Interview · · 7 min read

Over half your traffic is already machines. The human is still the buyer. See it first, then shape it.

John Searby, Chief Strategy Officer at HUMAN, explains why over half of all customer traffic is now automated and no longer human, and why agent traffic has grown over 7,000% since tracking began. He argues the first job is visibility, then controls to shape which agents can browse, shop or check out, and that the deeper problem is verifying the human behind the agent. A cybersecurity veteran and self-described AI optimist, he sees the LLM as the best surface we have ever had to reach a buyer, while warning that fraudsters were handed the same powerful tool.

John Searby, Chief Strategy Officer, HUMAN. Interviewed by Justin Cooke.

Episode page
Share
John Searby, Chief Strategy Officer at HUMAN
John Searby, HUMAN

Listen while you read:The Agentic Surface

0:0040:00

Over half of all web traffic is now automated, no longer human

The setup.

HUMAN is in the business of protecting the consumer journey. It ensures real human eyes are served ads across the advertising ecosystem, cleaning supply from invalid traffic, and it sits in line on advertisers' and merchants' websites, in the traffic, protecting against bots and automated fraud. That means seeing the whole journey, from the moment an ad is served to the traffic that arrives, the login, the shopping, the subscription. Searby puts the company in the centre of commerce, and with the arrival of agents it has seen a massive uptick. Over 50% of all the traffic its customers see now is automated and no longer human in some nature.

Agent traffic rose over 7,000%, and Muse is now 70% of it

On the agent wave.

The internet was built for humans, and for years automation meant search bots and an implicit deal with Google: scrape my site and I get human traffic back. AI skyrocketed that, first crawling to feed the LLM, then real-time retrieval that live searches a site when you ask about a product, and now agents that act like humans. Searby calls the launch of Muse a critical inflection point. Agentic traffic grew over 7,000% since the company started tracking it last year, and it has skyrocketed in the last few weeks, with Muse now about 70% of all agentic traffic.

Visibility was king, then came the controls

On telling good machines from bad.

The first step for most customers was simply making sure they could see it, because many had no idea how much was hitting their sites. HUMAN's intelligence teams have tracked bad actors for years, and as agents arrived the question changed: some of them might be good, and you might want them to shop. It is business dependent. A company building its own chatbot may want to control that surface entirely, while a publisher may want none of its content leaving in an LLM. So the same tools used to detect automation for 14 years now detect Muse, GPT, Claude and other consumer-side agents, and customers get controls to tweak the policies: this one can see products, this one can check out.

From Business of Marketing

Most companies try to reach senior decision makers. We give them a reason to engage.

Partner with Us →

The real break is identity: who is the human behind the agent

On the identity problem.

It comes back to how we define the audience and measure it, Searby says. He uses these tools every day, with agents shopping for him, but it is still him guiding it and still his interests. The unsolved question is how we understand the identity of the human as it relates to the agent. Across ad tech, martech, cybersecurity, fraud and commerce, everyone is trying to answer the same thing: who is the person, which agent are they using, can we verify it is them and that they guided it, so we can protect from fraud and still optimise for good commerce.

On getting everyone in the room.

When a CMO sees what is really hitting their site, the reaction is usually shock, and it quickly becomes a large integration effort. Because HUMAN works with security and fraud teams as well as marketing, it needs internal collaboration, and Searby says more than ever in his career multiple stakeholders are sitting in the same room agreeing to share tech and data they were not even aware of. This is a fundamental change in how humans interact with machines, so every part of a business needs to understand how it is affected.

Nobody has this nailed, and it is changing at the consumer level first

On how early we are.

We are early, and we all are, Searby says. This is the fastest moving news cycle and innovation cycle, where you feel you have missed out if you look away for 12 hours. It is unreasonable to expect anyone has this nailed, and anyone who tells you otherwise is not paying enough attention. Uniquely, it is changing at the consumer level first. He points to his 80 year old father, who never figured out a computer and now talks to his LLM every day and can give it agency to shop and do things he never managed, suddenly becoming a new audience that would not traditionally have been on most websites.

“Right now, everyone's flying fairly blind across this”

The customer is still the buyer, just reached a new way

On the AI optimist view.

If your agent shops for you, the brand is still advertising to you, the human behind the scenes, just through a new mechanism. SEO becomes AEO and GEO. Searby got into woodworking this year and used to Google a tool, read Reddit threads, then buy on Amazon out of trust. Then the LLM got good at directing him, presenting other merchants, sometimes cheaper, and telling him he could drive to a big box store and pick it up now. His behaviour changed materially, but it was still him, just how it was presented. He is very much an AI optimist on this.

On how brands show up.

The AEO world is still being built, and Searby goes back to incentive structures: whoever owns the LLM monetises through subscription, advertising space or a cut of the commerce thread, and it is on them to keep the supply clean, trackable and measurable with real humans at the end. The opportunity he sees is that what you present should change depending on whether it is a crawl, real-time retrieval or an agent. If it is a crawl, present this page. If it is a live search, present this page. If it is an agent, give it different rails. It is a hybrid approach aligned to your business plan, bringing hard traffic together with the analytics side so you see the full picture.

A new walled garden, and the best surface we have ever had

On the new supply.

Attention is moving, the same way Google Search changed it and the walled gardens came. This is the new search, and it will hold the best data on anyone's interest. The thing knows everything about Searby, every interest, his kids and details. So it is not only a new surface, it is the best surface we have ever had to gain attention, and dollars will flow that way. But you still need a good product and something people want, because the human comes back into the loop to buy. Retail and commerce media networks and open web advertising remain, just reached by a slightly different path, and the trust layer with the agent becomes critical.

We handed consumers a powerful tool, and the fraudsters the same one

On the fraud entry point.

Fraud has always been a cat and mouse game, but we just handed consumers an incredibly powerful tool, and we handed the fraudsters the same one. It became a very low entry point to spin up a fake merchant and a legitimate-looking site that promises shoes it never ships and disappears overnight, including in the ad space to pull attention to scam merchants. The old customer journey had mechanisms we expected over time, and all of that was compressed and replaced. Searby's answer is to link the traffic, device, identity and payment levels so fraud can be identified, with collaboration and integration at the core.

On the cybersecurity lens.

Searby comes from cyber intelligence, where the job was to hold the most data on what actors do and layer defences and controls so no incident slipped through. He applies that to ad tech: understand how the bad guys and the good guys operate, then decide where to add friction and where to add flow, with enough intelligence to inform how you present the ad, win the traffic, convert and enable the transaction. Security has long had rigid tools to verify you are who you say you are, and that now has to extend to hard verification of the human behind the agent and of the agent itself, moving left towards attention.

Marketing and security ignored each other too long

On convergence and the control plane.

Searby's buyer used to be the CISO, and it is now closer to the whole C-suite with a CMO lens, one side thinking about risk and the other about growth. His answer is to get close to customers' workflows and integrate where people work, in Adobe and Google Analytics and with the DSPs and SSPs where the full signal sits. He is blunt that he is not the solution, he is a piece of this puzzle, giving data and controls, redirecting traffic and acting as an integrated control plane. Teams ignored each other for too long in the tech stack, he says, and this is overdue. HUMAN's own priorities centre on its Agentic Trust offering and a new, data-driven approach to ad verification, with a partnership with OpenX announced that week.

On the advice.

For a marketer listening now, Searby's counsel is simple: get visibility first. Do you really see what is coming to your site and viewing your ads, and can you measure it? That may mean picking up the phone and calling your peers, because the data often sat with the CISO while marketing never knew it existed. Then develop the policies, then the controls and ongoing measurement, and be demanding of the platforms and agencies you work with. For the next generation he is just as direct: keep up with the speed of innovation, roll up your sleeves and try it, and get a minor in AI, because it is a reality of the world now.

Hear the full conversation. Follow Business of Marketing.

10 · The board question

Do we actually know how much of the traffic hitting our site and our ads is automated, and have we decided, agent by agent, which ones to welcome and which to block?

The Rewrite

Stay close to the conversation.

The conversations and intelligence worth knowing about.

About the guest

John Searby

Chief Strategy Officer, HUMAN · The Agentic Surface

John Searby is Chief Strategy Officer at HUMAN, the security business that sits in traffic on advertisers' and merchants' websites and across the advertising ecosystem to protect the consumer journey from bots and invalid traffic. He is known for bringing a cybersecurity lens to advertising and commerce at the moment agentic AI is walking the open web, and for HUMAN's Agentic Trust offering and its move into data-driven ad verification.

Episode page Full profile LinkedIn (opens in a new tab)

More interviews

All interviews
Share