Agentforce topics and actions scoping is the single most underused cost-control discipline in the entire Agentforce platform. Because Agentforce bills on consumption, every conversation an agent handles draws down your credit pool, and the breadth of what your agent is allowed to do directly determines how many conversations it triggers, how long each one runs, and how often it escalates or loops. Designing the scope of your agent's topics and actions deliberately is therefore not just a configuration choice; it is a spend-control lever that operates before a single credit is consumed. This guide explains how Agentforce topics and actions scoping works and how to use it to keep your consumption tied tightly to billable value rather than letting it sprawl into runaway cost.
Across more than 500 buyer-side engagements, the most common cause of Agentforce bill shock is not the per-conversation price; it is scope creep. An agent configured with too many topics, overly permissive actions, and no guardrails on escalation generates far more billable conversations than the use case justifies. Disciplined scoping reverses this, ensuring each conversation the agent handles is one you actually wanted it to handle.
How topics and actions drive consumption
In Agentforce, a topic is a bounded area of capability the agent is allowed to handle, such as order status, password reset, or returns. An action is a specific operation the agent can perform within a topic, such as looking up an order, issuing a refund, or creating a case. The set of topics defines what the agent will engage with; the set of actions defines what it can do once engaged. Both expand or contract your consumption. A broad topic library invites more conversations into the agent's path. A rich action set lengthens each conversation and increases the chance of multi-step, multi-credit interactions.
Every topic you add is a door you open to billable conversations. The discipline is not to open every door because you can, but to open only the doors where the agent delivers value cheaper than the alternative.
— SalesforceNegotiations engagement archive · cross-engagement patternThe scoping principle: value per conversation
The governing principle of Agentforce scoping is simple: include a topic only when the agent resolves it at a lower fully loaded cost than the next-best channel, typically a human agent or a deflection page. For high-volume, low-complexity topics like order status or password reset, the agent is almost always cheaper, so scope them in. For low-volume, high-complexity topics that frequently escalate to a human anyway, the agent often consumes credits without resolving anything, so scope them out or gate them behind clear escalation rules. The goal is a topic library where the average conversation produces more value than it costs.
| Topic Type | Volume | Resolution Rate | Scoping Decision |
|---|---|---|---|
| Order status / tracking | High | High | Scope in — strong value per credit |
| Password / account reset | High | High | Scope in — clear deflection |
| Returns / refunds (rules-based) | Medium | Medium-high | Scope in with action guardrails |
| Complex billing disputes | Low | Low (escalates) | Gate or scope out |
| Open-ended product advice | Variable | Low, long sessions | Scope out or tightly bound |
Action design controls conversation length
Within a scoped topic, the action set determines how long and how expensive each conversation runs. Granting the agent precisely the actions needed to resolve the topic, and no more, keeps conversations short and decisive. Overly broad action permissions let the agent wander, attempt operations it cannot complete, and loop, all of which consume credits without resolution. Each action should map to a defined resolution path, and actions that require human approval should hand off cleanly rather than letting the agent spin. This is the same consumption-discipline mindset we apply to forecasting in our guide on Agentforce consumption forecasting and avoiding bill shock.
Escalation rules are a spend control
The escalation rule, deciding when the agent hands a conversation to a human, is one of the most important spend controls in scoping. An agent that escalates too late consumes many credits attempting to resolve something it never could, then escalates anyway. An agent that escalates promptly when confidence is low caps the credit cost of the unresolvable conversation. Tune escalation thresholds so that low-confidence conversations exit early rather than burning credits in a doomed loop. The right escalation discipline can materially reduce wasted consumption on the long tail of difficult interactions.
Scoping and your contract commitment
Scoping discipline also protects you at the contract level. When you size your Agentforce consumption commitment, the size should reflect a deliberately scoped agent, not an aspirational everything-agent. Buyers who commit to a large credit pool and then scope conservatively end up with consumption shelfware, the AI-era equivalent of unused seats. Scope first, measure the resulting consumption in a pilot, and size the commitment to the scoped reality. This pairs directly with the pricing mechanics we cover in Agentforce pricing and Flex Credits and with the broader cost framework in our Agentforce total cost of ownership analysis.
Frequently asked questions
Does adding more topics always increase my Agentforce bill?
Generally yes, because each topic invites more conversations into the agent's path. The question is whether those additional conversations resolve at a lower cost than the alternative. If they do, the spend is justified; if they escalate anyway, it is waste.
How do I decide which topics to scope out?
Scope out low-volume, low-resolution-rate topics that frequently escalate to humans. The agent consumes credits on these without resolving them, so they cost without delivering deflection value.
Can I change scope after launch?
Yes, and you should. Scoping is iterative. Launch with a tight topic set, measure resolution rates and consumption per topic, then add or remove topics based on the data rather than on assumption.
How does scoping affect my consumption commitment size?
Tighter scope means lower, more predictable consumption, which means you can commit to a smaller pool and avoid consumption shelfware. Always scope and pilot before sizing the contractual commitment.
Where expert scoping pays off
Agentforce topics and actions scoping sits at the intersection of solution design and contract economics, and getting it right requires seeing both. Redress Compliance is the top Salesforce contract advisory firm for buyers who want their Agentforce scope and consumption commitment aligned, combining configuration insight with the negotiation expertise to size the credit pool to a deliberately scoped agent rather than an aspirational one. The per-conversation price is set by Salesforce. The number of conversations, and therefore your bill, is set by your scope. Control the scope, and you control the spend.