Most retail brands are not ready for what agentic commerce means. The ones who think they have a few years to prepare actually only have a few quarters. Retail is becoming increasingly tech-first, led by the rise of agentic commerce and the solutions and tools that enable it. Product discovery has shifted from keyword-led retrieval toward AI decisioning, where assistants interpret intent, find options, and influence what a shopper considers before a brand site ever loads.
The data confirms what I’m seeing in the field. A joint research report from Retail Economics, AWS, Botify, and DataDome found that AI-driven bot traffic to retail and e-commerce websites increased 5.4x in 2025. It builds on similarly high traffic growth from the years following the mass-consumer adoption of LLMs, and it shows no signs of slowing. Bots and agents are becoming the first readers and recommenders of the product story.
Visibility without visits
Here’s the framing I keep coming back to when I talk to business leaders: brands are now the map, not the place. Your CMO probably still thinks of your brand as a place where a customer arrives. The agentic shopping path is killing destination thinking: the map decides what shoppers see, not the destination. If your team is still focused on the destination, you’re already behind.
It helps to stop thinking of AI bots as web crawlers and start thinking of them as a type of consumer. AI systems crawl with an objective to ingest, validate, compare, and summarize information within their own interfaces, often long before a click occurs, and sometimes instead of one.
This is quantified by the data. The above report found that Google drove roughly one visit for every six crawls, while OpenAI required 198 crawls to generate one human visit. When a consumer uses AI to search, they no longer have to click and evaluate — the assistant does it for them. Thus, AI search requires that bots find, see, understand, and summarize the same information, 33 times more often than they ever send you a visit. The brands with bot-friendly, machine-readable content now are building a presence in the model’s mental map of their category. The brands that aren’t are ceding that map to competitors.
It’s a compounding risk that’s vastly underestimated. The cost of being invisible to AI isn’t just lost traffic today, but the lagging understanding of your brand and category for every quarter you’re not in the index. Invisibility compounds, and there are no quick fixes to win back visibility.
Platforms are setting the pace, and consumers are following in step
Over the last several months, we’ve seen major moves from the giants — including protocols like OpenAI and Stripe’s Agentic Commerce Protocol (ACP) and Google and Shopify’s Universal Commerce Protocol (UCP), as well as the launch of Google’s Universal Cart. I have a view on where this shakes out, and I think it’s worth committing to it.
Google has the resources and infrastructure to bring traffic to retailers, as well as the merchant data that helps agents plug into the AI shopping layer. While OpenAI’s ACP is payments-focused, Google has the existing foundation for AI discovery, checkout, and tracking; the partnership with Shopify only strengthens their lead. Consumers still instinctively look to Google for shopping, and it’s the standard most brands I talk to are focusing on. If I had to put my investment anywhere, I’d make my bet on UCP winning the protocol wars.
That doesn’t mean brands should optimize for one protocol and consider it done. Instead, ask: what actions will have a universal benefit over the next 2–3 years as agentic commerce grows? The answer is conversationally optimized, AI-enabled product feeds. Product catalogs for AI search enable discovery and have utility across any ecosystem. That’s where brands should be focusing their activities today.
Across the US, UK, and France, 73% of consumers report using AI assistants, with 38% using them specifically for shopping tasks such as discovery, comparisons, and recommendations. AI is the shopping intermediary; what buyers learn, shortlist, and purchase increasingly depends on what AI agents can find and understand about your brand.
If bots can’t see it, shoppers won’t either
For many retailers, the biggest gap lies in AI-ready product feeds, which need to be ACP- and UCP-compliant. Legacy product feeds don’t provide the context signals and depth required by these new AI protocols. It can leave even the best products excluded from AI recommendations, a price paid in revenue, discoverability, and competitive advantage.
AI-compliant feeds, automatically built and enriched with reviews, Q&As, and other contextual data, help AI systems understand and recommend products to shoppers. Even this, however, is subject to the new measurement problem: the impact may not show up in the metrics you’re used to.
Redefine success, then govern access
Organic search metrics were built for a click economy. AI-driven product discovery is a zero-click environment, and its influence may not look like a visit even when it drives revenue. There are new questions that retail teams should ask:
- Are our products being surfaced by AI?
- Are we being cited in relevant prompts?
- Are we represented accurately?
- Are we present when AI assistants narrow down product choices?
New KPIs that map to this environment include:
- Overall AI visibility: The presence rate in percentages for how often your brand or URLs appear in AI-generated responses, across AI platforms.
- Discovery visibility: Overall visibility across multimodal environments.
- Structured data coverage: The completeness of metadata, schema markup, and product information accessible to AI.
- Intent visibility: How visible your brand is across different customer journey intents — awareness, research, and decision.
Additionally, as AI traffic grows, the line between beneficial agents and malicious bots blurs. DataDome found that nearly 80% of websites are not protected against agent spoofing, exposing retailers to distorted analytics and inflated referral signals. Treat bot and agent traffic as a channel that needs governance; trusted bots and agents should be able to access what they need, with the benefit of greater visibility for the brand.
What changes now
AI-driven discovery changes where and how visibility translates into growth. Retailers now compete inside a discovery layer that reads first, summarizes next, and sends traffic last. Being crawlable is table stakes. Being machine-readable, technically accessible, and trusted is what determines whether the agent includes you in the conversation at all.
In the next 90 days, your task is to audit your product feed against ACP and UCP, stand up an AI visibility tracker, and ask your teams for an answer on agent readability. If they don’t have one, that’s the answer to the bigger question, too.
The brands that adapt now will protect their visibility and convert more consistently — even when customers never open a browser tab. Those that don’t will spend the rest of the year scrambling to regain ground they didn’t know they were losing.
