Bots are becoming customers: Why businesses need to rethink blocking automated traffic

For several years, most businesses have viewed Internet bots as a problem. Automated traffic has often been associated with credential stuffing, price scraping, denial-of-service attacks, and other cybersecurity risks. The standard response has been straightforward: identify the web bot and block it.

According to new analysis from web data specialist Decodo, automated traffic now accounts for 57.4 percent of all web requests, overtaking human traffic, which represents 42.6 percent. More significantly, AI-driven traffic is growing at an extraordinary pace, raising a challenge that many organizations have barely started to address: what happens if blocking bots also means blocking potential customers?

The rise of generative AI, AI assistants, and autonomous agents is creating a new form of digital visitor. Increasingly, software is acting on behalf of people—researching products, comparing prices, assessing services, and gathering information before a purchasing decision is made. According to Decodo’s analysis, agentic AI traffic increased by 7,851 percent year-on-year, while overall AI-driven traffic grew by 187 percent during 2025, approximately eight times faster than human traffic.

Historically, online commerce has followed a relatively simple model. A consumer enters a query into a search engine, reviews the results, visits a website, and makes a purchase.

Artificial intelligence is beginning to disrupt this process.

Instead of visiting dozens of websites directly, consumers are increasingly turning to AI platforms to perform research on their behalf. Whether using AI-powered search tools, shopping assistants, or integrated digital agents, users are outsourcing information gathering to software.

This means that websites are no longer being accessed solely by humans. They are being visited by intelligent intermediaries acting on behalf of humans.

As Vaidotas Juknys, CEO of Decodo, explains in a statement sent to Digital Journal: “The instinct to block every bot is understandable, but it’s like locking your storefront because some visitors don’t buy. The agents crawling your site today might be how your customers will discover you tomorrow.”

The implication is profound. Businesses may need to view some categories of automated traffic as opportunities rather than threats.

AI agents are already reaching commercial decision points

The Decodo findings indicate that AI systems are not merely exploring websites casually. They are interacting with the areas that are most closely linked to customer purchasing decisions. The pattern suggests that AI agents are increasingly evaluating products, comparing offerings, and gathering information from the same pages that human consumers traditionally use when making purchasing decisions.

For retailers, this means product descriptions, pricing information, stock availability, and technical specifications may increasingly need to be optimized not only for human visitors but also for AI systems interpreting and presenting that information elsewhere.

Retail is becoming AI’s preferred hunting ground

The greatest concentration of AI-driven activity appears within retail and eCommerce. Decodo found that retail and eCommerce represented 62.5 percent of training-crawler traffic during 2025. The company’s network data also identified retail as the largest commercial category for successful automated requests.

Retail websites contain highly dynamic information such as product prices, inventory level, and product reviews. These datasets are extremely valuable to AI systems attempting to answer consumer questions accurately. In contrast, sectors such as travel, aviation, real estate, jobs, and media attracted far smaller proportions of successful automated traffic.

The trend suggests that online retail may become one of the first industries where AI-mediated discovery becomes mainstream. However, businesses still face genuine risks since not all bots are beneficial. This means organisations continue to deal with credential-stuffing attacks, fraudulent account creation, web scraping and other risks.

Security teams therefore remain under pressure to restrict automated traffic wherever possible. The challenge is that traditional security approaches often treat all bots similarly. In an era of AI agents, that approach may become increasingly ineffective.

A blanket blocking policy could potentially prevent legitimate AI assistants from accessing information that customers have requested. This could reduce visibility within AI-powered search engines, shopping assistants, and recommendation platforms. The question is no longer whether a visitor is human. The more important question is whether the visitor is legitimate.

The consequences extend beyond cybersecurity. Marketing teams have spent years optimizing websites for search engines. A similar process may now emerge around AI agents. Product pages may need clearer structured data. Price information may need to be machine-readable. Content may need to be optimized for AI summaries and recommendation engines rather than solely for conventional search rankings.

The growth of AI agents signals a broader transformation in how consumers interact with information online. For two decades, businesses focused on search engine optimisation. The next competitive battleground may be ensuring that AI systems can accurately discover, interpret, and recommend products and services.