Salesforce launches AI shopping agents with ChatGPT

The economic case is already visible. Salesforce said AI influenced 20% of global online holiday sales in 2025, equal to $262 billion, while retailers running their own shopper agents grew sales 59% faster than those that did not. AI-referred traffic also converted at eight times the rate of social traffic, underscoring that the new channel is not just generating clicks but actual purchases.

For Salesforce, the launch is more than a product update. It is an attempt to position Agentforce Commerce as the operating layer between external AI interfaces and a retailer’s own data, inventory and order systems. In practice, that means a shopper who starts in ChatGPT or Gemini can be treated as the same known customer once they arrive on a brand’s site, app or messaging channel. That continuity is crucial for retailers trying to protect margins, reduce abandonment and avoid handing the customer relationship to platform owners.

The broader shift also helps explain why the story matters for the software and cloud stack around retail. If AI assistants become the front door to shopping, then the most valuable systems will be the ones that can verify stock, enforce contract pricing, promise delivery dates and resolve service issues in real time. Salesforce is betting its heritage in customer records, workflow and enterprise data gives it an advantage over lightweight chatbot add-ons.

Investors are likely to read the release as both opportunity and challenge. The bull case is that Salesforce can expand its commerce footprint as agentic AI drives more transactions through owned channels, deepening platform stickiness and supporting higher-value software sales. The bear case is that the company is entering a crowded race in which Google, OpenAI and others may control discovery, while retailers remain cautious about handing too much autonomy to software that can make pricing or fulfillment errors.

The stock action shows the market is still weighing those cross-currents. Salesforce shares have been volatile, and while its longer-term technical backdrop remains healthier than the sector’s weakest names, the recent pullback leaves investors focused on whether AI commerce can translate into measurable revenue acceleration rather than another marketing-led product cycle. For software peers such as ServiceNow, the message is similar: AI is moving from copilots to execution, but monetization will depend on whether enterprises trust these systems with real commercial decisions.

The next test will be adoption through the 2026 holiday season, when brands will need to prove that agentic commerce can lift conversion without sacrificing control, economics or customer trust. If it works, Salesforce could own a strategic layer in the next phase of digital retail; if not, AI commerce may remain a compelling demo rather than a durable growth engine.