GEN AI

THE BIG SHIFT: 2026 IS ALREADY THE YEAR OF THE AGENT

We are IN the Agentic Age.  

Last year, Google and Walmart confirmed something that should be on every brand marketer's radar; AI agents are now completing purchases on behalf of users inside closed ecosystems, without a human ever visiting a product page.  

Which means, if your brand is not architected for machine-readable decision-making, you will not appear in those flows.  

Not deprioritised: absent. 

We’ve moved past the era of simple content generation. LLMs are recommending, shortlisting and transacting on behalf of users, which means your competitive set is being shaped before an actual person ever sees it. 

This gives us a clear strategic steer: optimising for search visibility is no longer enough. The real question is whether your brand is visible to the agents making decisions upstream. 

There are two imperatives that follow... 

1. Design for agents, not just humans.

Your brand is now increasingly interpreted before it is experienced.

When an LLM model answers a question, it is synthesising patterns from product pages, press coverage, reviews and third-party listings simultaneously. It doesn’t have room for ambiguity. Fragmented or contradictory positioning across those sources creates conditions for your brand to be mischaracterised, or miss out on opportunities completely. 

The task here is ensure alignment and consistency. Ask yourself:

• Are your category descriptors consistent across channels?

• Do your product names clearly signal what they are and who they are for?

• Are your differentiators clear to everyone, or buried in brand language that only makes sense in your own boardrooms and meetings?

• Do independent sources reinforce the story you are trying to tell?

Then test it. Take real, high-intent prompts from your category and run them across leading models. Notice how your brand is described and how competitors are framed. Are you positioned as premium or generic? Specialist or interchangeable? Those answers may surprise you, but you’ll know exactly where to start. 

This then becomes a diagnostic layer, alongside search and share of voice tracking. AI doesn’t invent your positioning, all it does is reflect the patterns it finds, and if those patterns are inconsistent, that inconsistency becomes part of how your brand is portrayed. 

PRACTICAL STARTING POINT:

Google and Microsoft have both confirmed that Schema markup is used to power their generative AI features, making it a strategic requirement rather than an SEO consideration. Audit your structured data estate as a priority. 

2. Build intelligence advantage through GEO

If our first imperative is about fixing what agents read, this one is about understanding what they conclude.  

Even with your house in order, there’s another question to get curious about: when people ask AI for the best option, where does it place you, and why? 

LLMs are already compressing categories into quick, confident summaries (do a few “best X for Y” searches and you’ll see the pattern). Brands get labelled and filed away: premium, reliable, good value, innovative, niche. Sometimes they’re accurate, sometimes they’re lazy, more often than not they’re inconsistent—but they’re affecting how you’re perceived regardless. 

That’s where Generative Engine Optimisation (GEO) comes in. 

Paying attention to these ‘compressions’ reveals how your story is being told before anyone is even reading your website. It shows where you’re coming through clearly, where you’re being blurred into the middle, and where competitors are simply easier to describe than you are. 

Practical starting point:

Run a small set of high-intent prompts across a few models each quarter. Save the outputs. Look for themes: repeated descriptors, missing proof points, competitors getting credit for things you also do, or your brand being compared on the wrong attributes. 

Those patterns give you a practical to-do list: what to clarify, what to evidence, and what to own more deliberately. 

FIVE BEHAVIOURS SEPARATING LEADERS FROM FOLLOWERS

Beyond the technical shifts, we’re seeing some key differences in behaviour between those leading in the GenAI space and those who aren’t. Here are the five that we feel stand out most: 

1. LISTENING TO MEASURED EXPERTISE, over hype

We know how fast AI is progressing. And the commentary around it moves even faster. With new developments, commentaries and supposed ‘hacks’ bombarding us left and right. It’s easy to feel like you’re behind. 

However, those leading here aren’t twitching at every “game-changing” announcement. What’s needed is enough cynical judgement to distinguish between platform marketing and genuine capability shifts, which is a harder skill than it sounds when every press release is written to blur that line... 

Practical move: Before adopting anything new, stop and ask yourself: 

• Does this solve a defined business problem? 

• Is it proven in live environments, not just demos? 

• What changes operationally if we adopt it? 

If the answers are vague, wait. Advantage rarely comes from being first. It comes from being right. 

2. prioritising a PANORAMIC view of llm MODELS

A brand that appears authoritative in ChatGPT may be poorly represented in Gemini or Perplexity. Relying on just one model puts you at a distinct disadvantage. That said, the gap between outputs across different platforms is itself a strategic signal worth investigating.  

Practical move: Consider building a testing cadence across at least three models. This gives you a more holistic picture of where your brand stands in the AI information ecosystem, and where and your structured data needs a boost.

3. REFRAMING AI AS a CREATIVITY AND GROWTH opportunity

The efficiency case for AI has been made, repeatedly and loudly. And whilst time saving is nice, it’s also just a fraction of the opportunities available. 

The opportunity: The real win here is using AI to expand what’s possible. Tapping into these tools to create genuinely differentiated work. Greater personalisation, faster iteration, more creative routes explored before narrowing. Work that was structurally impossible two years ago. 

Think beyond doing the same work faster, and lean into the creative opportunities and imagination that AI tools afford us. 

4. USING NATURAL LANGUAGE CODING TOOLS TO PROTOTYPE FASTER THAN EVER BEFORE

Natural language coding tools have turned non-technical marketers into architects.  A senior marketer who can take an idea to a working prototype without waiting six months for a dev queue is a dangerous competitor. They can test, fail, and pivot with almost zero risk.

The opportunity: Prioritise creating the space and opportunities for teams to train and start prototyping their ideas. With more of your team armed and able to trial things, the fewer barriers to exciting possibilities. 

5. BRIDGING THE “AI SKILLS GAP”

Access to tools is no longer a differentiator. In fact, almost everyone has the same subscriptions. 

The real advantage in 2026 (which many AI strategies overlook) is whether your workforce can actually use them to produce work that is better, faster or more considered than it was before.  

Practical move: Growth will increasingly be constrained or unlocked by how quickly organisations can adapt skills, workflows and ways of working, not simply by access to new tools. Upskilling cannot be a 'one-off' training project; it needs to be treated as a permanent operational imperative. 

the bottom line:

Those that treat AI as a capability shift rather than a feature upgrade are the ones who will blaze ahead. The ones who treat it as a fancy new feature will find themselves being summarised, compared and ultimately deprioritised by systems that have no interest in waiting for them to catch up.