Not too long ago, I went looking for a pair of weightlifting shoes. This is a niche, specialist item that’s not available on every shelf, so I started my search at a category-leading global sports brands store. Straight out of the gate, the staff were not familiar with their own lifting range and referred me to their global website instead.
Perhaps the local operation and the global team are not well integrated. Or perhaps their sales training is lacking. Regardless, I was back at square one.
The customer experience waiting for me online was even worse. I engaged with five different chatbots in five minutes, none of which were able to help. The one human I managed to reach referred me back to a chatbot. In the end, I gave up and bought from a competitor whose experience was genuinely personal, human-led and quietly powered by AI rather than hidden behind it.
This is a small story; a microcosm of a very trickly problem, namely everyone now owns the same intelligence. For senior leaders, the only question left is whether theirs still sets them apart.
The brand that lost out on my singular shoe purchase did not lack tech. It lacked a decision about where the technology belonged, and a connected view of the customer to put behind it.
It had aimed AI against its own strategy and operational design, deflecting customers at the front door like a discount retailer, while selling aspiration and performance at a premium.
According to Forrester, roughly one in three companies will damage their own customer experience in 2026 this way, by pushing AI into places it does not belong, usually for efficiency and rarely with a clear strategy or data foundations beneath it. The technology worked as designed. The thinking did not.
companies will damage their own customer experience in 2026 by pushing AI into places it does not belong
Source: Forrester
Senior leaders are starting to feel this privately, as well.
A widely cited MIT study found that around 95% of organisations deploying generative AI have seen no measurable return so far, even as global spending climbs toward two and a half trillion dollars, the most expensive technology wave in history
of organisations deploying generative AI have seen no measurable return so far
Source: MIT study
If your AI has underwhelmed, you were most likely not careless; you were sold a story about the wrong thing. The story was never about AI. It is about how AI can bring your customers, partners, employees and ideas closer.
Everyone holds the same mirror right now
Picture AI as a mirror held up to your operation; undistorted and larger than life.
A clear, well-run business becomes sharper and faster. A confused one gets its confusion reflected at scale. The technology supplies no substance of its own; it only amplifies what is already there.
The mirror is everywhere, and basically free of charge. The intelligence that looked like an advantage eighteen months ago can be downloaded this afternoon. Intelligence, the resource we spent a century treating as scarce, has become a commodity. Your competitor holds the same mirror. So does the one you have not heard of yet.
That is the polarity at the centre of the moment: the more capable and universal AI becomes, the less it can distinguish you. When everyone has intelligence, intelligence is the cost of entry, not the advantage.
What distinguishes you is what you choose to put in front of the mirror, a strategy and an operating model built to serve it.
AI will not tell you what to be
The gap this exposes is not technical. 70% of CMOs say becoming strong at AI is critical to their future, yet only thirty per cent believe they are genuinely ready. This is a gap in clarity, not capability, the space between owning the mirror and deciding what to hold up to it.
A gap in clarity, not capability
Strategy is not a plan for the next quarter. It is the position you intend to hold two or three years from now, and until that is settled, AI has nothing to serve.
The same technology answers to a different master depending on the position you have chosen, a cost leader bends it toward speed and lower cost, a focused differentiator bends it toward flexibility and quality, kept well away from anything that would cheapen the experience.
This is precisely where the sportswear brand went wrong: it deployed AI like a discount operation optimising for cost, at the very moment its aspirational positioning demanded a human.
Which of the five is your AI actually driving?
Strategy only becomes real through an operating model, the balance between people, process, technology, information and product.
AI is not a sixth element bolted on top; it is layered into technology and information, shifting the whole equilibrium, amplifying a working system or amplifying dysfunction, faster and at greater scale.
Modern operational trade-off frameworks highlight every well-designed operating model produces five outputs, namely: speed, quality, flexibility, dependability and cost. No system maximises all five at once, so your strategy has to decide which two or three actually win you the business. The single most useful question for any AI initiative is the simplest one: which of the five is this actually driving, and is it the one my strategy needs? Most AI is quietly pointed at speed and cost, the easiest to automate and measure, whether or not the strategy asked for them.
The five outputs of an operating model
Orange: where most AI is quietly pointed. No system maximises all five at once.
Product design inside that model poses its own test: where does your offering sit between utility and emotional benefit, and which of the five outputs above does your AI need to protect?
A luxury brand like Rolls-Royce sells almost pure emotional benefit: status, craft, belonging. Its operating model weights quality and flexibility over speed, and keeps AI working invisibly behind the scenes, protecting human personalisation at every touchpoint.
A commodity provider like a mass-market telco sells largely on utility: speed, price, reliability. Its customers expect AI as the first line of response, because scale, not warmth, is the value being delivered. Get this backwards, as the sportswear brand did, and you hand a premium customer a discount experience, or a scale customer an expensively slow one.
The mirror is unforgiving here: if the operating model is unclear, AI reflects that back, which is why so much marketing AI disappoints.
It is rarely the model that fails, but the foundation beneath it. Gartner predicts that through 2026, organisations will abandon around sixty per cent of AI projects unsupported by the data they depend on. Before investing in a more sophisticated model, you need a data management strategy, because information is what the intelligence runs on.
of AI projects unsupported by the data they depend on will be abandoned through 2026
Source: Gartner
Once the operating model is clear, use cases stop being a guessing game. Take the centre of excellence model most global matrixed organisations run, where regional teams brief a central shared-services function to achieve scale.
Ask where AI belongs inside it and the answer turns concrete: sharpening the brief so regional intent survives translation. That is not ‘adopt AI’. It is a defined use case, anchored to the model, with a specific needle it exists to move.
Fast thinking is free. Slow thinking is the edge.
There is one input the mirror cannot give you, and it is the most human. After giving Daniel Kahneman’s “Thinking Fast and Slow” a read, your brains thinking system is split into fast, which is the automatic thought, and slow, more cognitive and deliberate thought.
AI now does the fast thinking brilliantly and for almost nothing. The slow and cognitive thinking, deciding what matters and why, and how it connect relationally and emotionally, cannot be handed over. It has become the most valuable capacity a leader holds.
Buyer behaviour shows the same pattern. Around 67% of B2B buyers now prefer a largely rep-free, self-service experience , yet sixty-nine per cent still turn to a rep to validate what the AI told them. People want the machine's efficiency and a human's reassurance at different points in the same journey.
of B2B buyers prefer a largely rep-free, self-service experience
still turn to a rep to validate what the AI told them
The machine's efficiency and a human's reassurance, at different points in the same journey
My own experience with that sportswear brand was a failure of exactly this judgement: I reached the moment I needed a person, and the system was built to keep me from one.
Get the sequence wrong and the cost is not abstract. Gartner expects more than forty per cent of agentic AI projects cancelled by the end of 2027, defeated by unclear value rather than weak technology. Optimise everything measurable while starving everything else, and you commoditise yourself with a healthy budget and the best of intentions.
Brand is the one thing you cannot download
If intelligence is now a commodity and products are converging, the question becomes what is left to compete on: the oldest idea in marketing, and the most underrated in technology. Brand.
Brand is where the human is, where emotion lives and relevance is decided. It cannot be copied from a model card or bought by an oversupply of tokens. Brand equity comes from a few durable things: awareness, association, perceived quality and loyalty. None of it is downloadable, and all of it compounds.
It matters more now, not less. At any given moment only around five per cent of your buyers are in the market; the other ninety-five are slowly forming an impression of who is worth considering when the need arrives, which is why brand means occupying that neurological shortlist long before the buyer is ready.
Your buyers at any given moment
Orange: in market now. Navy: the buyers your brand is reaching for tomorrow.
Brands with a share of voice above their share of market tend to grow into the gap, because attention compounds into memory and memory into preference.
There is now a second, non-human audience for all of it: AI has quietly become a reseller of your brand.
A key insight from a speaker at Cannes Lions 2026 was that in LLM environments, the journey from awareness to action is compressing rapidly, and brands can no longer predict where or how they will show up.
The machine's answer to a buyer draws on the salience, authority and trust your brand has built across the web.
If you are not in the human's memory, you will not be in the machine's recommendation either. Brand earns you a place on two shortlists at once, the buyer's and the algorithm's.
The commercial translation is direct. Brand salience lowers your cost of demand, shortens the path to the shortlist, lifts your win rate, and turns satisfied customers into advocates who carry your name to the ninety-five per cent not yet looking. Brand is not the soft end of marketing. It is the compounding asset that makes everything downstream cheaper and faster, the part of your business the machine can amplify but never replace.
Your move (not the machine's)
So, is your intelligence irrelevant?
The kind you can download certainly is. What remains relevant is the clarity to decide where you are going, how you are positioning your brand in the hearts and minds of those you serve, , the discipline to design an operating model that gets you there, and the slow thinking to know which of the five outcomes your AI should serve.
Decide on the position you intend to hold in three years. Design the operating model, the balance of people, process, technology, information and product, that will deliver it. Then, and only then, layer the intelligence in, asking of every use case the question that separates advantage from expense: which of the five is this driving, and does it bring my customers, partners and people closer, or push them away?
The mirror is in every leader's hand now. It will reflect your operation back exactly as it is, undistorted and amplified. The only question worth asking is whether you have built something worth amplifying? This is the kind of work we help B2B technology leaders with at ProximityB2B: helping them to discover, redesign, and optimise their marketing strategy so there is something worth amplifying before the technology ever gets involved.
Where AI belongs in your operating model, and where brand still needs to do the work no model can, is usually easier to see from outside the mirror. If this raised a question worth working through, I'm always up for that conversation.
Jonathan Booysen (CM)UK MBA is the Founder and Managing Director of ProximityB2B, a Global B2B technology marketing consultancy helping founders and growing teams bring customers closer by connecting brand, demand, and revenue to commercial outcomes.
Frequently asked questions
Why do so many AI projects fail to deliver a return?
A widely cited MIT study found around 95% of organisations deploying generative AI have seen no measurable return so far. The technology usually works as designed; what is missing is a clear strategy, an operating model and the data foundations for it to serve.
How should a business decide where AI belongs?
Start with the position you intend to hold in two or three years, design the operating model of people, process, technology, information and product that will deliver it, then ask of every AI use case which of the five outputs it drives: speed, quality, flexibility, dependability or cost.
Should B2B companies replace salespeople with AI?
Not entirely. Around 67% of B2B buyers prefer a largely rep-free, self-service experience, yet 69% still turn to a rep to validate what the AI told them. Buyers want the machine's efficiency and a human's reassurance at different points in the same journey.
Why does brand matter more as AI becomes a commodity?
When everyone has the same intelligence, brand is the one thing that cannot be downloaded. Only around 5% of buyers are in market at any moment, so brand is what puts you on the shortlist long before the buyer is ready.
How does AI search change the way B2B buyers find brands?
AI has become a reseller of your brand. Its answers draw on the salience, authority and trust your brand has built across the web, so if you are not in the buyer's memory, you will not be in the machine's recommendation either.
