---
title: "Who Is the Agent Buying For? The Agentic Commerce Stack"
description: "Agentic commerce works only when a shopper's agent can be identified, scoped, and audited at each step. This architecture map shows the delegation, consent, and audit layers that make an agent's purchase safe."
---

# Who Is the Agent Buying For? The Agentic Commerce Stack

**Type:** [Architecture](https://yomu.fyi/visual-explainers/category/architecture.md)

![A layered architecture diagram showing agentic commerce flowing top to bottom through Intent, Delegation, Consent Gate, and Execution layers, alongside a side rail for audit trail and durable state linked to every layer.](https://yomufyi.s3.us-east-1.amazonaws.com/visual-explainers/0c767512-1cc0-44dc-896e-ecd2827553b1/image-ea0d5900-fb07-46b9-b417-b6093299f150.webp)

> Agentic commerce works only when a shopper's agent can be identified, scoped, and audited at each step. This architecture map shows the delegation, consent, and audit layers that make an agent's purchase safe.

## Explanation

Agentic commerce means an agent acts on a shopper's behalf to compare, choose, and pay, which makes identity the load-bearing layer of the stack. The diagram reads top to bottom: intent starts at the shopper's agent, delegation turns that intent into a scoped credential tied to a known person, a consent gate decides whether an action is approved or reversible, and execution reaches merchant and payment systems. A side rail carries durable state and audit traces so every step can be replayed. Get the layers wrong and you get abandoned carts from excess friction, or credentials that outlive the relationship that justified them.

## FAQ

### Why give an agent its own identity instead of reusing the user's credentials?

A scoped, revocable credential limits what the agent can do and keeps every action attributable. MongoDB Atlas App Connections uses OAuth 2.1 delegation so an AI client acts on the user's behalf, with read-only or full access and revocation at any time, and each action attributed to both the application and the user.

### How much friction can an agentic checkout afford before shoppers abandon it?

In the cited survey of 5,250 consumers, 74% would abandon a task if forced to create an account before seeing key information, and 72% cited slow login or verification. Only 6% were ready for AI to handle as much as possible, yet most accepted agent actions that are approved first or reversible, which is why the consent gate sits before execution.

### What breaks when an MCP client issues its own identity assertion?

Claude can sign the assertion when an identity provider lacks the MCP extension, but it does not confirm an identity-provider sign-in, check multi-factor authentication, or apply the app's sign-in rules. Disabling the member in the identity provider no longer cuts off access, so they must be removed from both organizations, and issued tokens stay valid until expiry.

**Tags:** [Agentic AI](https://yomu.fyi/topic/agentic-ai.md), [Agentic Commerce](https://yomu.fyi/topic/agentic-commerce.md), [Auditability](https://yomu.fyi/topic/auditability.md), [Authorization](https://yomu.fyi/topic/authorization.md), [MCP](https://yomu.fyi/topic/mcp.md), [Payments](https://yomu.fyi/topic/payments.md)

## Related reading

### [Claude Platform now supports native OpenTelemetry observability in Databricks MLflow](https://yomu.fyi/post/claude-platform-now-supports-native-opentelemetry-observability-in-dat.md)
- Company: [Databricks](https://yomu.fyi/company/databricks.md)
- Author: Yuki Watanabe
- Published: Nov 3, 2026
- Type: Announcement

The Claude Platform now supports native OpenTelemetry export to Databricks MLflow, enabling developers to trace agent executions. Traces record model requests and tool calls as spans, ingested via Zerobus and stored in Unity Catalog tables linked to MLflow experiments. Setting up requires creating an experiment with Unity Catalog trace storage, configuring native export with a table-scoped OAuth token, and running a task to inspect traces. Developers can debug unexpected answers by analyzing traces with Genie Code, which identifies missing context or failed tool calls, and MLflow Issues detects recurring problems across traces. Agent Evaluation compares baseline and revised agents using the same traces to measure improvements or regressions, with domain experts able to review traces via Review Queues.


### [Enterprise-managed auth for MCP: Claude can now issue the identity assertion](https://yomu.fyi/post/enterprise-managed-auth-for-mcp-claude-can-now-issue-the-identity-asse.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: cemal\_kilic, wenbo
- Published: Nov 3, 2026
- Type: Announcement

Supabase’s MCP server now supports enterprise-managed authorization when an organization’s identity provider lacks the MCP Enterprise-Managed Authorization extension, because Claude can issue the signed identity assertion itself. After a Supabase owner or admin approves Claude and registers it as the issuer, members with matching Supabase SSO identities use Supabase from Claude without a per-member consent screen. Claude checks that the person is a current member of the Claude organization and which email address it uses, but it does not confirm an identity-provider sign-in, check multi-factor authentication, or apply Supabase’s sign-in rules. Supabase verifies the signature against the registered issuer, the member’s matching SSO identity, and the admin authorization before issuing a short-lived access token. Disabling a member in the identity provider no longer cuts off access, so they must be removed from both organizations, and tokens remain valid until expiry.


### [Consumers are Ready for AI. Are Companies?](https://yomu.fyi/post/consumers-are-ready-for-ai-are-companies.md)
- Company: [Auth0](https://yomu.fyi/company/auth0.md)
- Author: Auth0 Team
- Published: Nov 3, 2026
- Type: Explainer

The 2026 Customer Identity Trends Report, commissioned by Okta, surveys developers, senior leaders, and consumers on customer identity and AI. Among North American senior leaders, 23.2% named competitive disruption from AI-native competitors as the single biggest threat, second to cybersecurity and data risk at 31.4%, and 50.2% called AI a must-win capability. Consumer readiness remains limited: 37% are ready for AI to make final decisions and act (7%) or handle routine tasks (30%), though willingness exceeds 50% for most use cases when actions can be approved or reversed. More than three-quarters of organizations have AI agents or autonomous systems in production or under active development, but only 22.1% are moving as quickly as they would like. Delays include engineering capacity, cited by nearly 20%, and identity infrastructure that cannot support agent access patterns, cited by slightly more than 10%.


### [AI ambitions are blocked by identity debt](https://yomu.fyi/post/ai-ambitions-are-blocked-by-identity-debt.md)
- Company: [Auth0](https://yomu.fyi/company/auth0.md)
- Author: Auth0 Team
- Published: Nov 3, 2026
- Type: Problem & solution

Japanese organizations are investing in AI and agentic features, but outdated customer identity infrastructure is slowing delivery, according to the 2026 Customer Identity Trends Report. Nearly three-quarters of Japanese developers and senior leaders polled (73%) said their organization either had AI or agentic features fully in production (29%) or was actively building them (44%). Of the active builders, fewer than 28% are delivering at the speed they want. Nearly 60% of organizations that have not started cited outdated identity infrastructure as the primary cause, and respondents report spending more than 15% of engineering time on CIAM. The brief argues that modern CIAM platforms should provide authentication, authorization, and audit capabilities for AI agents, including a token vault, MCP authentication, asynchronous authorization, agent as principal, and fine-grained authorization for RAG.


### [Who Will Consumers Trust to Act, Shop, and Pay on Their Behalf?](https://yomu.fyi/post/who-will-consumers-trust-to-act-shop-and-pay-on-their-behalf.md)
- Company: [Auth0](https://yomu.fyi/company/auth0.md)
- Author: Auth0 Team
- Published: Nov 3, 2026
- Type: Explainer

As AI agents evolve from answering questions to taking action, the post examines what identity and security model retailers need for convenient, secure customer experiences. McKinsey estimates $3–5 trillion will flow through agentic commerce by 2030, with 25–40% of e-commerce agentic. In a survey of 5,250 consumers, 74% said they would abandon a task if asked to create an account before seeing key information, while 72% cited slow login or verification and 71% repeated identity checks. Only 6% of consumers said they were ready for AI to handle as much as possible, yet a majority would let agents act in limited-consequence cases when they approve actions first or can reverse them. The argument is that retailers need proportionate friction and a CIAM platform that links agents to customers, scopes credentials, and enforces fine-grained authorization with audit trails.


### [Financial Crime Mitigation with MongoDB / Part IV: Agentic AI for Case Management](https://yomu.fyi/post/financial-crime-mitigation-with-mongodb-part-iv-agentic-ai-for-case-ma.md)
- Company: [MongoDB](https://yomu.fyi/company/mongodb.md)
- Published: Nov 3, 2026
- Type: Explainer

The article describes an agentic AI pipeline for enhanced due diligence (EDD) in financial crime case management, built with LangGraph and MongoDB. It aims to replace manual, rule-based AML workflows with specialized agents that triage suspicious activity, gather evidence, and draft Suspicious Activity Report (SAR) narratives. The pipeline uses a team of agents for notification, data gathering, case analysis, SAR authorship, and compliance quality assurance. A human-in-the-loop pattern ensures each decision can be traced for auditability. MongoDB serves as the durable state backbone, while LangGraph handles orchestration and dynamic routing. The author emphasizes that enterprise agentic AI must centralize state in a persistent store rather than rely on the agent as the source of truth.


### [Introducing Atlas App Connections: Secure, One-Click Access for AI Tools](https://yomu.fyi/post/introducing-atlas-app-connections-secure-one-click-access-for-ai-tools.md)
- Company: [MongoDB](https://yomu.fyi/company/mongodb.md)
- Published: Nov 3, 2026
- Type: Announcement

MongoDB Atlas App Connections enables builders to securely connect AI coding tools to their MongoDB Atlas account with a single click, eliminating the need to share long-lived service account credentials. Built on OAuth 2.1, it allows AI clients to act on the user’s behalf through the Atlas MCP Server without generating or managing credentials. Supported tools include ChatGPT, Claude, Codex CLI, Cursor, Devin, and Grok, which can provision, manage, and query Atlas resources based on the user’s existing permissions. Administrators can opt in to delegation, set read-only or full access, and revoke connections at any time, with all actions attributed to both the application and the user. The feature is available now, with a five-step connection flow that includes installing a plugin, requesting connection, and approving via a consent screen.
