AI Shopping Agents at Cyber Week 2026: How Salesforce Commerce Cloud Brands Should Prepare

AI shopping agents at Cyber Week 2026: shoppers' personal agents compare prices, check stock and read reviews, so Salesforce Commerce Cloud brands should get product data, prices, inventory and agent access rules ready before Black Friday, by WeVerve Systems

AI Shopping Agents at Cyber Week 2026: How Salesforce Commerce Cloud Brands Should Prepare

Picture a shopper on Cyber Monday who asks an AI assistant to find the best price on a stand mixer, check whether the reviews recommend it, and flag anything that will not arrive before 20 December. Then they close the app and buy it themselves. Whether your store makes that shortlist depends on what the assistant can read about your products.

That scene opens new Salesforce research published on 12 November, and it is a fair summary of where holiday shopping is heading. AI shopping agents are becoming a real source of traffic, and in the last week a new open standard for how those agents identify themselves to stores reached version 1.0. This guide explains what is changing, what it means for brands on Salesforce Commerce Cloud, and what to fix and test before Black Friday on 27 November and Cyber Monday on 30 November.

Key takeaways

  • Salesforce projects that AI will drive $75.8 billion in Cyber Week spending, up 13% from 2025, and that traffic to ecommerce sites from AI tools is up 104% year over year.
  • Most shoppers use personal agents as advisors, not buyers. Only 19% of personal agent users have let an agent check out for them.
  • Sierra released version 1.0 of the Personal Agent Protocol (Poppy) on 13 November. Salesforce has joined and plans to bring it to Commerce Cloud on top of its existing UCP support.
  • The work that matters before Cyber Week is accurate product data, current prices and inventory, stable site performance under agent traffic, and clear rules on which agents you trust.
  • Shoppers forgive few mistakes. Most will abandon an AI assistant after two or three errors, so test your own shopper agent with hard questions now.

What the latest data says about AI shopping agents

Salesforce surveyed more than 2,500 shoppers in the US, UK, Canada and Australia for its holiday research. The headline is that AI is now part of how ordinary shoppers start their search, not only early adopters.

  • Almost 20% of shoppers say AI is their first stop on the way to the store, and 74% trust the recommendations it gives them.
  • 30% of AI shoppers say these tools often introduce them to brands they did not know.
  • 11% say they would skip a purchase if an AI assistant advised against it.
  • Shoppers who arrive through AI referrals are more than two times as likely to convert as typical ecommerce shoppers.

Salesforce’s peak holiday predictions go further. It expects 20% of all 2026 holiday ecommerce traffic to come from AI chat agents, and one in three ecommerce sites to offer their own site specific shopper agent by Cyber Week. These are Salesforce forecasts based on its own Shopping Index, so treat them as a direction of travel rather than a guarantee.

Advisors first, buyers later

The survey also shows clear limits. 31% of consumers have tried a personal AI agent, but only 19% of those users have used one at checkout. Just 9% give their agent unrestricted access to their financial details, and 73% of shoppers are not comfortable giving a personal agent their payment information. For this Cyber Week, the agent is mostly doing research. The human still presses the buy button.

That matters for priorities. If an agent reads your catalogue and reports a wrong price or a missing size, the sale is lost before the shopper reaches your site.

Personal agents and brand agents are not the same thing

Two kinds of AI agent will interact with your store this season, and they need different preparation.

Personal agents work for the shopper

A personal agent belongs to the shopper and works across every brand they buy from. It compares prices, reads specifications and reviews, and checks delivery. You do not control it, but you control what it finds: your product data, your prices, your stock and how your site responds to it.

Brand shopper agents work for you

A brand shopper agent runs on your own storefront. On Commerce Cloud this is Carter, the Shopper Agent in Agentforce Commerce, which Salesforce made generally available in June alongside its Buyer and Merchant agents. According to the Commerce Cloud innovations page, product comparisons are available now, while cross-sell and upsell suggestions, page awareness and authenticated shopper context (past orders and saved preferences for signed in shoppers) became generally available in October 2026.

Comparison of a shopper personal AI agent and a brand shopper agent such as Salesforce Carter: who it works for, where it runs, what it does, and what the brand controls
Two kinds of agent will touch your store this season. You control one directly and set the rules for the other.

Salesforce reports that 39% of shoppers have used a brand’s shopper agent, and 80% of them had a good experience. Shoppers who engage with one show an 8% higher average order value and buy 23% more items than the average order. Those are Salesforce’s own figures from its research, but they show why the brand agent is worth getting right.

The new standard: how agents will knock on your door

Until now, agentic commerce has mostly grown one integration at a time, with each AI platform connected separately to each brand. On 13 November, Sierra released version 1.0 of the Personal Agent Protocol, known as Poppy, an open standard for how personal agents work with companies through their websites, APIs and business agents. It is built largely on existing standards such as MCP, OpenAPI and UCP, and uses OAuth so customers control what their agent can do. Sierra says more than 100 companies are now design partners, and Meta’s Muse agent automatically looks for Poppy first on every site it visits.

Where Commerce Cloud fits

On 16 November Salesforce confirmed it has joined Poppy and explained how the pieces fit. Poppy covers the first steps of an agent’s visit: finding the store, proving its identity and, with the shopper’s consent, signing into their account. For cart, checkout, payment and orders, it hands off to the Universal Commerce Protocol (UCP), the open standard built by Google with industry partners.

Flow of a personal AI shopping agent visiting a store: the shopper asks their agent, the agent finds the store and identifies itself, the shopper approves access by signing in on the brand login page, cart and checkout run through the Universal Commerce Protocol, and the order lands in the brand systems with the brand as merchant of record
How the two standards split the work, as Salesforce and Sierra describe it: Poppy covers arrival, identity and consent, and UCP covers cart, checkout, payment and orders.

UCP is already part of Commerce Cloud. The UCP integration, generally available since October 2026, lets shoppers buy directly on Google surfaces such as the Gemini app and AI Mode in Search while the brand stays the merchant of record. Poppy support is not live yet. Salesforce says its intent is to bring Poppy to Commerce Cloud on top of UCP and that it will share more as plans take shape.

What stays in your control

  • Which agents get in: Salesforce says brands should be able to welcome trusted agents, limit access by use case or block agents that do not meet their standards.
  • Shopper privacy: with Poppy, a store sees a private ID for each shopper, so it can recognise a returning customer but cannot compare IDs with other companies.
  • Account access: shoppers sign in on the brand’s own login page. The agent never gets their password, only the access they approve, and they can revoke it at any time.

One caution. As The Next Web notes, the specification is still labelled a draft and allows changes that could break existing setups. It is worth following, but not something to rebuild your storefront around during peak week. If payments are your main concern, our earlier piece on PayPal’s agentic commerce terms covers how agent orders settle and refund.

What to prepare before Cyber Week

Salesforce’s Caila Schwartz put the advice simply in the survey release: “Be the brand that AI can find and humans can trust.” In practice, most of the work is data and site health, not new features.

Cyber Week checklist for AI agent traffic. Product data: titles and attributes complete, sale prices shown on product pages, inventory current and accurate, delivery estimates easy to find. Site and agents: bot rules let trusted agents in, site stays fast under extra traffic, shopper agent tested on real questions, AI referral traffic tracked
A short readiness list for the next ten days. Most of it is data quality and site health work your team can do now.

Product data agents can read

  • Check that titles, attributes, sizes and specifications are complete and consistent. Agents compare these line by line.
  • Make sure promotional prices show correctly on product pages, not only in banners or cart rules.
  • Keep inventory current. Salesforce describes Carter as pulling live inventory and pricing, so stale stock data affects your own agent too.
  • Show delivery estimates clearly. Commerce Cloud’s shipping and delivery estimates feature can be configured in Business Manager.

Site performance under agent traffic

Salesforce says retailers will need to handle nonhuman traffic without slowing their sites for human shoppers. Review your bot management and rate limiting rules so legitimate agents are not blocked by accident, while abusive traffic still is. For teams using AI coding tools, Commerce Cloud’s MCP and agent plugins, generally available since October, can help investigate bot attacks and inspect site performance.

Your own shopper agent

Trust is fragile. Most shoppers will abandon an AI assistant after two or three mistakes, and 27% of personal agent users say their agent has bought something that was too expensive. If you run Carter or another shopper agent, test it now with the questions your customers actually ask: sizing, delivery cut off dates, returns, and promotion terms.

What to do now and what to leave until January

Do before Cyber WeekPlan for January
Fix product data gaps on top sellersFull catalogue enrichment and attribute clean up
Check prices, stock and delivery dates on product pagesImplementing Poppy once Commerce Cloud support is available
Review bot and rate limiting rulesNew channels such as the Commerce Cloud TikTok Shop integration, due January 2027
Test your shopper agent with real customer questionsNew shopper agent use cases and storefront redesigns
Track AI referral traffic separately in analyticsReviewing peak results and agent driven sales
Data fixes and testing cannot wait. Larger integration and platform changes are safer after the code freeze.

How WeVerve helps

WeVerve Systems builds and supports eCommerce stores on Salesforce Commerce Cloud and other major platforms. Before Cyber Week, we can audit product data on your best selling categories, check that prices, inventory and delivery information reach product pages correctly, and review site performance and bot rules ahead of peak traffic. During Black Friday and Cyber Monday, our 24/7 eCommerce support team can watch site health and fix issues fast. In the new year, we can help you plan shopper agent improvements and agentic channel integrations such as UCP and Poppy as they mature.

If you want your store ready for both human shoppers and their AI agents this season, talk to the WeVerve team.

Sources

No Comments

Post A Comment