What it costs to run

This recording explains the factors that influence Agentforce operating costs for the agents in this series, and distinguishes ongoing usage from one-time implementation and licensing. It is an explainer for agency leaders and finance teams, not an agent demonstration.

Overview

This recording is for insurance agency leaders and finance and operations teams evaluating Agentforce. It is an explainer, not a demonstration of an agent's workflow. It looks at what drives the ongoing cost of running AI agents in Salesforce, using the agents shown in this series and an example agency.

What you'll see

The recording is shot in the Agent Catalog, an app inside the Salesforce demonstration org, as a sales leader sees it. It explains how Agentforce usage is counted, which kinds of agent work drive that usage, how the choice of AI model affects it, and how the different ways of buying Agentforce compare. It then separates ongoing usage from the one-time implementation and from the staff time needed to look after the agents.

What drives usage

As the recording explains, usage is counted each time the AI writes something and each time an agent takes an action, rather than by the length of the text. An agent that runs on a schedule for every producer is used more as the team grows, and overall use rises with adoption. Testing in a sandbox also counts as usage.

Keeping the parts apart

The recording treats three things separately: ongoing usage, which depends on which agents run and how often; the one-time implementation, which depends on how much the solution is tailored to the agency; and licensing, which depends on the Salesforce edition and how Agentforce is bought. Base Salesforce licences and Microsoft licensing are outside its scope. Salesforce pricing and packaging change, so current terms should be confirmed with Salesforce for your own contract.

Configured for your organization

The recording is based on PS Advisory's foundational agents in its Agentforce for Insurance accelerator, which are configured for each customer's Salesforce org. Which agents you switch on, how often they run and which model they use are configuration choices, so usage is planned around your agency rather than assumed.