Salesforce Agentforce Pricing 2026: Conversations, Credits, and Per-User Plans

Salesforce Agentforce pricing now comes in three different models: $2 per conversation, Flex Credits at $500 per 100,000 credits, or per-user licensing, from a $5 Agentforce User License that still draws Flex Credits to $125 per user per month for add-ons with unmetered employee use. Which one applies depends on how and where you deploy it. Agentforce runs on the Salesforce Platform, its per-user add-ons attach to existing Sales, Service, Field Service, or Industries cloud licenses, and deployments carry, per independent 2026 reporting, a typical implementation timeline of five to eleven months, with fewer than 10% of customers reported to have it fully scaled. Prices checked against Salesforce's Agentforce pricing page on November 3, 2026.

If you are evaluating Agentforce in 2026, here is how the pricing evolved, what each model actually charges for, and how it compares to a standalone AI coworker.

Salesforce Agentforce pricing at a glance

Source: Salesforce's official Agentforce pricing page, checked November 3, 2026, lists all three models as current options. Confirm which model applies to your deployment and what counts as a "conversation" or a "credit" for your specific use case before you buy, since a single customer interaction can trigger multiple backend actions.

How Agentforce got to three pricing models

Agentforce's pricing has changed more in two years than most enterprise software changes in a decade, a sequence SaaStr has tracked model by model, and the history explains why the options look the way they do today.

2024 launch: $2 per conversation. The original model billed per conversation. It was simple to understand but priced out smaller organizations and made costs hard to predict, since what counts as a single "conversation" when one customer query triggers several backend processes was never entirely clear.

May 2025: Flex Credits. Salesforce introduced consumption-based Flex Credits, priced at $500 per 100,000 credits, with individual actions like updating a record or drafting a summary consuming credits based on complexity. This shifted billing from the conversation level to the action level.

More recently: per-user licensing. Salesforce added a third option, per-user licenses. Today they range from a $5 per user per month Agentforce User License, which still consumes Flex Credits, to $125 per user per month add-ons ($150 for Industries clouds) with unmetered employee usage, giving buyers a flat-fee alternative to usage-based billing.

Having three live pricing models at once is unusual, and it reflects genuine difficulty translating agentic AI usage into a bill customers can predict in advance. If you are quoting Agentforce, get clarity on which model your quote uses and model your expected volume under that specific structure before committing.

What actually drives your Agentforce bill

1. Which pricing model you're on. The three models measure different things. Conversations count interactions, Flex Credits count granular actions, and per-user licensing counts seats. The same usage pattern can cost meaningfully different amounts depending on which model applies.

2. Existing Salesforce licenses. Agentforce runs on the Salesforce Platform, and the per-user add-ons attach to Sales, Service, Field Service, or Industries cloud licenses, so the add-on price is added cost on top of those licenses, not a standalone platform fee. The $50 per user AI-teammate license is the exception: Salesforce sells it as a standalone SKU.

3. Implementation timeline and cost. Independent reporting on Agentforce deployments describes typical setup timelines of five to eleven months. That is a real cost center beyond the licensing fee, in consulting time, internal resourcing, and delayed time to value.

4. Action complexity under Flex Credits. Not every action costs the same number of credits. More complex agent actions consume more of your 100,000-credit allotment, so the effective cost per task varies by what the agent is actually doing.

Hidden costs to check before you sign

• Underlying Salesforce licenses. Confirm the all-in cost including the Sales, Service, or other cloud licenses the add-ons attach to.
• What counts as a conversation or a credit. Get the specific definition in writing for your use case; ambiguity here directly changes your bill.
• Implementation cost and timeline. Budget for a multi-month rollout, not a self-serve setup, per independent reporting on typical deployment length.
• Which model actually fits your usage. High-volume, simple interactions may favor one model; complex, lower-volume workflows may favor another. Model your specific case under each option before choosing.

Agentforce vs Coworker

Agentforce is a Salesforce-native agent platform with usage-based or per-user pricing layered on top of Salesforce's clouds. Coworker is a standalone AI coworker.

The practical difference: Agentforce extends an existing Salesforce investment with agent capability, priced and implemented at Salesforce's enterprise scale. Coworker is a context layer for enterprise AI that works across Salesforce and the rest of your stack, with a proof of concept in 48 hours. For the category view, see enterprise AI pricing compared and AI agent use cases.

If what you need is Claude, ChatGPT or Cursor reading and updating Salesforce records, Salesforce's hosted MCP servers are a third route. Our Salesforce MCP guide covers their tools, setup and billing.

Who Agentforce is a good fit for

Honestly: Agentforce makes the most sense for organizations already deep in the Salesforce ecosystem, running Sales Cloud or Service Cloud, with the budget and patience for a multi-month enterprise rollout and the internal resourcing to manage a usage-based billing model. For CRM-embedded service and sales agents inside an existing Salesforce deployment, that depth is real.

It is a weaker fit for teams without an existing Salesforce investment or teams that want an agent that works across tools beyond the Salesforce ecosystem. For that, see best enterprise AI platforms and agentic AI vs generative AI for how agent platforms differ in scope.

For a shortlist of specific tools, I sorted the best Agentforce alternatives by the job you need done.

Frequently asked questions

How much does Salesforce Agentforce cost? Agentforce currently offers three pricing models: $2 per conversation, Flex Credits at $500 per 100,000 credits, or per-user licensing from $5 per user per month (plus Flex Credits) to $125 per user per month for add-ons with unmetered employee use. Which applies depends on your deployment, and the add-ons are priced on top of existing Salesforce cloud licenses.

Does Agentforce require other Salesforce products? Mostly. Agentforce runs on the Salesforce Platform, and its per-user add-ons attach to Sales, Service, Field Service, or Industries cloud licenses, so their price is additional to those licenses. Salesforce also sells a standalone Agentforce AI-teammate license at $50 per user per month.

Why does Agentforce have three different pricing models? Salesforce introduced $2-per-conversation pricing at launch in 2024, added consumption-based Flex Credits in May 2025 for more granular action-level billing, and later added per-user licensing as a flat-fee alternative. Having all three live reflects the real difficulty of pricing agentic AI usage predictably.

How long does an Agentforce implementation take? Independent reporting describes typical implementation timelines of five to eleven months, and notes that under 10% of customers have reportedly scaled it fully. Budget implementation time and resourcing as a real cost beyond the license fee.

What is a Flex Credit in Agentforce? Flex Credits are Salesforce's consumption-based unit for Agentforce, priced at $500 per 100,000 credits. Individual agent actions consume credits based on their complexity, so the effective cost per task depends on what the agent is doing.

Is there a simpler alternative to Agentforce? Coworker has no Salesforce license prerequisite and offers 50+ tool connectors across your actual stack, not limited to the Salesforce ecosystem.

Related reading

• n8n pricing
• Guru pricing
• Moveworks pricing
• Enterprise AI pricing compared
• AI agent use cases
• Best enterprise AI platforms 2026