As AI agents, crawlers, and automated traffic surge online, marketers are confronting a more complicated question than simple fraud detection: who...or what...is actually interacting with their brand?
At last month’s POSSIBLE Conference in Miami, one of the more unsettling conversations wasn't about streaming economics, retail media, or AI-generated creative. It was the reality check I got recording an episode of Insider Interviews with Jay Benach, about whether brands can trust that the audiences they think they're reaching online are even human.
Benach, GM, Media Security at Human, described a digital ecosystem where bots don’t just inflate traffic numbers or trigger fake ad impressions. Now they expertly shop, scrape, click, compare prices, fill carts, and interact with websites in ways that mimic real consumer behavior. That means influencing retargeting data, just as marketers are pouring bigger investment into automated buying systems and AI-driven optimization. And that means it’s harder to distinguish between human and machine activity and judge outcomes.
Viewability had been a bit of a contentious point, so standards were introduced to solve a legitimate issue: ads loading below the fold or appearing onscreen too briefly to be seen. Now, “viewable doesn’t necessarily mean human,” Benach said. Just as standards evolved, fraudulent actors adapted.
Bots have been programmed to scroll pages, move cursors, and create behavioral signals specifically designed to satisfy those measurements. Technically, the ad was “seen.” Just not by anyone a marketer actually wanted to reach.
From “Bot or Not” to Human
Human’s own evolution reflects how the problem itself has changed. Initially launched as “Bot or Not, LLC,” later “White Ops,” it landed on “Human Security” because the mission expanded beyond just fraud detection. The goal now is helping brands identify the real and right humans interacting with their business. Or, as Benach plainly pitched me, “"Do you want to know if you're being robbed? If the answer is yes, there is a solution to that."
Benach came to the category through an unusual route: running a video game ad network where certain publishers seemed able to generate suspiciously massive audiences almost overnight. “Something didn’t make sense,” he recalled. Then, meeting the founders of what was Bot or Not, he was sold on their mission to determine with precision whether a browser session was being operated by an actual person or by automation.
At the time, even acknowledging widespread bot activity in advertising made many companies uncomfortable. Today, the conversation has become far more nuanced because automation itself is no longer automatically viewed as malicious.
"The question is no longer 'Is this a bot?'" Benach said. "It's 'Who sent it, and what does it want?'"
That shift reflects the reality of today's internet. There’s a but about bots: Not all are harmful. Search indexing bots help organize the web. Shopping comparison bots may act on behalf of real consumers researching products. LLM crawlers continuously scrape websites so AI platforms can provide current answers and conversational experiences that feel timely and informed.
The “big but” is that malicious automation has become more sophisticated and more commercialized. Benach described "sniping bots" that intentionally fill shopping carts to distort retargeting signals, automated systems designed to drain competitors' paid search budgets through repeated clicks, and networks built specifically to manipulate campaign metrics. The result? Marketers are operating within a mix of humans, AI agents, useful automation, and outright fraud.
Instead of simply blocking all automation, brands have to try to understand which automated activity represents fraud, which represents opportunity, and how to optimize for both. Benach refers to that as a new "vertical learning curve" for marketers as they adapt to increasingly machine-mediated customer journeys.
"How do you serve the good and right bots?" he asked.
Why Audience Quality Matters More
The irony is that automation itself is not inherently the problem. One of the more interesting ideas Benach raised is that certain bots may eventually represent valuable intent signals. A shopping assistant bot operating on behalf of a real consumer, for example, could indicate stronger purchase intent than many traditional forms of browsing behavior.
If impressions can be inflated, clicks manipulated, and engagement simulated, audience quality suddenly becomes more valuable than audience volume alone. That is contributing to renewed advertiser interest in premium, transparent environments, particularly within professionally produced streaming and video content.
Across the industry, marketers are placing greater emphasis on contextual quality, verified audiences, attention metrics, and brand safety rather than relying solely on raw reach. In many ways, the conversation resembles a broader recalibration around trust.
The Next Advertising Battleground
Benach is particularly bullish on advertising opportunities emerging inside LLM-driven environments. Unlike banner ads or sponsored search results that consumers have largely trained themselves to ignore, AI-generated conversations create a more embedded and interactive form of engagement. Recommendations appear within active dialogue, often during moments of high intent and discovery.
"If I can target my spend toward the highest level of intent and precision, that's incredibly valuable," he said.
In that sense, the next evolution of advertising may not simply be about better targeting. It may involve understanding how brands communicate inside machine-mediated conversations where AI agents increasingly shape discovery, comparison, and decision-making before a consumer ever reaches a website.
This POV reframes one of marketing's oldest assumptions: that the audience is always human. Increasingly, the path between brands and consumers may include automated intermediaries gathering information, filtering choices, and influencing decisions long before a person ever sees the message directly.
If it wasn’t enough to learn that your web visitors might be faux humans then check out my previous “POV:Possible” conversations, with “the COOL company” which might make you rethink your ad tech stack, or with the VAB, which might spur another look at your media plan. Next, just wait till you hear what Alembic has to say.
The good news? All these companies are solving for today’s very modern media problems.
Posted at MediaVillage through the Thought Leadership self-publishing platform.
Click the social buttons to share this story with colleagues and friends.
The opinions expressed here are the author's views and do not necessarily represent the views of MediaVillage.org/MyersBizNet.
•