Private Equity AI Strategy
Breakfast Briefing #6
8-11am | Tuesday 3 March 2026 | London
Agentic AI:
What's Real, What's hype, What to do next
Our breakfast briefings are carefully crafted to cut through the hype, provide accessible and actionable insights for Private Equity professionals. Dive deep into real case studies, engage in hands-on workshops, and develop valuable connections with experts. ​
Our Partners
Visionary Partner
Activation Partner
Innovation Partners
Speakers
.png)
Delegates
Panel
Agenda
Agentic AI: What It Is, Where It Works, How It Scales
Moderator: Graeme Cox, CEO & Founder, Attercop
Speakers:
John Gunn, Lead Data Science & AI, IK Partners
Tom Pearson, Head of Data, BGF
​​
We’ll start the morning by bringing clarity to one of the most widely used and often misunderstood terms in AI today: agentic AI. The objective is to give everyone a shared, practical understanding of what people mean when they say “agents”, what’s genuinely new, and what is simply a re-labelling of existing automation or AI capabilities. This is designed to help PE leaders ask better questions, assess opportunities more confidently, and avoid getting pulled into hype-led conversations.
​
From that shared starting point, we’ll move into a panel discussion with three AI leaders from private equity firms who are working hands-on with AI across their organisations and portfolios. They’ll share what they are seeing in practice: where momentum is building, where progress is proving harder, and where agentic approaches are starting to show promise. The panel will also explore the practical “how” at a business level - how initiatives are being shaped, supported, and governed; how teams are thinking about partners and capability-building; and what it takes to move from early experiments to something repeatable across a portfolio, without getting lost in the technical weeds.​
Case study
From tech-led growth to agentic ambition
Moderator: Isabelle Fayad, Senior Channel Partner, Ramp
Speaker: Mike Jones, CTO, loveholidays
In this fireside chat, Mike Jones will share the story of how a PE-backed, high-growth business put technology at the centre of its value-creation plan - building the foundations that helped loveholidays outpace competitors, and positioning the company to take real advantage of the AI wave.
​
We’ll cover:
-
The foundation that enabled speed and scale: the practical infrastructure, engineering discipline, and data capability investments that made the business more resilient and made AI adoption easier (and safer)
-
How they’ve applied AI so far: where AI is already creating tangible impact across teams and workflows, and what’s been harder than expected
-
From AI use-cases to agentic ambition: why they see agentic AI as a meaningful next step, what “agentic” means in their context, and where they think the near-term value is
-
What they’ve done to date on agents: early experiments, learnings, and the conditions they believe need to be in place before scaling
-
Pragmatic execution lens: how they think about controls and guardrails, integration into real processes, and how to judge ROI signals versus “innovation theatre”
​
Expect an operator’s perspective on sequencing: getting the basics right, delivering value with AI today, and laying credible groundwork for agentic capability tomorrow.
Workshop
Build an AI agent for a real PE workflow
Moderator: Tom Hewitson, Chief AI Officer, General Purpose
​
To close the morning, we’ll shift from listening to doing. In this fast-paced, table-based workshop, delegates will work with their peers to design and build a simple AI agent in Claude Code for a private equity workflow — all within a 50-minute sprint. The goal isn’t to create a “perfect” production system; it’s to give everyone a practical, shared understanding of what agentic AI actually is, how it works, and where it can (and can’t) create value. What's included:
​
-
A quick orientation: what we mean by “agentic”, what Claude Code is doing under the hood, and the core building blocks of an agent (task, steps, tools, constraints)
-
Workflow selection: pick a PE use case (e.g., diligence support, portfolio performance insight, IC memo drafting, value creation tracking, vendor screening)
-
Define the agent spec: objective, inputs/outputs, success criteria, escalation points, and human-in-the-loop controls
-
Build + test in Claude Code: prompt structure, tool calls, memory/context, and iteration cycles to improve reliability
-
Guardrails and risk: handling sensitive data, approval checkpoints, auditability, and avoiding “confident wrong” outputs
-
Debrief: what worked, what broke, where ROI might show up first, and what it would take to operationalise safely
​
Tom brings a training-led, pragmatic approach - General Purpose focuses on hands-on enablement and safe, practical adoption, with an emphasis on measurable outcomes.