B2A, AEO, GEO and beyond: your guide to marketing’s new AI language

You’ve mastered B2B. You understand B2C. You may even have survived several years of people talking enthusiastically about D2C. Bad news. B2A has arrived.

B2A. What’s that? Business-to-AI. Ahh yes. Of course. 

And, annoyingly, it actually makes sense. As AI agents start researching, recommending and eventually transacting on our behalf, brands may find themselves marketing not just to people, but to the machines acting for them.

We can thank Rich Curtis for this one. The FutureBrand CEO casually dropped the term into fireside chat at the official unveiling of earned-first agency VOCAL (formerly Weber Shandwick) last week with the confidence of something the rest of us apparently should already have known.

Curtis later admitted the term had popped into his head over breakfast while he was thinking about potential talking points for said panel. Thanks, Rich.

Because just when we thought artificial intelligence might simplify our working lives, the marketing industry has done what it does best and started inventing an entirely new vocabulary to explain it.

And things heated up quickly.

Not long after B2A entered the room, VOCAL co-founder and managing partner Angela Malkin raised the stakes with another doozie: B2A2B.

Business to AI to business. Gulp. To be fair to Curtis and Malkin, there is a serious idea sitting underneath all of this alphabet soup.

We are moving towards a world where brands will not only market to humans. Increasingly, they will also need to communicate with the AI agents researching, recommending, comparing and potentially buying things on our behalf. A game-changer for the comms and marketing industry. 

This has implications for everything from brand discoverability and search to ecommerce, customer experience and the way companies structure their data.

So B2A may sound like something invented five minutes before a conference panel. It may also turn out to be a fairly useful way of describing what comes next. And it is hardly alone.

AI is rapidly producing a whole new language for marketers, much of it arriving before most of us have properly worked out the last batch.

So, in the interests of keeping up at industry events, surviving your next agency presentation and arming yourself for the potentially magnificent LinkedIn posts ahead, here are a few terms worth knowing.

B2A: Business-to-AI

The basic idea is simple enough: businesses will need to market not only to people, but to the AI agents acting for them.

Us human beings are leaning on AI agents to find the best home loan, compare cars, book a holiday or choose a new pair of running shoes. Which means your next customer may not be a customer at all. It may be software.

Likelihood you’ll hear it at Cannes next year: Extremely high.

Likelihood somebody puts “B2A strategy” in a LinkedIn post before Christmas: Also extremely high.

B2A2B: Business-to-agent-to-business

This is where things got out of hand. Malkin’s extension of the idea imagines agents sitting between businesses themselves, researching suppliers, comparing services, sourcing information and potentially helping negotiate or transact.

Once again, there is a legitimate shift underneath the acronym.

But B2A2B is also the point at which we may need to establish some industry-wide rules, otherwise, by Cannes 2027, someone is inevitably going to unveil B2A2A2B and insist it requires a 47-slide presentation.

AEO: Answer engine optimisation

Remember SEO? Of course you do. It’s still here. But now there’s AEO, which is essentially about making sure your brand turns up when someone asks an AI system a question rather than typing keywords into a traditional search engine. This is now mission critical for brands wanting to show up in the right way in LLM's, or large language models.

Instead of fighting to rank first on Google, brands want ChatGPT, Gemini, Perplexity, Claude and other AI systems to know who they are, understand what they offer and include them in the answer.

Quite important, actually.

Unfortunately, that means we now need another acronym.

GEO: Generative engine optimisation

This occupies roughly the same neighbourhood as AEO.

Ask five people where AEO ends and GEO begins and there is every chance you will receive six answers.

The broad ambition is the same: make your company, content and products visible to generative AI systems. There are already agencies, platforms and consultants building businesses around it. Naturally.

Agentic commerce

This one sounds particularly good when delivered slowly on a conference stage. But again, the concept matters.

Instead of AI simply suggesting what you should buy, agents could help research products, compare options and complete parts of the transaction.

The internet gave us ecommerce. AI would now quite like to do the shopping itself.

Agentic enterprise

Keen readers of this trade publication will have noticed this term "agentic enterprise" cropping up quite a lot lately. I've become sort of obsessed with it, no thanks in part to my annual pilgrimage to Salesforce's annual tech bonanza Dreamforce, where the agentic enterprise is a hotly contested topic.

This is also one of the safest phrases available to anyone presenting a marketing trends deck in 2026. Broadly, it describes the shift from marketers using AI as a tool to AI agents carrying out multi-step tasks across marketing systems.

Planning. Personalisation. Optimisation. Content. Reporting.

Potentially useful.

Also almost guaranteed to appear in the sentence: “We are only at the beginning of our agentic journey.”

Share of model

Marketers spent decades worrying about share of voice. Now comes share of model: how visible your brand is inside the answers produced by AI. Ask an AI assistant for the best running shoes, banks, universities or insurance companies.

Does your brand appear?

If not, somebody in marketing is going to want a dashboard explaining why.

MCP

Model Context Protocol. You do not necessarily need to understand this one in enormous technical detail. You merely need to nod thoughtfully when somebody mentions it and say: “The interoperability piece is going to be really interesting.”

Then move swiftly towards the coffee cart.

A2A: Agent-to-agent

AI agents communicating with other AI agents. At this point, there is a reasonable argument the humans can simply leave the meeting.

And perhaps that is the larger point. For all the fun you can have with language, the underlying change is moving quickly.

Search is changing. Commerce is changing. Brand discovery is changing. The interface between companies and customers is changing.

Some of these acronyms will disappear as quickly as they arrived. Others may become as unremarkable as SEO or CRM. B2A might even stick.

And just when you think the acronym factory has finally shut for the day, along comes US president Donald Trump's rebranding of AI to SI, or superintelligence, which gives us B2S.

Or, as the rest of us may eventually call it: BS.