AI Agents, Protocol Wars, and the Future of Shopping: Inside Agentic Commerce with Botify

Is agentic commerce the next big thing, or just hype?

 

In this interview with AJ Ghergich, Global VP at Botify, we break down the emerging landscape of agentic commerce, what it means, and how brands should adapt to win in this new era even amid the ‘protocol wars’ between OpenAI and Google.

Agentic commerce represents a significant shift in the future of retail.

Chatbots are reactive, working in a turn-based way: you say something, and the chatbot responds, reacting to user input in a fixed workflow. Meanwhile, AI-powered shopping might personalize recommendations or curate options, but you, the human shopper, make the actual browsing and buying decisions.

Agentic commerce is fundamentally different. The entire shopping experience rests with AI agents. It’s goal-based, powered by fully autonomous agents that own every step from discovery to purchase to returns and even subscription management. These agents don’t just respond like chatbots; they proactively plan, sequence steps, and crucially, use tools (like APIs) to execute. We’re not technologically there yet, nor are consumers, but this is where the retail industry is heading.

The industry is not quite there when it comes to fully autonomous, end-to-end transactions. The research and discovery phases are moving quickly—shoppers already use conversational AI for product discovery but the transactional side remains a major challenge.

With OpenAI’s Instant Checkout, it was clear that the discovery phase worked, but actual transactions didn’t scale. There are many real-world complexities, like real-time inventory, sales tax integration, and fraud detection—all the unglamorous but absolutely critical logistics for agentic commerce to work.

Plus, there isn’t enough adoption among retailers and merchants yet. Think about the launch of electric vehicles. The vision, direction, and interest were there, but the charging-station infrastructure wasn’t. It doesn’t mean the EV vision failed; it means we put the cart before the horse. For OpenAI, they proved people want and will use AI for discovery and research, but the underlying transactional and infrastructure stack needs to catch up before agentic commerce can really deliver on the full promise.

Right now, we’re seeing what I’d call a “protocol war”—Open AI’s Agentic Commerce Protocol (ACP) vs. Google’s Universal Commerce Protocol (UCP). Both ACP and UCP are designed to be a universal language for agents to communicate with existing tech stacks, similar to how we have one standard for apps to communicate via APIs. The big players are jockeying for the technical framework that will win and be used by the masses.

Amazon, interestingly, is holding its cards close and hasn’t come out to support ACP or UCP or introduced its own protocol. They’re showing signs of leaning into agentic commerce, but crucially, they’ve also walled off a lot of their content, blocking outside AI platforms from finding and crawling it. If you don’t have access to Amazon, Target, Walmart, and a few others, you’re missing the heart of retail. Ultimately, the winning protocols will need buy-in (or at least access) from these giants to come out on top.

 

AI-powered research and discovery isn’t going away—full stop. Consumers are embracing LLMs, and there’s no turning back.

But for brands, being crawlable to search and AI platforms is no longer enough. Behind the scenes, structured data and product feeds become even more critical. No matter which technologies win the protocol war, agents will need clean, structured, up-to-date data from retailers’ product feeds to perform discovery and (eventually) transactions.

Product feeds need to be highly structured, AI-optimized, and adaptable, or brands risk being left out of AI-driven recommendations. Think of feeds and structured data as the sitemaps of agentic commerce: they’re foundational and agnostic to who wins the standards battle. And this is why we just launched Botify Agentic Feeds, to automate the creation and delivery of AI-ready, protocol-compliant product feeds. With Agentic Feeds, retailers can ensure their products are always accessible and compelling to AI agents, whether by enriching feeds with reviews and Q&A content or by adapting immediately as protocols evolve. The brands that win will be those that become the trusted data source the agents turn to.

More broadly speaking, I also think AI visibility as a KPI will stick. Retailers want to know: “Is my brand or product appearing in AI-powered shopping experiences?” That’s different from traditional rank tracking, and it’s rapidly becoming top-of-mind for CMOs.

 

To remain competitive, brands must focus on infrastructure and feeds that power AI responses. What OpenAI tried to do with Instant Checkouts will likely not become a reality this year—the technology must mature, and consumer trust must grow. But so much can be done today to ensure your brand has the right foundation and competitive edge to succeed, no matter where the industry goes next.

 

If AI agents are crawling retail sites more aggressively, what new opportunities and risks does that create for marketing and e-commerce teams, and how should they adapt their crawling/traffic policies?

 

The explosion in bot traffic is massive. Retailers experienced +5.4x increases year-over-year in AI bot visits, and that’s on top of the massive growth we saw the year prior. Even more notably, for every one OpenAI user session, brands see an average of 198 bot crawls—for Google, it’s one session per six crawls. What that shows us is that discovery isn't necessarily happening on brand sites anymore. For marketing teams, the risk comes from failing to adapt measurement. Traditional metrics, like website traffic, no longer capture the full picture. As consumers discover products directly through AI platforms like ChatGPT, they may never visit your site until they’re ready to purchase. Without understanding where and when you appear in AI search, and without optimizing for it, you miss opportunities to strengthen visibility and generate new revenue.

More broadly, the upside is that AI platforms can be seen as a new discovery channel with massive potential. Every time a new discovery platform has emerged, like social in the 2010s, early adopters have won market share. The brands investing in high-quality, structured product data and robust site infrastructure today will be the ones that win as these agentic channels mature. Retailers should monitor bot policies, ensure their site is crawlable and data-rich, and use the opportunity to outmaneuver slower competitors.

 

Brand leaders are excited first by the promise of a new, measurable revenue channel and the chance to outpace competitors by adopting new, AI-driven models early. And the savvy CMOs realize that you can’t build visibility in AI search (AEO/GEO) unless you already have a strong SEO foundation. The promise is that investments in search visibility now compound across both human and AI/agentic channels. It’s exciting to know that so much of what we’re already doing by traditional means, like SEO, will still have gains tomorrow.

But there’s also a lot that keeps leaders up at night. Loss of control over the customer journey is one of the biggest concerns. For years, CMOs have painstakingly mapped personas and multi-step journeys. Now, the 12+ step journey is collapsing to one or two, and it’s not even a human making the purchase. Suddenly, the customer is an algorithm, using its own tools and reasoning to make decisions. 

 

It’s forcing CMOs to wrestle with tough questions: How do I serve an AI customer? How do I shape its opinions or decisions? Are we moving fast enough as an organization? To this last question, the answer is almost always no. The remedy is to focus on the foundational strategies that will produce results, no matter which standards or platforms win: structured data, feeds, and infrastructure.