New job energy is a real thing. I’ve been at TrunkBBI for just under a month, and I’m already knee-deep in the fun stuff. Seventeen years in SEO across agency and in-house, including LEGO and THG, teaches you a few patterns. One of them is this: every few years the way people find and pick brands quietly shifts, then all at once it doesn’t feel quiet at all.
Earlier this week I sat down with our Performance & Innovation Strategy Director, Matt Holmes, to compare notes on where AI is taking us. This is the write-up.
The short version is that AI is changing the decision itself. Old school search gave you ten blue links and a promise that the best would rise. Now you have AI systems that summarise, compare, and increasingly suggest. You’re not just trying to be seen, you’re trying to be the brand that gets pulled into the answer, the one an assistant name-checks, the one that feels like the safe and best recommendation.
Discovery isn’t happening in one place anymore. Google remains the trusty all-rounder and remains the default for most, but people hop between YouTube for demos, TikTok for social proof, Amazon for purchase signals, and yes, LLMs for quick sense-checks and to support decision making. If your brand and story only lands in one of those places, you’ll be invisible in the others.
On the SEO side, building highly relevant links still matters, but acquiring brand citations and mentions are now an increasingly important part of the mix. The systems that build AI answers learn from the open web, so the fundamentals still matter: solid content, clean tech, real expertise.
What’s new is the added emphasis on building evidence and consistent brand signals beyond your own site. Original research, credible reviews, named authors, honest comparisons, coverage in places your audience already trusts. That’s what earns a mention when a model assembles an answer. If you’ve been treating E-E-A-T as a checkbox, this is a good moment to evolve and to treat it as a publishing style.
Measurement needs a tweak too. Keep tracking positions, because they’re a decent proxy for visibility and quality, but add indicators of presence. Are you being referenced by others? Are your videos being surfaced? Do your products show up accurately? Are you part of the sources that answers lean on? If the only number you look at is keyword ranking, you’ll miss the game that’s being played across multiple platforms.
Paid social has its own flavour of change. Creative is quietly becoming the new targeting. The platforms can now assess your ads at scale, decide what looks credible, and learn faster than any manual audience setup. If that scares you into generating hundreds of cookie-cutter variants, resist the urge. People can spot auto made. The brands that are cutting through are using AI to speed up exploration, then finishing with taste, craft and a human editor. Same testing discipline, just higher standards for what earns a place in the feed.
First-party data is the fuel that makes all of this smarter. Post-cookie, broad targeting is blunt. Your own data helps algorithms find the right people and helps you understand intent with more precision. On organic, you won’t paste a CRM export into Google, but you can use what you know about real customers to shape the questions you answer and the objections you handle. When paid and organic share that intent model, you start turning up at the right moments with the right proof.
There’s also a very real reason to keep humans involved. AI still struggles with cultural nuance. Matt had a personal example as he tried to name a new dog with the help of an assistant. Let’s just say no one wants to be the person shouting the word “Bark” across a park! Machines can be confidently wrong in ways that make people wince. That’s fine if you treat AI as an assistant to brainstorm or accelerate ideas, it’s can be dangerous if you treat it as the author and the strategist.
So what does good look like right now? The brands turning visibility into significance tend to publish original, helpful content that actually answers questions. They collect and show real reviews, they compare options honestly, showing up where their audience is, not just where the marketing plan says they should be, and crucially, they’re referenced and cited by others. It’s an ecosystem play. Be the expert on your own site and be the expert other people point to.
Looking a year out, assistants will act as pre-qualifiers. More people will ask a model to narrow the field before they ever land on a results page.Who are the three best options? Which one ships faster? Which one has fewer returns?
If you don’t make that early short-list, you may never be considered. As agentic workflows grow up, expect assistants to do more of the heavy lifting, from comparison to checkout. That means optimising for humans and for agents: ensuring clear and accurate product data, transparent policies coupled with clean, structured content.
If you want a simple playbook, here’s mine. Keep the craft high. Publish real expertise from credible, expert authors who add real value. Structure your knowledge so machines can understand it. Earn mentions beyond your own backyard. Use AI to move faster, not to flatten your voice.
Visibility got you into the room. Significance gets you chosen.
