Nobody Saw the Ad. Everybody Bought the Product.

Somewhere between your first sip of coffee and your second, your dedicated personal AI agent had already done more work and made more decisions than you can in a week.

It reviewed 47 brand pitches targeted towards you, otherwise known as advertisements. Rejected 44. Negotiated better terms on two. Completed a purchase on one. You didn't see a single ad. You didn't compare a single price. You didn't even know it happened.

By the time you sat down at your desk at work, the transaction was done. A product was on its way to your door because your AI agent decided, based on everything it knows about you, that you needed that particular one. And the brand that won? Well, it never spoke to you. It spoke to your agent. And your agent spoke back.

This sounds like something out of a pitch deck for a startup that hasn't shipped yet. It isn't. The infrastructure for this world is being built right now. Google launched its Agent to Agent protocol in April 2025, backed by over 100 companies and now governed by the Linux Foundation[1]. Anthropic released its Model Context Protocol around the same time. These aren't announcements. They're plumbing. The kind you lay before you build the house.

I have personally spent over 20 years in marketing. I've survived the internet, social media, programmatic advertising, and the first wave of AI tools. Each time, the disruption changed the channel. A new place to show up. A new format to learn. A new audience behavior to study. This one doesn't change the channel. This one changes the participants. For the first time in the history of commerce, the entity on the other side of your message might not be human. And it might not care about your storytelling.

When Machines Learned to Talk to Each Other

The idea of AI agents communicating with other AI agents sounds futuristic until you realize we already did it once and barely noticed.

Enter: Programmatic Advertising. For my non-marketing readers, the majority of digital display ads are bought and placed by machines, in milliseconds, without a human ever seeing the creative, approving the bid, or choosing the placement. We called it "automation." We built entire careers around it. We never stopped to ask what happens when that same logic extends beyond ad buying into the entire customer journey.

Now it has. The AI agent market grew from $5.4 billion in 2024 to $7.84 billion in 2025, with projections indicating it will reach $52.62 billion by 2030[2]. Gartner predicts that by 2028, 60% of brands will use agentic AI to deliver one-to-one interactions[3]. McKinsey found that 62% of organizations are already experimenting with AI agents, though only 23% have begun scaling[4].

The protocols exist. The investment is real. And the question is no longer "will this happen?" The question is "what does the world look like when it does?"

Because when a company's AI agent can talk directly to a consumer's AI agent, negotiate terms, customize offers, and close a deal without either human touching the process, we are not describing a new marketing channel. We are describing a new kind of marketplace. One where the stalls are run by algorithms and the shoppers are, too.

Where Did the Humans Go?

Do you remember that Black Mirror episode "Joan Is Awful"? The one where Annie Murphy basically discovers an AI has been quietly living her life, making choices she never actually made. Genuinely unsettling stuff. We all watched it and thought, "Well, that's terrifyingly creative." Great TV. Definitely not real life.

Funny thing is, companies are now building the commercial version of that exact premise. Not the creepy Netflix dramatization, obviously, but the core idea. AI agents that act on your behalf. That decides what you see, what you buy, who earns your attention and your money. The question is no longer whether this will happen. It's whether you'll even notice when it does.

That thought opens a cascade of questions that nobody in a boardroom seems to be asking yet.

If an AI agent rejects your brand in 0.3 seconds, does your brand story still matter? If creativity exists to move humans, what moves a machine? Does "emotional appeal" mean anything to a language model that processes sentiment as data points, not feelings?

What about the digital divide? Wealthier consumers will afford smarter AI agents. Better filters. Better negotiators. Better taste, algorithmically speaking. Does that mean premium brands will need to impress a more sophisticated machine before they ever reach the human behind it?

And what about noise? Social media already feels overwhelming. Too much content, too many voices, too little signal. Now multiply that by billions of AI agents generating, filtering, and responding to content at speeds no human can follow. The volume doesn't decrease. It becomes deafening in a frequency only machines can hear.

What really makes my head spin is that we are, presumably, building these systems to serve us. But at what point do they start shaping us instead? Not in a dramatic, dystopian way. In a quiet, comfortable way. The way we stopped memorizing phone numbers when smartphones arrived. The way we stopped navigating when GPS took over. The way we stopped questioning search results after the third click. Slowly, imperceptibly, until one day we realize we've become a species that defers, by default, to the machine. Not because we were forced. Because it was easier.

Marketing Will Feel This First

If there's one industry that should be watching the agent to agent future with both eyes open, it's marketing. Not because marketing is more exposed than other industries, but because marketing already lives in the space between human desire and commercial persuasion. And that space is about to be occupied by machines on both sides.

We already touched on how programmatic advertising has proven the model. Machines buying ad space from machines, autonomously, at scale. That was the proof of concept. Agent to agent commerce is the full execution: extending that logic across targeting, content creation, negotiation, personalization, and purchase.

Here's what I believe will happen to brands, and this is a personal developing view. Brands will be split into two. One identity for humans: emotional, visual, narrative driven. The brand building we've practiced for decades. And a second identity for machines: structured data, trust scores, and reputation signals that an AI agent can parse in milliseconds. A machine readable brand. Think of it as your brand's resume, formatted for an audience that doesn't have feelings but does have preferences.

The CMO of 2028 won't just ask "does our campaign resonate with people?" They'll ask "does our brand data structure give us an advantage when a consumer's AI agent evaluates us against 400 competitors in under a second?"

Now, this is where I have to ask the uncomfortable question closer to home. We declared 2026 the Year of Artificial Intelligence in Saudi Arabia[5]. Our enterprises are deploying AI solutions at an impressive rate[6]. We're scaling automation, attracting AI funding, and building the infrastructure[7]. All of that is real. But are we, as marketers and business leaders in this region, actually ready for the shift I just described? Because being in the middle of an AI investment boom and being prepared for agent to agent commerce are two very different things. Most marketing teams I see across the GCC are still figuring out how to use AI for content and campaign optimization. Nobody is talking about what happens when the customer's AI agent is the one you need to convince. That gap should worry us.

There's also an ethical dimension we can't sidestep. If your AI agent decides what you buy, are you still making a choice? If a brand's AI agent learns to manipulate your AI agent's decision criteria, who is responsible? Who programs the filter, and whose interests does it serve? These aren't philosophy seminar questions. They are the questions that will define consumer trust in the next decade.

What to Learn Before the Window Closes

This isn't the internet arriving in 1995. It isn't social media emerging in 2007. I said it before, and I'll say it again: those disruptions changed channels. A new place to reach people. A new format to master. This one changes the participants. A human is no longer guaranteed to be on the other end of a transaction, a negotiation, or even a conversation. That's a category of disruption we simply have not experienced before. And the window to prepare is measured in months, not decades.

So what does preparation look like? Three major steps have to be taken immediately:

First, start experimenting with AI agents now. Not next quarter. Not after the board approves a budget line. The tools already exist, and the learning curve rewards early starters. Build internal muscle memory. Let your team test, fail, and iterate. The organizations that wait for perfect clarity will find themselves fluent in a language the market has already moved past.

Second, hire differently. The marketing department of 2028 will include people who understand agent architecture, AI systems thinking, and the mechanics of how machines evaluate, compare, and choose. Not as a line on a job description, but as a core competency. We learned SEO. We learned social media algorithms. We learned programmatic buying. This is the next literacy, and it carries higher stakes.

Third, study the protocols. A2A, MCP, and the agentic frameworks emerging across the industry are to the agent economy what HTTP was to the web[1]. You don't need to become an engineer. But you need to understand what these systems do, how they connect, and what they mean for how your brand shows up when machines are the first audience.

The companies that treated the internet as a fad in 1995 didn't make it to 2005. The brands that dismissed social media in 2008 spent the next decade trying to catch up. This shift will move faster. Gartner estimates that 40% of enterprise applications will feature task-specific AI agents by 2026, up from less than 5% in 2025[8]. That's not a gentle evolution. That's a wave.

I started this by describing a morning where your AI agent handled everything before you finished your coffee. 47 pitches reviewed. 44 rejected. One purchase completed. All without you.

It sounded almost peaceful, didn't it? Efficient. Maybe even liberating.

But sit with it for a moment. What did you actually do that morning? You drank your coffee. You went to your desk. And somewhere between those two things, a machine made a dozen economic decisions on your behalf. Not because you asked it to in that moment. Because you trained it to. Because the convenience became so natural that you stopped questioning it. And then you stopped noticing it altogether.

I never expected AI to move at this speed. A year ago, we were discussing tools that could write emails and summarize reports. Now we're building protocols for machines to negotiate with other machines on our behalf. This isn't the same conversation anymore. It isn't even the same room.

I don't know where this ends. I don't think anyone does. But I know that the marketers, the leaders, and the builders who will navigate this shift are the ones asking these questions right now, while the answers are still ours to shape.

The machines are already talking to each other. The question is whether we're still part of the conversation.