• 4 min read
While some brands are more proactive than others, agencies that represent brands in sensitive categories are monitoring against “bad” bots and looking to leverage more useful ones.
Getty Images
On June 3, bots made up 58% of all worldwide traffic to HTML pages, according to data compiled by the network security provider Cloudflare. Human-generated traffic, meanwhile, accounted for 42%.
“Welp, that happened faster than I predicted,” Matthew Prince, Cloudflare co-founder and CEO, wrote on X at the time. “Agentic traffic [is] growing so fast that bots have now passed human traffic online for the first time in the Internet’s history.”
It hasn’t exactly been a secret that bots were slowly taking over the internet; automated traffic overall across the web increased nearly eight times faster than human activity last year, according to a report from cybersecurity firm Human Security. “Bad bots,” aka those that execute actions ranging from data scraping to transaction fraud, made up 37% of internet traffic a year ago, according to an April 2025 report from the network security firm Imperva cited by the Video Advertising Bureau in a February report.
What’s a marketer to make of the sea change? Jeff Eisenfeld, director of activation at the agency Media by Mother, told Marketing Brew that there’s still a ways to go before most clients are considering all—or any—of the potential ramifications.
“There is going to be a whole AI cybersecurity concern movement when it starts to get [to be] really mature AI and folks want to leverage it for the wrong reasons, but [at] this moment, it hasn’t been top of mind for our clients,” he said. “A lot of brands are still understanding how to use AI and what’s the importance. A lot of folks are still at that beginner infancy stage, where they’re not concerned about cybersecurity because they haven’t really started to even test the capabilities of AI.”
With that said, some agencies and brands are leading the charge as they think about ways to protect against bad bots—and optimize for better ones.
But there’s at least one vertical for which filtering bot traffic is a “paramount” concern, Eisenfeld told us: financial services. At Media by Mother, a bot traffic-filtering tool helps companies that collect sensitive customer information to prevent malicious bots from accessing sensitive financial information. The agency similarly monitors financial service brands’ lead forms, where consumers input personal information; in those instances, traffic verification tools like DoubleVerify can be particularly useful, Eisenfeld told us.
But not all bots are bad bots. At Omnicom, executives are thinking about content optimization holistically, making brand content accessible “for any touchpoint,” whether it’s LLMs or web crawlers and other bots that could affect consumer discovery, Paolo Yuvienco, the company’s CTO, told Marketing Brew. It’s an imperfect system: updating web content to make it more likely for an LLM to ingest information from it doesn’t necessarily work every time, since many LLMs have training runs that end. Some of those models, however, have crawlers that ingest data upon the point of query, he noted—and in those instances, brands can optimize for specific LLMs.
“The contextual relevance is different than the traditional SEO listing of Google,” he said. “It becomes far more contextual because it parses the language through its LLM.”
The effects of bots on the internet have other brands taking a more serious look. Amazon is suing Perplexity over Perplexity’s shopping agent Comet, which could automatically place orders on e-commerce sites; The company Human Security aims to guard against agent fraud. And Parallel Web Systems, an AI agent infrastructure startup founded by former Twitter CEO Parag Agrawal, rolled out Index, a product that aims to compensate certain publishers and creators whose content is crawled by non-malicious bots. Bots can also sometimes access paywalled content, another concern for publisher brands aiming to protect their businesses.
“If you believe that agents will use the web 1,000 times more, that the constraint is shifting from human attention to maximizing value and using compute effectively on all the data you have available,” Agrawal said in an interview published in Stratechery last month. “The big challenge we’ll encounter is ‘Does everyone have incentive to participate?’ Or will a bunch of content owners just opt out of this cooperative game?”