Costi Perricos
Global Generative AI Leader, Deloitte
Co-founder, Deloitte AI Institute
About
I’m Global Generative AI Leader at Deloitte, where I’m a partner based in the UK. I’ve spent more than 30 years working in data, technology and transformation, around 15 of them focused on government and public services. I co-founded the Deloitte AI Institute and previously led Deloitte’s AI and data practice in the UK and globally.
My work in artificial intelligence began with a PhD in information engineering at the University of Cambridge, undertaken when no one was particularly interested in AI or data. Today I spend my time with business and government leaders on putting AI to work in practice: redesigning the work itself rather than bolting technology onto old processes, and keeping people at the centre of the change. I speak regularly at forums such as the AI for Good Global Summit in Geneva and during the World Economic Forum Annual Meeting in Davos, and I write the LinkedIn newsletter Actioning AI.
What I focus on
- Redesigning work around AI, rather than adding it on to existing processes.
- Measuring the real value of AI, beyond cappuccino ROI, including what compute and tokens actually cost.
- The human side of adoption: bringing both enthusiasts and sceptics into pilots and scaling.
- Governing AI agents as a digital workforce, with accountability resting with people.
- Trusted data as the foundation for any AI that works.
- Hiring and training young people so tomorrow’s experts still have a way in.
Cappuccino ROI
Cappuccino ROI is a phrase I coined for what happens when you add AI to each step of an existing process. Everyone saves ten minutes here and ten minutes there, just enough to go and grab a cappuccino. People feel more productive, and they probably are, but the savings are scattered across the organisation and never really hit the bottom line—a point I discussed with ISMG at HumanX Amsterdam and in this note on LinkedIn. The organisations seeing the biggest gains do something different: they redesign the work first, with AI at the centre, aiming it at growth, new business models and customer acquisition so the value can actually be measured and monetised.
Speaking & media
- The Era of the AI Workforce — Track Record podcast by TRAMLINES, episode 9
- Doctors and teachers to benefit from AI but safeguards needed to protect sensitive information — The National
- Redesigning government services is key to unlocking AI’s potential in the Gulf — Forbes Middle East
- Why staff using ChatGPT, Claude and Gemini for work could create problems for employers — The National
- Why ‘cappuccino ROI’ keeps AI gains from the bottom line — ISMG, HumanX Amsterdam
- Lessons learnt from agentic AI — AI for Good Global Summit, Geneva
- Businesses face up to budget-busting AI bills — Financial Times
- Davos Discussions: The Human Coworker — Workiva, The Pre-Read
- Deloitte GenAI Summit 2026, São Paulo
- Harnessing AI for Good: emerging AI trends and global responses — unDavos
- ‘Our goal is to create a culture that doesn’t stigmatise the use of GenAI tools’ — The Hindu
- AI: The biggest tech infrastructure buildout in history — AI House Davos
Writing
I write Actioning AI on LinkedIn to translate developments in AI into practical lessons for business and government leaders. Recent editions:
- From individual assets to connected systems: AI’s next infrastructure opportunity
- Four industry signals for the next phase of AI adoption
- AI at work: Insights from Gen Zs and Millennials
- From tool to teammate: AI and the reimagination of work, governance and enterprise economics
Also: How agentic, physical and sovereign AI are rewriting the rules of enterprise innovation (World Economic Forum, January 2026).
Elsewhere
For speaking or media enquiries, the best way to reach me is via LinkedIn.