Track Record by Tramlines

Costi Perricos, Deloitte's Global Generative AI Business Leader, joins Albert Azis-Clauson to discuss how AI is changing the way businesses are organised, how people work and develop their expertise, and the economics of professional services.Drawing on his journey from rehabilitation engineering and academia into consulting, Costi explains why human judgement remains central as AI becomes more capable. They explore what leaders need to understand themselves, how junior people can gain experience alongside agents, and who takes responsibility for managing a digital workforce.Key Moments:00:00 Introduction01:18 Engineering, academia and the route into consulting09:32 Leading GenAI at Deloitte11:44 AI education, fearers and reverers15:39 Why human judgement still matters19:36 Developing the next generation of experts24:47 Leadership and managing a digital workforce28:04 Human creativity and the limits of AI30:41 Thinking of AI agents as teammates32:24 What small and large businesses can learn from each other37:24 The changing economics of professional services41:39 Ownership, agent architecture and the agentic harness48:53 Startups, large firms and business transformation53:33 Practical advice for leaders adopting AI55:06 Honest conversations about failure58:33 Where Costi would invest £100 million1:00:50 Costi’s advice to Albert

Craig Donaldson grew up during the miners' strike, ran 84 bank branches in East Africa at 25, and went on to be the founding CEO of Metro Bank. Now he is the Executive Chairman of Tramlines. He tells Albert what pubs taught him about people, why he started a bank when everyone else was cutting costs, and why domain experts with the right support deserve a chance even if they never become unicorns.Key Moments:00:00 Introduction01:44 A background to Craig's early life03:10 From technologist to banker04:13 Pubs, tills and "business is people"05:08 84 branches in East Africa at 2506:14 Why people took a chance on him07:37 RBS and the global financial crisis08:08 Learning from the hard times09:08 Intellect, interpersonal skills and hard work10:15 Resilience and boss watching11:56 From doing to leading13:50 Dave Craggs and the curve on the graph16:24 Energy from someone else closing the deal17:29 Why start a bank?18:17 The zig and the zag19:38 A call from Vernon Hill20:52 The first new high street bank in over 100 years22:23 Printing cards in store24:04 Create fans24:51 Fail fast25:31 Seven years to IPO27:17 Proof points and stage gates28:48 Statements, post boxes and mirrors30:41 What Metro taught him32:21 Human-led businesses in an AI world34:24 Leaving Metro Bank38:37 Meeting Albert and telling him everything41:37 How Tramlines started44:33 Discipline, and AI as a tool46:17 What excites Craig now48:54 Why experienced people come to Tramlines51:44 The £100m question52:52 Advice for Albert54:52 Close

Daniel Lanyon started out as a journalist in Istanbul in 2010, joined The Times' foreign desk as a freelancer, and followed the money into financial journalism. He went on to help build AltFi, a publication covering fintech, from around 2016 until it was bought by a family office. Along the way he met the early teams at Revolut and Monzo, when Revolut had just one product and a team of 10 or 20.He is now Managing Director of Funds at Tramlines. In this episode of Track Record, Albert asks how that career shaped the way Daniel now thinks about founders, investing and AI. They cover how his years as a reporter taught Daniel to judge founders, why he thinks domain experts could overtake the 'Stanford dropout', and how to keep a clear head about AI.Key Moments:00:00 Introduction01:43 Daniel's role at Tramlines02:40 Istanbul, Tintin and becoming a journalist05:28 The Times, The Grocer and following the money06:31 From financial markets to fintech08:19 Why UK Fintech took off09:15 Branches: from moat to millstone10:08 Revolut, Monzo and Starling in the early days11:50 Three tracks: a business, a cycle and investors13:28 A football hedge fund and a buoy on a beach14:17 AltFi and learning to sell16:49 Inside versus outside17:18 The 2021 boom and the downturn18:49 Deciding to become an investor20:12 Meeting Albert21:23 VC as a five-year-old's football match22:24 Where the value lies now23:31 Why the best founders are domain experts25:12 When the big names caught on25:53 "I will never have an app-only bank"27:25 Scaling domain expertise with AI28:52 Going earlier: meeting founders before they quit31:10 Consultancy as a way in32:59 What AI changes for investors35:09 Everyone at Tramlines was at the front36:19 Keeping two sets of books38:29 Analog in an AI world39:13 How to get to the front of AI40:40 Think, plan, do41:40 Three ways to use two sets of books43:06 The £100m question45:21 Advice for Albert46:24 Closing remarks

Francesca Hodgson has built a venture-backed fintech, coached hundreds of founders in the last few months alone, and runs the SPARK incubator at King's College, Cambridge. She talks to Albert about why founders should start with the problem rather than the idea, what raising money really cost her, and why AI makes it more important, not less, to know what you are looking for.Key Moments:00:00 Introduction01:23 Francesca's background02:57 Relationships, trust and discovering your purpose05:43 How Fran coaches: starting with the problem07:56 The trouble with the big idea10:31 Vision, the minimum viable business and the beachhead11:25 Build with, not for12:55 The seven validation gateways15:11 Raising 15 million with 5% of the equity16:32 Revenue, the funding people overlook17:58 There's always a cost19:27 Women founders and the funding gap24:40 Why venture capital is a hammer25:58 The loneliness of a venture-backed founder27:45 Separating governance from management29:01 Psychological safety and leadership styles31:03 Shared ownership and collaboration33:41 AI amplifies what you put in36:15 The false sense of security39:02 Domain expertise plus AI expertise40:32 Thought processing, AI slop and checking the output42:43 Where the opportunity is for founders44:13 The barrier to innovation is now the ability to speak45:38 Where Francesca would invest £100m47:31 Closing remarks

As AI is makes software faster and cheaper to build, it changes what investors should look for, how companies sell, how enterprises adopt technology and where defensibility comes from.In this episode of Track Record, Albert Azis-Clauson talks to Edward Keelan, a partner at Octopus Ventures, about the shift from traditional SaaS metrics to a world where pricing, retention and outcomes are still being worked out. They discuss why revenue and the speed of revenue growth matter earlier than many founders expect, why enterprise AI adoption is closer to workflow transformation than buying software off the shelf, and why customer relationships are becoming more consulting-like.Ed also shares his experience of moving from Rolls-Royce and a startup into venture, and they discuss socio-economic access to capital, education, personal relationships and the risk of relationship-led investing reproducing existing networks.The conversation ends with a shared challenge to early-stage VC: as technology becomes easier to build, domain expertise, customer knowledge and the ability to deliver in a specific vertical may become the more durable advantage.Key Moments:00:00 Introduction: Ed Keelan on AI and investment02:29 From Rolls-Royce and startups to Octopus05:24 Why entrepreneurs and small companies matter09:39 AI outreach and the return of personal relationships12:53 Socio-economic access, education and aspiration20:29 How AI is changing growth-stage investing23:54 Pricing, product-market fit and unsettled metrics27:04 Enterprise adoption and transformation30:16 Where AI creates value and changes work33:57 Customer relationships, services and workflow design36:54 Domain expertise and the changing sales team41:09 Vertical software, defensibility and scale44:28 What investors should measure now46:58 The future of VC and the role of operating partners49:03 Octopus Ventures and closing thoughts

In this conversation, Albert Azis-Clauson speaks with network scientist and Director of Networks at Tramlines, Erica Young.Erica explains how a personal project to map the people she met after moving to London led her into network science, and why relationships create value that compounds over time.We examine how AI could change relationship management, why trust becomes more important as products become easier to build and online content harder to verify, and why traditional networking events do not work for everyone.The conversation explores how to approach new relationships without making them transactional, how a case of a junior job offer led to a board seat, and how different network shapes support discovery, coordination, and resilience. Erica closes by explaining why she would invest £100m in social capital and why one relationship-building habit, sustained for 10 years, can have an extraordinary effect.Key moments:00:00 Introduction01:25 From a personal project to network science05:33 Why network value compounds07:39 How VC incentives weaken collective networks12:11 How AI can transform relationship management15:49 Why relationships become more defensible with AI18:25 Networking beyond extrovert-designed events19:47 Why investment-or-nothing is too narrow21:31 How a junior job opportunity led to a board seat22:48 Explore versus exploit27:43 The daisy and the ball of yarn32:29 How to assess your network34:34 How a chance Airbnb encounter changed Erica's career39:38 Why the teenagers were the AI experts in the room41:14 How to cultivate and maintain relationships45:39 Why some work friendships disappear48:42 Where Erica would invest £100m50:03 One habit sustained for 10 years