The news: AI-referred visitors to retail websites converted at higher rates, engaged more deeply, and generated more revenues per visit than those from non-AI sources in September, according to Adobe’s latest AI Traffic Trends Report.

• AI traffic converted at a 90% higher rate than non-AI traffic, the largest level since Adobe began tracking the metric.
• Those who visited from AI sources spent 73% more time on retail sites, viewed 25% more pages per visit, and added products to their carts at a 53% higher rate than those coming from non-AI sources.
• The gap in revenues per visit between AI and non-AI traffic widened in September, with AI traffic generating 83% more revenues per visit.

Behind the numbers: While AI referrals account for a miniscule proportion of overall visits to retail channels, their influence is growing. AI traffic to US retail sites jumped 71% YoY in September and is expected to rise significantly during key shopping periods like the Cyber Five, per Adobe. That is both an opportunity—mainly for retailers that rank highly on AI visibility—and a challenge, as brands struggle to ensure their sites, product pages, and overall shopping experience are optimized for the AI-influenced customer journey.

One major challenge that retailers face as AI-assisted commerce gains traction is consistently measuring attribution. AI-driven discovery is the biggest attribution challenge for US marketers, according to a survey by EMARKETER and Partnerize. A recent PYMNTS report found that none of the surveyed merchants could consistently connect AI-driven visits with completed purchases. Most struggled to consistently identify visits originating from AI sources, while others were unable to associate those visits with transactions. Despite these limitations, companies are prioritizing AI visibility to get their products in front of as many shoppers as possible. But lack of insight into whether those investments are performing could make it harder for retailers to determine which AI initiatives and platforms deliver meaningful sales.

Retailers must also contend with how to balance being prepared for agentic commerce with having adequate safeguards to prevent unauthorized or unwanted activity. Secondhand shopping platform The RealReal was recently forced to update its bot defense system to prevent Meta’s Muse from using its reservation system to keep other shoppers from purchasing a product for hours at a time, per The Verge. And Home Depot is restricting agents’ access to its site until they can prove they are authorized and connected to a specific shopper, Angie Brown, the company's executive vice president and chief information officer, told Business Insider.

Implications for retail: Third-party AI platforms are expected to play a larger role in holiday shopping this year as consumers seek inspiration, guidance, and deals. Adobe expects AI’s revenue-per-visit advantage to double during the holiday season, increasing the need for data accuracy, completeness, consistency, and LLM readability on brand and retail channels.

But visibility is only part of the battle. Retailers also need to think about how to reinvent the shopping experience on their sites to make it easier for agents to shop on customers’ behalf while preventing unwanted or unauthorized activity. Companies will need to address the need for better AI attribution to better determine which AI platforms are worth prioritizing and where investments produce the highest returns.