Salesforce Agentforce sells on conversation credits. The list price runs $2 per conversation. The Data Cloud prerequisite, the action metric, and the agent type choice each move the realized rate by twenty to forty percent at renewal.
Agentforce is the Salesforce platform for building and deploying autonomous AI agents. The product launched in late 2024 and rolled out across Service Cloud, Sales Cloud, and Commerce Cloud through 2025 and 2026. The license metric is the conversation credit, with a list price of $2 per conversation in the Standard tier.
Two prerequisites push the bill up. Data Cloud is required to ground the agent on enterprise data. The action metric counts each external system call the agent makes, with a small additional fee. The combined effect lifts the realized cost per conversation thirty to fifty percent above the headline $2.
Read this guide alongside the Salesforce knowledge hub, the Salesforce advisory practice, the AELA reference, the 2026 pricing guide, the renewal playbook, and the Vendor Shield subscription.
A conversation credit is a single end to end interaction with an Agentforce agent. The interaction can include multi turn dialogue, tool calls, and reasoning steps, but it counts as one credit. A new conversation starts when the user begins a new session.
The most common counting trap is the multi agent workflow. A retail agent that hands off to a billing agent then to a returns agent burns three credits for one customer issue. The cost model needs to count credits at the workflow level, not the agent level.
Salesforce ships two default agent types and a custom builder. Service Agent runs on Service Cloud data. Sales Agent runs on Sales Cloud data. Custom agents run on Agentforce Studio with developer defined skills.
| Agent | Underlying cloud | List per conversation | Common use |
|---|---|---|---|
| Service Agent | Service Cloud | $2.50 | Tier 1 customer support |
| Sales Agent | Sales Cloud | $2.00 | Inbound lead handling |
| SDR Agent | Sales Cloud | $2.00 | Outbound prospecting |
| Commerce Agent | Commerce Cloud | $2.50 | Product discovery |
| Custom Studio | Any | $2.00 | Internal workflow agents |
Each agent type ships with default skills. Custom skills run inside Agentforce Studio and add no incremental license fee beyond the conversation credit. The trade off is build versus buy: default agents deploy in weeks, custom agents take quarters but bend to the workflow.
Agentforce cannot ground on enterprise data without Data Cloud. The agent retrieves customer context, account data, case history, and product information through Data Cloud queries. The Data Cloud commit becomes part of the Agentforce envelope.
Most Agentforce conversations consume two to four Data Cloud Credits in addition to the conversation credit. The combined cost per conversation lands at $2.30 to $2.80 in the Service Agent tier, not the headline $2.50. Model the combined credit consumption before signing, not after.
The Salesforce account team rarely volunteers the combined math. Demand the worked example before the renewal goes to legal.
List price is the starting point. Volume commits, multi year terms, AELA absorption, and EA alignment each move the realized rate. The mid market negotiation benchmark sits at twenty to thirty percent off list. The large enterprise benchmark sits at thirty to forty percent.
| Annual conversations | List value | Discounted target | Realized per conversation |
|---|---|---|---|
| 100,000 | $200,000 | $170,000 | $1.70 |
| 500,000 | $1.0M | $750,000 | $1.50 |
| 1,000,000 | $2.0M | $1.3M | $1.30 |
| 5,000,000 | $10.0M | $5.5M | $1.10 |
The Salesforce AI Enterprise License Agreement (AELA) bundles Agentforce, Data Cloud, and the underlying clouds at a flat enterprise rate. AELA discounts the combined envelope further but locks the buyer into a three year minimum and a defined seat count. AELA is the right path for buyers carrying one million plus conversations annually.
The lever set runs at renewal, not in between. Conversation volume commits, agent type rationalization, Data Cloud right sizing, and AELA absorption each contribute to the realized rate.
Agentforce conversation credits look like a clean per call meter. The combined Data Cloud and action cost runs thirty to fifty percent above the headline rate. Model the combined credit before signing, because the rate holds for the term and renegotiation only happens at renewal.
The seven step checklist below is the buyer side starting position for any Agentforce engagement.
Agentforce conversation credits list at $2.00 per conversation in the Standard tier and $2.50 in the Service and Commerce tiers. The list is the starting point for negotiation. Volume commits, multi year terms, and AELA absorption typically pull the realized rate to $1.10 to $1.70 per conversation at enterprise scale.
Yes. Agentforce cannot ground on enterprise data without an active Data Cloud subscription. The agent retrieves customer context, account data, case history, and product information through Data Cloud queries. Each grounding query consumes Data Cloud Credits in addition to the Agentforce conversation credit.
A conversation credit is a single end to end session with one agent. Internal escalation from one agent to another inside Agentforce starts a new conversation and burns a new credit. Hand off to a live human ends the agent conversation. Multi agent workflows therefore consume more credits than they appear to.
The Salesforce AI Enterprise License Agreement (AELA) bundles Agentforce, Data Cloud, and the underlying clouds at a flat enterprise rate. AELA usually wins for buyers carrying one million plus conversations annually because the bundled discount lands deeper than the standalone Agentforce order. AELA locks a three year minimum and a defined seat count.
Each external system call counts as an action and bills at additional cents per action. A typical agent conversation runs two to four actions, adding $0.05 to $0.20 to the cost per conversation. Combined with Data Cloud Credits the realized cost per conversation runs thirty to fifty percent above the headline credit rate.
Redress runs Agentforce engagements inside Vendor Shield, the Renewal Program, and the Benchmark Program. The work covers the conversation volume forecast, the combined credit cost model, the agent type mix, the Data Cloud right sizing, the AELA absorption decision, and the multi year ramp negotiation. Always buyer side, never Salesforce paid.
Redress runs Salesforce Agentforce engagements inside the Vendor Shield subscription, the Renewal Program, the Benchmark Program, and the Software Spend Assessment.
Read the related benchmarking framework, about us, locations, and contact pages.
A buyer side reference on the Agentforce credit math, the Data Cloud right sizing, the AELA absorption decision, and the multi year ramp negotiation. Built from hundreds of Salesforce engagements.
Independent. Buyer side. Written for CIOs, CFOs, and procurement leaders carrying Salesforce estates. No Salesforce influence. No sales kickback.
Open the white paper in your browser. Corporate email only.
Open the Paper →Agentforce conversation credits look like a clean per call meter. The combined Data Cloud and action cost runs thirty to fifty percent above the headline rate. Model the combined credit before signing, because the rate holds for the term and renegotiation only happens at renewal.
We have run 500+ enterprise clients across 11 publishers. Every engagement starts with one conversation.
Agentforce volume forecasts, Data Cloud absorption maths, AELA versus standalone benchmarks, multi year ramp shapes, and the wider Salesforce commercial leverage signals across every renewal we run.
Once a month. Audit patterns, renewal benchmarks, vendor commercial signals across Oracle, Microsoft, SAP, Salesforce, IBM, Broadcom, AWS, Google Cloud, ServiceNow, Workday, Cisco, and the GenAI vendors. No follow up sales pressure.
Free providers (Gmail, Yahoo, Outlook) cannot subscribe. Work email only. Unsubscribe in one click.
