Salesforce, a cloud software company, recorded traffic referred from AI chats rising between 150% and 428% year over year in every quarter measured from the first quarter of 2024 through the first quarter of 2026. Overall commerce traffic grew by single to low double digits, according to its fourth “State of Commerce” report.
Use of “agentic search” as the first step in a shopping journey increased 200% year over year. Salesforce used the term for shoppers beginning with a question in a large language model conversation or an AI assistant on a retailer’s site. The survey found 28% of commerce organizations used agentic AI, 44% planned adoption within six months and 86% expected large language models to become essential to product discovery within a year.
“A customer can adopt AI overnight. A business has more to get right: trusted data, connected systems, teams ready to use them,” Caila Schwartz, Salesforce’s head of agentic commerce shopper insights, said in the research summary.
Businesses lag behind in data readiness
Only 27% of surveyed organizations reported fully unified customer data across sales, service, marketing and commerce, while 32% had fully defined AI success metrics and key performance indicators.
Among organizations with fragmented customer views, 40% reported slow or ineffective responses to customer issues and 39% struggled to measure commerce investments. Multichannel organizations most often cited inconsistent pricing and promotions, at 41%, and inventory that was not synchronized in real time, at 40%.
Meeting platform data requirements
Platform requirements detail the operational work. OpenAI, an AI developer, limits product-feed onboarding to approved partners and asks them to provide structured catalog data including titles, descriptions, images, prices and availability. Its guidance generally recommends sending the entire feed once a day by file upload, then using the API for updates during the day.
Google uses a separate Universal Commerce Protocol that includes transaction requirements. Among the steps to go live, merchants must configure product feeds, shipping and returns, implement three core REST endpoints for creating, updating and completing checkout sessions and push order-status updates to Google.
Integrations require approval before going live in AI Mode and Gemini. Google also notes not every feature in the evolving standard is supported on its services.
Adobe reports a surge in AI conversions
Adobe recorded AI-originated traffic to U.S. retail sites rising 393% year over year during the first quarter of 2026. AI-referred visits converted 42% better than non-AI traffic in March, according to Adobe Digital Insights data covering more than 1 trillion visits.
A separate Adobe analysis using its proprietary AI Content Visibility Checker gave individual product pages an average score of 66%, compared with 75% for homepages. Adobe’s public report and accompanying blog did not disclose how many sites or pages were scored.
Research scope and survey methodology
Salesforce’s traffic analysis covered more than 1.5 billion shoppers across 37 countries. Its organizational findings came from a separate survey of 3,450 commerce professionals across 20 countries and 13 industries, while a third survey covered 4,689 consumers in eight countries.
Salesforce’s organizational findings were self-reported and did not establish that unified data caused greater AI visibility, conversion or revenue.