trust Forum brought together industry leaders, practitioners and thought leaders in London and New York. Their shared aim was to explore what it really takes to make AI trustworthy, governed and genuinely useful. In this episode, host Ioana talks with two forum co-hosts about what they heard in the room and what it means for organisations adopting agentic AI.
0:00 Lucy Bourne, Oaka Studio (London)
Lucy Bourne of Oaka Studio helps Microsoft partners strengthen their relationship with Microsoft, sharpen their go-to-market and scale delivery. She shares three trends reshaping the partner ecosystem:
Business applications partners moved first.
Managed service providers are now catching up fast.
Partners from competing automation platforms are turning to Microsoft.
Lucy explains why governance now features in almost every conversation she has. She also revisits a memorable analogy from the London forum: aspirin was in use for decades before anyone fully understood how it worked. Finally, she reflects on AI overwhelm, professional identity and why tolerance to ambiguity matters for people as much as for organisations.
31:40 Arnav Gupta, Damco Solutions (New York)
Arnav Gupta of Damco Solutions works with enterprises across industries, including many highly regulated. He shares a practical guide to taking AI agents into production, drawn from his session at the New York forum, which was hosted at Microsoft’s offices. He makes the case for multiplayer AI agents, which have their own goal and identity and work alongside whole teams in tools like Microsoft Teams. He also explains why connecting agents to core systems of record is where many Copilot initiatives fall short. On AI slop, he argues that the fix is clear accountability: every output belongs to a human, not to the AI that helped produce it.
Both conversations close on the same theme: technology choices matter, but the bigger question is what humans should be doing, and what kind of society we want to build.
Key takeaways
Governance is no longer optional. Line-of-business, managed services and security partners are all now offering agentic AI, and all of them need a credible governance story.
Waiting carries a cost. The risk of moving too slowly can outweigh the risk of not fully understanding the technology yet. Good governance means balancing the two.
Multiplayer agents unlock real transformation. Agents with a defined name, role and goal, working across a team, deliver more than personal productivity tools alone.
Agents need access to your systems of record. Opening up platforms like Dynamics and Dataverse to agents, with the right controls, is where much of the value lies.
Adoption is a mental model problem. People engage more readily when they treat an AI agent as a colleague they delegate to, rather than as a tool.
Accountability beats AI slop. A policy that makes every person responsible for their output keeps quality high, whoever or whatever drafted it.
Aptitude may matter more than skills. As agents take on more of the work, adaptability, creativity and the ability to finish what you start become more valuable.
In-person connection still matters. As AI handles more of the routine work, real-life human interaction becomes more important, not less.
Enjoy
Ioana Tanase 👉 LinkedIn
Lucy Bourne👉 LinkedIn
Arnav Gupta👉 LinkedIn
