What I believe

Five positions I will defend in public.

“Not because AI will change everything — that is too vague to be useful. The real shift is structural.”

Commerce

Conversational commerce will be bigger than e-commerce.

For thirty years online shopping has followed the same pattern: visit, browse, add to cart, check out. We optimised it; we never reinvented it. The next shift is not better websites. It is a customer saying what they need in a messaging app and an agent finding it, buying it, delivering it and handling support. The technology exists. What is missing is trust, payments, and businesses willing to redesign the journey around the conversation.

Read the note on LinkedIn →

Strategy

AI is infrastructure, not a feature.

Most enterprises ask “what AI features should we add?” and build smarter versions of what they already have. The companies that will define their categories in 2030 ask a harder question: if AI can handle the reasoning that used to need human judgment, what does the rest of the product need to be? That question is uncomfortable because it puts years of decisions back on the table. It is also the only one worth asking.

Read the note on LinkedIn →

Agentic AI

A chatbot answers questions. An agent completes tasks.

An agent queries the CRM, triggers the API, updates the record, sends the follow-up and escalates on a condition — across many steps, with no human in between. One message, one intent, full resolution. That is a different customer experience and a different operating model, and it only works on platforms built event-driven and API-first. Everyone else will bolt it on, and it will show.

Read the note on LinkedIn →

Telecom

Telcos had a choice in the 2010s. Most chose defence.

When WhatsApp and iMessage displaced SMS, operators held assets nobody could copy: trusted identity, billing relationships, regulatory standing, carrier-grade infrastructure, the SIM itself. They could build platform value or protect SMS revenue. Most protected. The lesson reaches far beyond telecom: when the interaction model changes, the real risk is not lost revenue — it is losing the position from which future revenue is created.

Read the note on LinkedIn →

2036

Ten years from now, business-to-customer communication will be unrecognisable.

Apps and websites matter less; conversations through messaging, voice and assistants matter more. Identity gets simpler and more trusted, with fewer disconnected logins. Agents handle the routine — questions, purchases, service issues — and the winners are the companies that stop treating communication as a cost line and start treating it as the core of the customer experience. I do not know exactly what 2036 looks like. I am confident it is faster, simpler, more personal and more trusted than today.

Read the note on LinkedIn →

Newsletter · monthly

The Next Conversation

One essay a month on how businesses will talk to customers in 2030 — written by someone building the infrastructure, not reporting on it. No hype, no summaries of other people's takes.

Opens your email app with a subscription request. Or subscribe on LinkedIn.