Agentic commerce has not yet taken over the full shopping journey; near-term value is concentrated in discovery, merchant optimization, and payment infrastructure
AI summary card
Agentic commerce has not yet taken over the full shopping journey; near-term value is concentrated in discovery, merchant optimization, and payment infrastructure
The core conclusion of Bernstein's inaugural Agentic Commerce Day is that consumers have already begun using AI for product discovery, but true agent-driven transaction completion remains far off. Payment networks, leading platforms, and large retailers with fulfillment and data advantages are better positioned.
- At present, referral traffic from AI chatbots to large players is below 1%, while some DTC brands can reach as high as 10%, showing that AI discovery already has early value but has not yet become a mainstream transaction entry point.
- OpenAI's decision to drop Instant Checkout highlights the difficulty of changing consumer behavior, merchant integration, real-time pricing, inventory synchronization, fulfillment, and multi-cart settlement.
- A BCG survey shows that 70% of merchants are already working on agentic-commerce discovery initiatives, and this figure is expected to rise to 90%; catalog maintenance, GEO, and product-content optimization are becoming critical.
- The report is more constructive on Visa and Mastercard, arguing that agentic transactions could create more opportunities in tokenization, security, consent management, programmable payments, dispute handling, and value-added services.
- DTC brands may benefit in the near term from AI improving discoverability in niche categories, but in the long run large brands, leading marketplaces, and retailers with fulfillment networks and pricing power may regain the upper hand.
Report interpretation
Overview
This report summarizes the cross-industry discussion at Bernstein's inaugural Agentic Commerce Day in New York, with participants including representatives from Wayfair, Mastercard, BCG, former Walmart executives, DTC brands, and Paxos. The report focuses on the impact of agentic commerce on retail, internet platforms, payment networks, retail media, and stablecoin infrastructure. The core view is that agentic commerce is still in an early iteration phase: consumers are increasingly using AI tools for discovery and search, but have not yet widely delegated the full shopping and payment process to agents.
Core views
The report argues that the short-term winners are not autonomous agents that fully replace merchants and marketplaces, but rather participants that embed AI into discovery, search, product recommendations, product-page optimization, customer support, marketing efficiency, and payment security workflows. Large platforms and leading retailers have stronger defensive and offensive positions because they own traffic, data, catalogs, fulfillment networks, and AI experimentation capabilities; mid-sized merchants may be the easiest to lag behind. On the payments side, Visa and Mastercard are expected to become key rule-makers for standards around risk, liability, identity, authentication, tokenization, and dispute management, putting them in an advantageous position within the agentic transaction ecosystem.
Analysis framework
The report uses a conference-note and expert-interview style thematic research approach, integrating views from Wayfair, Mastercard, BCG, retail practitioners, DTC brands, and stablecoin infrastructure participants into investment implications, and cross-checking them against Bernstein's coverage rating table, industry disclosures, and prior research in the same series.
Methodology notes
Assessing the commercialization stage of a new technology through cross-industry conferences and expert views
The report is not a single-company financial model update; rather, it forms a thematic investment view around the maturity of agentic commerce, consumer behavior, merchant integration, payment liability, and advertising-budget migration.
Inferring e-commerce advertising and marketplace economics from changes in search entry points
The report examines whether AI-assisted shopping will change search entry points, query-intent density, and conversion rates, and thereby affect how high-margin retail media budgets are allocated across marketplaces, search platforms, and omnichannel retailers.
Agentic transactions require stronger identity, authorization, audit, and dispute-handling mechanisms
The report argues that if agents recommend or purchase on behalf of users, the ecosystem will need standards around liability, risk, authentication, and authorization, which may expand token and VAS opportunities for payment networks.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- WAYFAIR INC (US.W)Conference participant and covered internet-retail name
- Strengths
- Wayfair is building AI-assisted shopping experiences and has partnered with Gemini to launch UCP, giving it a venue to test conversational and AI-assisted discovery in high-consideration categories such as home goods.
- Weaknesses
- Fully agentic direct purchasing is harder to deploy in medium-to-high consideration categories, and consumers still tend to keep relationships with familiar marketplaces and retailers.
- Comparison
- Compared with long-tail or mid-sized merchants, Wayfair as a large vertical retail platform is better positioned to integrate directly with leading AI labs; but compared with even larger platforms such as Amazon, its traffic, fulfillment, and ecosystem scale may still be weaker.
- Risks
- Changes in AI shopping entry points may affect traffic sources, conversion paths, and ad efficiency; if consumer behavior shifts slowly, returns on investment may lag.
- Visa Inc (V) / Mastercard Inc (MA)Potential beneficiaries of agentic-commerce payment infrastructure
- Strengths
- Payment networks have an ecosystem position in tokenization, identity, authentication, liability standards, dispute management, and value-added services, and may become the standard-setters for agentic transactions.
- Weaknesses
- Machine payments and microtransactions for agents remain at a very early stage, and the business model is unproven.
- Comparison
- Compared with a single payment service provider or a merchant-built solution, the networks have broader acceptance, standard-setting power, and cross-ecosystem connectivity.
- Risks
- If liability allocation, fraud, user consent, audit trails, and regulatory requirements are not resolved, the scaling of agentic payments may be delayed.
- DTC brandsEarly beneficiaries of AI discovery
- Strengths
- AI tools can improve the discoverability of DTC brand products in niche use cases; the report notes that referral traffic from AI chatbots can reach 10% for some DTC brands.
- Weaknesses
- In the long run, large brands and large retailers may catch up through fulfillment, pricing, and data advantages.
- Comparison
- In the near term, they may benefit more from incremental discovery than large platforms, but their long-term defensiveness is weaker than that of large retailers with scale in fulfillment and advertising budgets.
- Risks
- If AI discovery rules tilt more toward catalog completeness, fulfillment reliability, and pricing competitiveness, the traffic tailwind for DTC brands could narrow.
- Stablecoin-linked cards / stablecoin infrastructurePotential infrastructure for agent and machine payments
- Strengths
- Stablecoin-related participants are beginning to acknowledge the feasibility of card- and stablecoin-linked cards in consumer-to-merchant use cases, and card networks can provide acceptance, standards, and monetization models.
- Weaknesses
- Machine payments, microtransactions, and related business models are still early.
- Comparison
- Compared with purely on-chain payments, a combination of card networks and stablecoins can more quickly tap into existing merchant acceptance networks and dispute-resolution systems.
- Risks
- Settlement, orchestration, fiat on/off ramps, regulation, consumer protection, and the expansion of B2B/B2C/P2P use cases still require substantial infrastructure build-out.
Key data
- Share of referral traffic from AI chatbotsBelow 1% for large players; as high as 10% for some DTC brandsShows that AI discovery has early impact, but has not yet become a mainstream transaction entry point.
- Merchants advancing agentic-commerce discovery70% already underway, expected to rise to 90%Based on the merchant survey referenced by BCG at the conference.
- View on agentic transaction tokenizationCould be close to 100% tokenizedThe report argues that the complexity of intent, disputes, audit trails, identity, and authentication will drive tokenization.
- Bernstein stock rating definitionOutperform is more than 15 percentage points above the market index; Market-Perform is within plus or minus 15 percentage points of the market index; Underperform is more than 15 percentage points below the market indexApplies to Bernstein's standard common-stock ratings, on a 12-month horizon.
- WAYFAIR INC ratingMarket-PerformThe report lists W as Market-Perform in its U.S. emerging internet coverage list.
Impact & implications
In investment terms, agentic commerce is more likely in the near term to strengthen existing e-commerce and payment infrastructure rather than immediately disrupt the shopping chain. If leading marketplaces can make AI experiences more conversion-efficient, habit-forming, and data-rich, they still have a chance to preserve or even expand retail-media budgets. Payment networks may benefit from demand for tokens, identity, security, consent management, and dispute handling. For merchants, product-catalog quality, real-time inventory, promotions and loyalty-system integration, and product-description and image optimization will become more important. For DTC brands, AI discovery can deliver a short-term exposure tailwind, but in the long run large brands and large retailers may catch up through fulfillment, pricing, and data advantages.
Risks
- Consumer behavior may change slowly, with AI tools remaining at the discovery and assisted-search stage and failing to convert into fully delegated transactions.
- Merchant catalog, inventory, pricing, promotions, fulfillment, and multi-cart settlement integration are complex and may limit the scaling of agentic commerce.
- When agents recommend or purchase on behalf of users, liability, risk, fraud, dispute handling, and audit trails have not yet been fully standardized.
- If AI shopping entry points change search traffic sources, budget allocation across marketplaces, search platforms, and retail media may be disrupted.
- Mid-sized merchants may fall behind in GEO and AI shopping experiences because of insufficient technology, talent, and integration capabilities.
What to watch
- Whether AI applications such as ChatGPT and Gemini move beyond product discovery to a scalable closed-loop transaction experience.
- Progress on Google's shopping graph, OpenAI's checkout strategy, and merchant catalog integration.
- Whether AI-assisted shopping experiences at large retailers and marketplaces improve conversion rates, repeat purchases, and ad monetization.
- Standards formed by Visa, Mastercard, Stripe, Google, OpenAI, and merchants around tokens, liability, identity, and dispute handling.
- Whether referral traffic from AI chatbots to DTC brands can remain above large players and convert into sales.
- The pace of rollout for stablecoin-linked cards, B2B/B2C/P2P settlement, and on/off-ramp infrastructure.