E-commerce
August 26, 2026
E-commerce is undergoing a silent but profound mutation: AI agents and large language models (LLMs) are beginning to influence the purchasing journey, from search to transaction. Growth figures make headlines — AI traffic up 4,700% in some segments — but they are starting from a tiny base. What really matters is channel sustainability, user satisfaction, and the infrastructure being put in place. Adobe, Salesforce, and Shopify confirm that traffic generated by AI assistants is holding steady week after week in 2026, beyond promotional peaks. When 85% of users state that AI improves their shopping experience and 73% use it as their primary search source, the behavior becomes structural. This guide decrypts the statistics that really matter to you, the e-commerce merchant: traffic, conversion, transaction protocols, and what you need to integrate right now.
Summary
AI traffic growth: impressive but still marginal
The figure of +4,700% growth in AI traffic is head-turning, and it is real. But it measures a channel that still represents well under 1% of total retail traffic in most datasets. Adobe Analytics confirms that traffic generated by ChatGPT, Perplexity, and other AI assistants is climbing fast, but from an almost zero base in 2024.
However, this curve should not be ignored. Shopify reports that its merchants are seeing traffic referred by AI agents arrive consistently in 2026, with steady weekly progress. Capital One Shopping notes that during Prime Day 2025, AI assistants generated significant traffic spikes in certain categories, particularly electronics and home. The real question is not the slope of growth, but sustainability: does this channel remain active outside of promotions? All sources agree: yes.
Convert over 2,000 customers on average per month with Qstomy.
The world’s 1st Shopify AI dedicated to customer conversion
Empowering 200+ e-commerce merchants
Conversion and AOV: results remain mixed
Does AI traffic convert better than other channels? The data varies depending on the source. Alhena, which analyzed ChatGPT and Perplexity in 2025, shows that traffic originating from these assistants displays a lower conversion rate than organic search and affiliate marketing, but higher than paid social media. The AOV (average order value) varies significantly by vertical: electronics and fashion are leading the way.
Kaiser and Schulze, in their peer-reviewed study of 973 retailers, observe long sessions and deep browsing, but lower conversion. Their interpretation: users are in the research phase, not yet in the purchasing phase. They compare, explore, and then return later through another channel to finalize. Salesforce, for its part, reports $14.2 billion in revenue "influenced" by generative AI in 2025, a broad attribution figure that counts any transaction where AI touched the customer journey.
User satisfaction: the real signal to watch
While conversion remains to be confirmed, user satisfaction is stellar. Capgemini Research Institute reveals that 71% of consumers want to see generative AI integrated into their shopping experience. Capital One Shopping reports that 85% of users who have tried an AI assistant for shopping say it improved their experience, and 73% now make it their primary source for product research.
These satisfaction figures are the strongest signal: when the experience is good, adoption becomes sticky. Users return, explore more, and eventually buy, even if it's not always in the same session. For you, the merchant, this means that your presence within these assistants, and the quality of your product data, becomes a critical lever for visibility and consideration. E-commerce SEO is evolving: it is no longer just about ranking on Google, but about being cited by ChatGPT, Gemini, or Claude.
Agentic commerce protocols: ACP, UCP, MCP
Agentic commerce is not limited to sending traffic: it aims to enable AI agents to perform transactions on behalf of the user, without ever leaving the conversational interface. Three protocols structure this landscape in 2026: ACP, UCP, and MCP.
OpenAI launched ACP (Agentic Commerce Protocol) on September 29, 2025, with Stripe, enabling Instant Checkout in ChatGPT for Etsy and Shopify merchants, for a 4% merchant fee. Google replicated on January 11, 2026, with UCP (Universal Commerce Protocol), co-developed with Shopify and supported by more than 20 retailers (Home Depot, Best Buy, Macy's, Visa, Mastercard). Anthropic, via the Linux Foundation, maintains MCP (Model Context Protocol), a general standard used by Salesforce, Adobe, and other integrators. tunnel-de-commande.com tracks the adoption of these protocols and notes that merchants view them as complementary, not competing.
Amazon and the exception of AI recommendations
Amazon remains the exception that proves the rule. About 35% of its revenue is attributed to AI-driven product recommendations, a ceiling that the rest of the industry does not even approach. During Black Friday 2025, RealityMine measured that ChatGPT traffic to Amazon was significant, but the majority of conversions still came from Google Search.
This dominance of Amazon reflects a mature data, personalization, and AI infrastructure built over fifteen years. For DTC or multi-channel merchants, the challenge is to build their own AI capability, or to rely on open protocols (UCP, MCP) to keep control of the customer relationship. Small brands under $100k/month can leverage these tools without a massive infrastructure, provided they invest in product data and conversational experience.
Cyber Week 2025: the advantage of branded agents
Salesforce and Digital Commerce 360 published one of the most telling statistics of 2025: during Cyber Week, retailers who had deployed their own AI agent grew 32% faster than those who had not. This gap is not just explained by external AI traffic, but by the compounding effect of an AI assistant capable of guiding, recommending, responding, and converting in real time.
Retail Dive, in its 2025 Black Friday review, highlights that the winners all have one thing in common: a seamless conversational experience integrated into the catalog and checkout funnel. The losers, on the other hand, did not invest in AI or restricted it to basic FAQ chatbots. The difference is measured in millions of dollars of GMV. Conversion rate optimization is now driven by conversational AI, not just the A/B testing of buttons.
Attribution methodology: vendor vs. academic
"AI-influenced" revenue figures vary wildly depending on the methodology. Salesforce counts $14.2 billion influenced in 2025, a figure that includes any transaction where generative AI touched the journey, even indirectly. This is a broad attribution, useful for measuring the global footprint of AI, but it does not reflect direct causality.
Kaiser and Schulze, in their SSRN study, adopt a stricter approach: they track transactions with first-party attribution across 973 retailers, and conclude that the direct contribution of AI remains modest, albeit growing. Both visions are true, but answer different questions. For you, the merchant, the important thing is to track your own data: set up UTM tracking on links coming from ChatGPT, Perplexity, or Gemini, and measure conversion, AOV, and LTV. Increasing traffic and conversion relies on the rigorous measurement of each channel.
AI Visibility: How to Be Cited by Assistants
Ranking well on Google is no longer enough: you must also be cited by ChatGPT, Perplexity, Gemini, and Claude. Erlin, an AI tracking platform, analyzed 500+ brands in 2026 and shows that citation depends on the quality of structured data, the richness of the content produced, and presence in sources that LLMs deem reliable (reviews, articles, technical sheets).
Concretely, this means: impeccable schema.org markup, detailed product descriptions, numerous and authentic customer reviews, and presence on authority sites. Skai, in its study on AI awareness among US shoppers, reveals that 64% of consumers trust an AI recommendation more than a display ad. If you are not cited, you are invisible. Mobile-first design and structured data are becoming levers of AI visibility.
AI traffic vs organic search: where do we stand?
AI traffic is growing fast, but it remains far behind organic search on all academic conversion indicators. Kaiser and Schulze note that ChatGPT sessions are long (5 to 8 minutes on average), with many page views, but a high bounce rate and low conversion. They interpret this pattern as research behavior, not immediate purchasing.
Digiday, analyzing ChatGPT traffic to Walmart in 2026, observes the same trend: lots of exploration, little direct checkout funnel. But users who return through another channel (direct, email, retargeting) convert better than average. Conclusion: AI traffic is a consideration and intent lever, not (yet) a direct conversion channel. Popular Shopify stores have understood this and are optimizing their AI presence to capture intent, then converting via email or retargeting.
Agentic commerce: the infrastructural bet of 2026, 2027
McKinsey, in its 2025 outlook on agentic commerce, estimates that the real shift will not come from AI traffic, but from the ability of agents to transact autonomously. When a user asks "buy me the best running shoes under €150," the agent will need to be able to compare, select, negotiate, and pay, without human intervention.
The ACP, UCP, and MCP protocols are the rails of this infrastructure. Merchants who integrate now are getting a head start: they are testing friction, adjusting their catalogs, and calibrating their margins. Those who wait until 2027 will have to catch up in an already structured market. tunnel de commande.com reports that early adopters (Etsy, Shopify, Home Depot) are already showing transaction completion rates above 60% on agentic checkout funnels, compared to less than 40% for new entrants.
Qstomy vs. generic AI assistants: strategic complementarity
Generic AI assistants (ChatGPT, Gemini, Perplexity) excel in broad search and multi-brand comparison, but they do not know your catalog in real-time, your inventory, your promotions, or your return policies. This is where Qstomy comes in: a specialized Shopify AI agent, it guides your visitors toward purchase by recommending the right products, offering contextual upsell and cross-sell, assisting with the cart, and handling package tracking and customer service.
Unlike generic LLMs that drive traffic but do not directly convert, Qstomy lives inside your store and optimizes every step of the funnel: from product discovery to customer service resolution. More than 100 Shopify merchants use it to reduce cart abandonment, improve customer satisfaction, and increase AOV. When a visitor arrives via ChatGPT or Google, Qstomy takes over to turn intent into a transaction. Complementarity, not competition: external assistants bring the attention, Qstomy converts and retains. Is Shopify still the best platform? Yes, especially with native AI agents like Qstomy that fully leverage the ecosystem.
In brief: what you need to remember and do now
AI traffic is exploding, but remains marginal. User satisfaction is high and the channel is stabilizing. ACP, UCP, and MCP protocols are structuring agentic commerce. Merchants with a branded AI agent are growing faster. Here is your immediate checklist:
Track AI traffic: UTM on ChatGPT, Perplexity, Gemini, and measure conversion and AOV.
Optimize your product data: schema.org, rich descriptions, and numerous customer reviews.
Integrate ACP or UCP: if you are on Shopify, explore agentic checkout options.
Deploy a branded AI agent: Qstomy or equivalent, to convert AI traffic into real sales.
Measure influence, not just direct attribution: AI often touches the customer journey without being the last click.
In short: conversational AI is not replacing organic search yet, but it is becoming an essential consideration channel. Invest now in AI visibility, transaction protocols, and a branded agent to capture and convert this new wave of qualified traffic.
FAQ: AI e-commerce statistics 2026
Will AI traffic surpass organic search in 2026? No, not yet. It is growing fast but starts from a very low base and remains below 1% of total retail traffic. Sustainability matters more than the slope.
Do AI assistants convert better than other channels? Not directly: they excel in the research phase, but the final conversion often comes from another channel (email, retargeting, direct). Their strength is influence, not last-click attribution.
Should I integrate ACP or UCP right now? If you are on Shopify and have significant volume, yes: early adopters show better completion rates and learn quickly. Otherwise, prepare your product data and monitor benchmarks.
Enzo
August 26, 2026
