Agent-Completed Revenue is revenue generated from sales in which an AI agent directly carried out the shopping process or completed a transaction on behalf of the user. It includes purchases where the agent selected the product, entered order details, moved through checkout, or led directly to payment confirmation.
Unlike Agent-Assisted Revenue, this metric applies to transactions where the AI agent’s role goes beyond recommendation or referral. The agent actively participates in executing the purchase.
Agent-Completed Revenue = total value of transactions directly completed or carried out by AI agents
For example, if an AI agent finds a product matching the user’s requirements, selects a seller, creates the cart, and places the order after receiving the required approval, the transaction value may be included in Agent-Completed Revenue.
What Does an Agent-Completed Transaction Mean?
Depending on the system and available permissions, an AI agent may:
- select a product and seller,
- verify price and availability,
- add the product to the cart,
- choose a variant, delivery method, or collection point,
- apply a promotional code,
- provide the information required to process the order,
- initiate payment,
- request user confirmation,
- place the order,
- receive transaction confirmation.
The level of agent autonomy may vary. In some processes, AI performs most actions while the user only approves the purchase. In others, the agent may operate within predefined permissions and spending limits.
Why Is Agent-Completed Revenue Important?
Agent-Completed Revenue measures the value of sales generated through a new purchasing model in which the customer does not have to navigate the entire online store and checkout process independently.
As Agentic Commerce develops, some transactions may take place:
- within an AI interface,
- through an agent connected directly to the store,
- without a traditional website session,
- with fewer manual actions performed by the customer,
- through external payment systems or agentic commerce protocols.
In such cases, sessions and traditional website conversion rates may not reflect the actual value of sales.
How to Measure Agent-Completed Revenue
Measurement should primarily rely on transaction data and reliable identification of the order source.
Relevant data may include:
- The agent or platform identifier.
- An order status such as agent-initiated or agent-completed.
- Checkout system data.
- Payment platform data.
- Information from Agentic Commerce APIs and protocols.
- The source of cart creation.
- The status of the purchasing process.
- Information about who confirmed and completed the transaction.
- Data on cancellations, returns, and failed agent transactions.
The metric can be reported as:
- total Agent-Completed Revenue,
- share of total sales,
- number of agent-completed transactions,
- average order value,
- revenue by agent or platform,
- revenue by product category,
- net revenue after returns and cancellations.
Agent-Completed Revenue vs. Agent-Assisted Revenue
- Agent-Assisted Revenue includes sales where an AI agent influenced the customer journey.
- Agent-Completed Revenue includes transactions directly carried out or completed by the agent.
- AI Referral Revenue includes purchases completed after the user visited the store from an AI system.
Every transaction included in Agent-Completed Revenue can also be considered agent-assisted. However, not every transaction included in Agent-Assisted Revenue was completed by an AI agent.
Example
A user asks an AI agent to purchase a specific brand of dog food for less than €50 with next-day delivery. The agent finds a suitable offer, verifies availability, selects the correct variant, adds it to the cart, and places the order after the user’s approval.
The value of this order is included in Agent-Completed Revenue because the agent directly carried out the purchasing process.
Summary
Agent-Completed Revenue is the value of sales directly carried out or completed by an AI agent. It shows how much revenue comes from shopping processes in which the agent actively selects the offer, creates the order, and leads to the transaction.
