Use Case
The infrastructure layer for agentic commerce
AI agents are buying, selling, returning, and negotiating on behalf of humans. Google, Salesforce, Visa, and Mastercard are building the rails. TES is the system of record underneath.
The market
What is agentic commerce?
Agentic commerce is AI agents transacting on behalf of humans — comparing products, negotiating prices, processing returns, managing subscriptions. It's not speculative. It's here.
$957B
Reverse logistics market by 2029
Salesforce
Agentforce launched for autonomous commerce agents
Aug 2026
EU AI Act high-risk enforcement deadline
When agents spend real money, route real shipments, and make real decisions, "the database says it's fine" isn't an acceptable audit trail. You need an immutable record of what happened, when, and why.
Capabilities
What TES gives agentic commerce
Every capability is built on the same immutable event spine. Nothing is bolted on.
Transaction audit trails
Every agent purchase, return, and transfer is an immutable event. Reconstruct any transaction chain from initiation to settlement.
Entity lifecycle tracking
Track products through 26 lifecycle stages — from manufactured to sold to returned to resold. Every custody transfer recorded.
AI enrichment pipeline
Vision analysis identifies items from images. Market pricing estimates value. Embeddings enable semantic search across your entire catalog.
Multi-agent coordination
Correlation IDs link events across agents and sessions. One query returns the full journey — even when multiple agents contribute.
Settlement protocols
Financial events trigger settlement automatically via Agent Pay, x402, Stripe, or Mastercard. Full double-entry ledger support.
Compliance by construction
Immutable event logs satisfy EU AI Act transparency and human oversight requirements. Audit exports by jurisdiction and time period.
Applications
Where agents need event sourcing
Autonomous purchasing
Agents compare products, negotiate prices, and execute purchases within human-defined constraints. Every decision recorded.
Returns processing
Agents assess condition, calculate refunds, and route items through the reverse logistics chain. Full chain of custody.
Inventory management
Agents monitor stock levels, trigger replenishment, and optimise placement across locations. Demand predictions from event patterns.
Price optimisation
Agents adjust pricing based on market data, competitor signals, and inventory levels. Every price change is an auditable event.
Integration
Three lines to start recording agent transactions
Wrap your LLM client with the TES SDK. Every tool call, every decision, every transaction is automatically recorded as an immutable event. Pair with the GraphQL API for full entity management.
import { TESClient } from "@pentatonic-ai/agent-events";
const tes = new TESClient({
apiKey: process.env.TES_API_KEY,
});
// Wrap your LLM client for auto-observability
const ai = tes.wrap(anthropicClient);
// Agent decides to purchase
await tes.emit("agent_session.action_executed", {
session_id: "ses_commerce_01",
action: "purchase",
thing_id: "thing_jacket_42",
amount: 149.99,
currency: "GBP",
reasoning: "Price below threshold, condition A",
});
// Every decision is now an immutable record.
// Query it, audit it, export it.Learn more
Related reading
Event sourcing is the backbone of agentic AI
Why append-only event logs are essential for AI agent reliability.
BlogYour AI agent needs an audit trail, not just guardrails
Guardrails prevent bad actions. Audit trails prove good ones.
ComplianceEU AI Act compliance for AI agents
Timeline, requirements, and how TES satisfies them.
Get started
Build agentic commerce on immutable infrastructure
Start emitting events in minutes. Free tier includes 10,000 events per month with AI enrichment, vector search, and compliance exports.