Amazon Quietly Builds a 40-Person Agentic Commerce Team to Connect Its Marketplace with ChatGPT

The outline of Amazon's quietly assembled 40-person team

Amazon quietly builds an agentic commerce team to connect with ChatGPT

Amazon posted a job listing for a Principal TPM to lead a dedicated 40-person team integrating its marketplace with third-party AI agents including ChatGPT.

ppc.land

In May 2026, a single job listing on Amazon Jobs surfaced one of the company's most significant strategic shifts of the year. Job ID 10411992, Seattle-based, titled "Principal Technical Program Manager, Agentic Commerce Experiences," with a base salary range of $177,000 to $239,400. Sitting inside Amazon.com Services LLC, the hire will lead one of three specialized teams within a 40-person engineering organization.

The listing was first highlighted on LinkedIn on May 8 by e-commerce analyst Juozas Kaziukenas, who noted he had "not seen any other mentions of agentic commerce teams at Amazon" before. Amazon itself did not announce the organization. The structure came to light through the most operational channel possible: a job posting.

The scope is specific. According to the description, the hire will oversee "the development of robust, scalable APIs and integration layers that connect Amazon's services with third-party platforms or on-site/off-site customer experiences." Internally the role spans Search, Personalization, and other core service teams; externally it coordinates with third-party platform teams to drive the technical roadmap. The experience bar — seven-plus years of technical program management, ten-plus years working with engineering teams, five-plus years in software development — suggests Amazon is staffing for production-grade execution from day one, not exploratory prototyping.

From "block" to "controlled integration"

The timing matters. Amazon blocked AI bots from OpenAI, Anthropic, Meta, Google, and Huawei via robots.txt on August 21, 2025. In November the company sued Perplexity AI in federal court over the unauthorized deployment of AI agents on its platform. In February 2026, Amazon updated its Business Solutions Agreement to introduce a new Agent Policy effective March 4, 2026, requiring AI agents operating on Amazon to self-identify and comply with contractual terms.

That posture — keep unsanctioned external agents out, build proprietary alternatives — remained consistent throughout that period. The newly visible agentic commerce team does not contradict it. It runs alongside it as a second track with different rules of engagement. As Kaziukenas put it, blocking scrapers and building controlled APIs are not opposing strategies. They are complementary: a scraper takes data without permission or commercial arrangement, while an API integration lets Amazon set the terms on what data is exposed, at what latency, under what commercial conditions, and with what attribution of purchase intent.

The economic logic is straightforward. Amazon's advertising revenue reached $17.2 billion in Q1 2026, up 24% year over year. Protecting that revenue stream while still accessing new AI-mediated discovery channels requires gated, paid access rather than open scraping.

The UCP connection

The listing aligns with Amazon's recent participation in the Universal Commerce Protocol Tech Council. On April 24, 2026, Amazon, Meta, Microsoft, Salesforce, and Stripe joined the UCP Tech Council, alongside founding participants including Google, Shopify, Etsy, Target, and Wayfair. Amazon had been notably absent from UCP governance until that point, despite the protocol's January 2026 launch.

Joining the Tech Council is a governance move that places Amazon inside the room where specifications are written, but it does not automatically produce technical integration. This Principal TPM listing suggests the engineering layer that turns governance participation into shipped code is now being staffed.

The Amazon Ads MCP Server, which opened in beta at the IAB Annual Leadership Meeting in February 2026, provides a useful precedent. That server lets AI agents interact with Amazon advertising APIs in natural language. The agentic commerce team appears to be building an analogous capability on the commerce side — exposing product inventory, pricing, and purchasing flows to external agent platforms through a controlled API surface, rather than building one-off integrations for each AI partner.

How this differs from Rufus and Join the Chat

It helps to distinguish three separate Amazon initiatives that all sit under the "agentic" umbrella but point in different directions.

Rufus reached 300 million users in 2025 and reportedly drove approximately $12 billion in incremental sales. It narrows recommendations to roughly five named products through Amazon's proprietary COSMO knowledge graph and runs entirely inside Amazon-owned surfaces. Join the Chat, announced in early May, takes the reverse approach within Amazon's perimeter: invites external models like Claude onto Amazon's customer-facing pages.

The newly surfaced Agentic Commerce Experiences team points outward. Its job is to expose Amazon product data and purchasing flows to external AI agent surfaces such as ChatGPT, where a consumer might ask for product recommendations and receive Amazon listings as candidates, with checkout and fulfillment completed downstream. The three tracks together cover three faces of the same shift: AI systems becoming the dominant interface through which consumers discover products.

Competitive context

Amazon is not alone in building this infrastructure. OpenAI and Stripe co-developed the Agentic Commerce Protocol, which underpins ChatGPT's Instant Checkout that launched on September 29, 2025. Google is a central architect of UCP. Microsoft launched its Copilot Merchant Program in April 2025.

Amazon's structural position differs from other retailers because it is simultaneously a marketplace operator, an advertising platform, a fulfillment network, and a major AI infrastructure provider via AWS. Decisions about which AI agents can access its product data, and under what terms, affect not just Amazon's own commerce flows but the hundreds of thousands of third-party sellers who list on its platform.

Shopify, which we covered separately on May 6, positions its ChatGPT and Claude integration as open, cross-merchant infrastructure. Amazon's approach is to expose its in-house marketplace inventory to external AIs under controlled conditions — a design philosophy quite different from Shopify's distributed network model. For sellers, the question of which platform will deliver more AI-mediated traffic increasingly shapes channel strategy.

The supporting data is striking. AI bot traffic to retail sites grew 5.4x during 2025, according to figures cited in PPC Land's UCP coverage. A March 2026 joint report from Retail Economics, AWS, Botify, and DataDome reported that OpenAI's systems perform 198 crawls per single visit delivered to a retail site. Amazon's pivot from blocking crawlers to opening APIs is a coherent response to that asymmetric "high crawl, low conversion" pattern.

What sellers and merchants should do now

For brands selling on Amazon, two priorities stand out.

The first is the completeness of structured product attributes. The COSMO graph that powers Rufus reads structured backend attribute fields. External agents consuming Amazon's integration APIs are likely to depend on similar structured signals. Catalogs with missing attributes or inconsistent naming will be structurally disadvantaged in AI-mediated recommendations, regardless of advertising spend.

The second is to anticipate the rebalancing of advertising surfaces. Amazon's ad revenue depends on consumers comparing products on Amazon.com. If decision-making shifts to external AI surfaces, the reach of traditional sponsored placements changes. New API-tied ad products will almost certainly emerge, and sellers should track those releases carefully.

A third, underrated step is multi-channel data consistency. Once an agent like ChatGPT can receive product information through both Amazon and Shopify pipelines, brands need their product pages, pricing, and inventory data to remain coherent across both. Mismatches across channels will confuse agent reasoning and erode purchase intent.

Closing thoughts

The 40-person agentic commerce team Amazon quietly assembled is a concrete signal of a strategic shift from "block everything" to "open under controlled terms." Rufus polishes the AI experience inside Amazon. Join the Chat invites external AI in. The newly surfaced Agentic Commerce Experiences team pushes Amazon inventory out to external agent surfaces. The three tracks let Amazon hedge across multiple possible futures of how AI agents become the primary product-discovery interface.

The harder question — whether external integration ultimately complements or competes with Amazon's advertising revenue and on-site discovery — will only be answerable once the APIs ship. Until then, sellers should focus on structured product data and multi-channel data consistency: investments that retain value in every plausible outcome.