The present and future of AI in business: Key insights from the Generative AI Summit

17/04/2025

Grayce recently attended the Generative AI Summit, a two-day conference bringing together the latest AI expertise and businesses wanting to adopt and implement the benefits that generative AI can offer.

The key themes were centred around the continuing risk, ethics and governance conversation, plus the advent of agentic AI. In this blog, we share some of our key findings from the event.

Agentic AI adoption will transform businesses

The summit kicked off with an address by the programme organiser Mhairi Macbride, followed by a panel discussion led by Paul Dongha, Head of Responsible AI & AI Strategy at NatWest Group. Paul’s keynote highlighted the current and ongoing importance of risk mitigation and governance in the context of AI. He touched on the new versions of Gen AI models, pointing out that we seem to hear from vendors and platforms daily about the latest improvements, the advent of AI is extremely rapid. And, amongst the pace of development, is the rise of agentic AI.

In some ways, agentic AI can be viewed as a wraparound on top of generative AI. These models can think deeper and more comprehensively, and show their working, and (this is the crucial change), they are now able to carry out actions autonomously. They can send and reply to emails, make bookings and a host of other administrative business tasks that are relatively simple. As the year progresses, the tasks will get more complex with the potential to further transform business operations and outputs.

Agents as staff, focus on the technical outputs and capabilities, not human skills

It was interesting attend a panel session and hear about agentic AI and how it allows AI LLMs to be considered ‘virtual employees’ with a range of skills and competencies. Some companies already package and sell agentic capability pitched as extensions to your team.’The panel, including Daniel Hulme, Chief AI Officer at WPP, Brian O’Reilly, GM of International at Writer, and Dara L. Sosulski, Head of AI and Model Management at HSBC, discussed this at length.

They talked about agents as Staff but the discussion was firmly focused on the technical capabilities agents have rather than trying to suggest that Agents should have employment rights or develop human skills. It posed an interesting question around the use of agents long-term and the impact on businesses.

New AI capabilities are impressive, but they need the checks and balances to ensure businesses aren’t putting themselves at risk

Governance was a theme that was threaded through most of the talks on the day. Steve Holyer, Data Platform Leader EMEA of Informatica, reinforced the themes of risk, governance and accountability. Their product, Claire GPT was extremely competent. He showed a demo that looked initially like their GPT product had fetched a database table back for review, which was impressive alone because it was able to find the right table based only on an English-language (non-SQL) prompt.

But, on closer inspection, we saw that it hadn’t retrieved a table; it had actually written a complex SQL query against the Informatica Data Model that pulled out exactly the right information formatted in exactly the right way. Demos are carefully prepared and scripted, but the general consensus in the room was that this was a more sophisticated tool with new capabilities.

While this was technically competent, it was clear that expertise is still needed across the business to ensure accuracy, and to also govern the way that this technology could be used.

With so much more to cover, we’ll share some more of the key themes, including Graph Databases and how they can work with AI to model the real world in a way that SQL databases can’t.

To understand how Grayce Analysts can support your AI adoption journey, speak to us today.

    .