How I see it
How I see it
Five things I believe about communications and content, and where each one came from.
Nobody arrives at a position like these from reading. Each of them cost something to learn.
If narrative precedes strategy, then the brief precedes a plan.
The usual order is backwards. A team is given an objective, builds a plan to hit it, and writes the narrative afterwards to wrap around what has already been decided. The narrative becomes packaging.
Before anything was drafted for the divestiture, I took a fifteen-person working group of C-suite, marketing, commercial and client engagement leads through what we were actually trying to achieve – not the deliverables, the outcomes. Avoid negative coverage. Sustain client confidence. Build awareness of a French banking brand in a market that barely knew it. Only once the room agreed on the story did anyone start planning how to tell it.
It took longer at the start. It meant that when the transaction moved faster than the plan, which it did, everyone was making decisions from the same story rather than from the same document.
In practice: A clean break.
I once prepared a chief information officer for a media programme by bringing in a friendly journalist to stress-test him. The feedback was that he would be eaten alive.
The obvious response is more training – rebuild the executive until he matches the profile. I changed my approach instead. Rather than making him into the confident, quotable technology leader the cookie-cutter version of that role demands, we found the profile he actually wanted and could sustain: fewer interviews, better matched, on subjects where his depth was the asset rather than his delivery.
Engagement with his thought leadership grew 250%, and he ended up on stages he would never have reached as a worse version of somebody else.
The lesson generalises. An executive who is being coached towards someone else's personality will perform it badly under pressure, and pressure is the only condition that matters. Your job is to find the version of them that is both authentic and effective, not to close the gap to a template.
The thinking in practice: a keynote I drafted for the executive it was built around.
Authority earns moments. Moments create advocates. Advocates generate the language the engines surface. And the engines reinforce the authority.
The funnel model is a hundred and twenty-eight years old and still the shorthand everyone in the room will reach for. It's useful for deciding what to make and what to measure. It's no longer a description of how people find things.
Authority comes first, because a brand with no standing cannot assert its way into a market – it can only demonstrate. Timing turns authority into impact: activating content at the moments audiences are already engaged and seeking guidance, whether that's the World Economic Forum, SIBOS, Money20/20 or a room of thirty-five pension schemes over lunch. Advocates are the multiplier, because high-NPS customers and genuine communities produce the organic, natural-language content – forum threads, peer recommendations, reviews – that AI discovery engines actually surface.
Then it turns. Which is why nurturing a community is a discovery strategy, not a nice-to-have.
In practice: Patient money.
Large language models are exceptional for research, textual analysis and wireframing. Paired with automation they can increase production efficiency by orders of magnitude. They excel at mechanising solutions when given accurate prompts in specific contexts.
I've twice secured board-level approval to design, procure and implement a GenAI-infused content workflow. Aggregated, they cut time-to-market for external narratives by over 55% – and the point was never the efficiency. It was that the team got back the capacity to make the rich, difficult content the machine can't.
Because what these models lack is empathy and genuine novelty. To lift a value proposition in a crowded market, the narrative has to stand alone and break new ground. Models amalgamate what already exists. Show me an agent that combines audience listening, product specification, brand narrative and brand equity, and from that designs a strategy that deconstructs market friction into a new way of thinking, and I'll add an em dash to my name by deed poll.
Every AI-assisted workflow also needs a human filter – editorial, factual, cultural. A model cannot be held responsible for a published message. Only a person can own it, and that ownership matters most on the day something goes wrong.
Legal wanted the lead message to be renegotiated pricing. Commercial wanted bespoke materials for every client. Product wanted to hand over a deluge of technical criteria.
Nobody in that room was wrong. But every one of them was facing inwards, describing what the change meant to their function rather than what it meant to the person receiving it.
The instinct in that situation is to pick a winner quickly and move. It's faster, and it produces a message that three departments spend the next six months quietly undermining because they were never heard. The alternative costs a meeting: let each function make its case properly, in front of the others, and then make the call in the open with the reasoning visible.
People will accept a decision that went against them far more readily than one that was made without them. That's not a soft principle, it's an operational one – the messaging holds afterwards because nobody in the building has a private reason to work around it.
> The Stories - Four moments that capture putting my thinking into action.
> About me - Learn more about me and the career that has shaped how I see things.