Most marketers are about to have a bad year

Many just don't know it yet.

In the last year, LinkedIn lost 60% of its B2B traffic to AI search.

Email open rates climbed but click-throughs fell. The emails your team spent hours on are being opened, summarised, and sorted by AI before a human reads them.

Some marketing teams looked at the open rate, called it a win, and moved on. Others are still celebrating.

That’s the small version of the problem.

AI has changed how people find information. When someone asks ChatGPT or Perplexity which vendor to use, certain names come back. The brands that surface tend to be the ones AI can read and verify clearly. Ad spend doesn’t factor in much.

You’re no longer competing for a click. You’re competing to be the answer when no one’s searching.

Most marketing teams know this version of the story, even if they haven’t updated their approach for it yet. What they’re less aware of is the second game.

It’s not just how people find things that’s changing. It’s who decides.

Google now runs shopping ads inside its AI search interface, with 75m daily users. OpenAI is building ad campaigns that set their own budgets, create their own assets, and optimise without anyone making a call.

Sometimes the buyer isn’t a person anymore. It’s an agent or system, operating on someone’s behalf, checking criteria they set up once and largely forgot about.

In B2B, that changes something fundamental.

When a business evaluates a vendor today, a person does it. They read case studies, sit through demos, call references, form impressions in meetings. Human buyers respond to story, to warmth, to the sense that you understand their problem. You can have a rough first conversation and recover in the second one.

A system doing that same evaluation has none of that. It reads data. It checks whether evidence can be verified. If there are two vendors with similar pricing and one has cleaner documentation, structured case studies with verified numbers, and consistent third-party mentions, and the other has a better looking website and a slick sales rep, the system picks the first one. Gut feel doesn’t factor in and rapport doesn’t carry over.

Most B2B marketing is built around the sales rep and the slick website. Built to impress humans in rooms and build trust. Which is fine, right up until the room has no humans in it.

Nobody knows exactly when AI buying agents become the dominant mode. But the direction is already clear.

In B2B payments in MENA, this has already moved from theory to roadmap. The conversation isn’t about AI evaluating vendors anymore. It’s about AI executing the transaction. The buyer and the payer are becoming the same thing, and neither of them is human.

Gartner surveyed 402 senior marketing leaders last year. 65% said AI will fundamentally change their work. 20% said they personally need zero skills updates to deal with it. Only 15% of CEOs believe their CMO is AI savvy right now.

By 2027, Gartner predicts AI illiteracy will be one of the top three reasons large-enterprise CMOs get replaced. If not sooner.

The easy response is that marketing leaders need to learn the tools. That’s true but it’s the smaller problem. The harder one is leaders who understand the tools just fine and still haven’t grasped that the job changed. They’re treating AI as efficiency software. Something that does the same things faster.

It isn’t. AI is changing who finds you and who chooses you. Running the same strategy at higher speed doesn’t help when the race moved.

The gap in that Gartner data is worth sitting with. The same leaders who expect AI to fundamentally change their industry don’t think they personally need to change. Somewhere between “this transforms marketing” and “I probably don’t need to upskill much,” the logic fell apart. Most of them are thinking about AI as a better version of what they already had. A faster horse.

What marketing needs to become is genuinely different from what got most people into senior roles in the function. Being cited by AI systems requires different thinking from being ranked on Google. Being selected by a buying agent requires different trust signals and proof from persuading a procurement lead. And doing both simultaneously, with the same team and budget, while the original game keeps running in parallel, is the actual problem. Nobody’s written that job description yet.

The CMOs who work this out first won’t be the ones who learned the tools earliest. They’ll be the ones who understood the job changed and adjusted their thinking, not just their software.

That’s a harder thing to see from inside a function that’s been doing roughly the same thing for 20 years, and largely gotten away with it.