Salesforce
Also known as: Agentforce, Agentforce 360, Atlas Reasoning Engine, Salesforce Inc, Data 360, Salesforce Data Cloud
CRM platform whose Agentforce layer builds and deploys autonomous agents across sales, service, and marketing, grounded in customer profiles from its Data 360 customer data platform.
Agentforce is Salesforce's platform for building and running AI agents on top of its CRM applications and data. Agents are defined in Agentforce Builder and in Agent Script, where every request passes an agent router that classifies it and hands it to the right subagent, and variables carry context through the session. Agents act through Salesforce's own automation, including flows, Apex and prompt templates, and through external MCP servers, and are grounded in the customer's material through Agentforce Data Libraries alongside Data 360 records.
Data 360, formerly Data Cloud, is the customer data platform underneath. It unifies interactions across sales, service, marketing and web into one profile per customer, resolving identities across devices, channels and anonymous sessions, and it segments and activates those profiles across Salesforce apps and channels. More than 200 connectors bring data in, and Zero Copy works in both directions, querying Snowflake, Databricks, Redshift and BigQuery in place and sharing Data 360 data back out to them. Agentic Memory in Data 360 keeps session summaries, profile facts and preferences from agent conversations for each unified individual.
Agents do not have to wait to be asked. A flow action can invoke any active agent, and headless agents run from an Apex trigger or a scheduled job as part of a background process, while customer facing agents converse through web and in app messaging, mobile and voice. When a conversation needs a person, a built in escalation hands it to a human service rep through Omni-Channel routing.
Agents are tested before release with test specs that score utterances against expected behavior and can run in continuous integration. Every session is traced step by step, from turns and model calls to actions and metric scores, and traces export as OpenTelemetry data.
Administrators choose the model for the org, and Agent Script can override it for an individual agent, across models from OpenAI, Anthropic, Google and Amazon that mostly run inside the Salesforce trust boundary. Customers can also bring their own.
Salesforce holds attestations scoped to Agentforce itself, including SOC 2 and ISO 42001, and the Einstein Trust Layer masks sensitive data, detects toxicity, keeps an audit trail and holds zero data retention agreements with model providers. Hyperforce keeps customer data in the org's country where services allow, with a paid EU Operating Zone. Agentforce is bought through Flex Credits at $500 per 100,000 credits, at $2 per conversation, or through per user licensing, on top of Salesforce edition subscriptions, and Data 360 is priced separately, with Profiles from $240 per 1,000 a year.
From our comparison video
Hightouch vs Salesforce CDP: Your Warehouse or Data 360?
Our 8 minute video puts Salesforce side by side with Hightouch. It covers what each one does and how each is priced, then when to choose which.
Recorded in October 2026. The profile here is live and updates as research is verified.
Vendor details
Canonical URL
https://salesforce.com
Category
Enterprise operations agent
Subcategory
CRM Agentforce agents
Funding status
Public company listed on the New York Stock Exchange under the ticker CRM.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Agentforce Builder connects agents to Salesforce data and channels and to existing assets including Flows, Apex, MuleSoft APIs and prompt templates, and agents can consume external MCP servers. From outside, the Agent API opens conversations with an agent synchronously or as a stream, an invocable action runs any active agent from Flow or Apex, and Salesforce Hosted MCP Servers expose Salesforce data and automation to external AI agents. Customer facing conversations run through Messaging for In-App and Web, the Agentforce mobile SDKs and voice.
In practice
Your service team drowns in routine cases. Agentforce agents answer and act across web, in app and voice channels, grounded in your own knowledge and records, and hand the conversation to a human rep through Omni-Channel when it needs one.
You want agents to act without anyone asking. A record triggered flow, an Apex trigger or a scheduled job can invoke an agent in the background, so work reaches the agent the moment the data changes.
You need agents you can test and defend. Test specs score agent behavior before release and run in continuous integration, every session is traced step by step and exports to your observability tools, and attestations such as SOC 2 and ISO 42001 are scoped to Agentforce itself.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
13.0 / 14 capabilities · 93%
| Integrations & Tool Calling | Full |
|---|---|
|
Agents act on the customer's systems through actions. Agentforce Builder connects agents to data, channels and existing Salesforce assets including Flows, Apex, MuleSoft APIs and prompt templates, and out of the box actions cover common jobs such as order and reservation management. Sourcesalesforce.com/agentforce/agent-builderread 2026-09-18 |
|
| Workflow Orchestration | Full |
|
Agent Script defines how work moves through an agent. Every request begins at an agent router that sets initial variables and classifies the request, then the agent processes a subagent and transitions between subagents along paths the script defines. Across agents, an orchestrator agent can invoke specialized agents in other Salesforce orgs and outside Salesforce through the Agent API, and a flow action can invoke an agent as one step of a longer automation. Sourcedeveloper.salesforce.com/docs/ai/agentforce/guide/ascript-flow.mdread 2026-09-18 |
|
| Knowledge Grounding & RAG | Full |
|
Agents are grounded in the customer's own material. An Agentforce Data Library connects an agent to trusted unstructured or semi structured sources, such as files or a knowledge base, which the customer adds to the library and the agent retrieves from when it answers, and new material is added to the library with no retraining. Sourcedeveloper.salesforce.com/docs/ai/agentforce/guide/headless-examples-agent-adl.mdread 2026-09-18 |
|
| Human Oversight & Guardrails | Full |
|
Escalation to a person is a built in agent utility. The utils.escalate call hands the conversation to a human service rep through an Omni-Channel routing connection the customer defines, and instructions tell the agent when to call it. Agent Script structures what the agent may do, with every request passing an agent router that classifies it before any subagent runs and explicit transitions between subagents, and Salesforce's own guidance recommends a person check model responses before they are shared in most use cases. Sourcedeveloper.salesforce.com/docs/ai/agentforce/guide/ascript-ref-utils.mdread 2026-09-18 |
|
| Security, Identity & Governance | Full |
|
Salesforce holds attestations scoped to the agent layer itself, including a SOC 2 report and a SOC 3 report for the Einstein Platform and Agentforce on Hyperforce, an ISO 42001 certificate covering the Salesforce AI Platform and Agentforce, a PCI attestation of compliance for Data Cloud, the Einstein Platform and Agentforce, a C5 report, an IRAP assessment and Spain ENS High for the Einstein Platform and Agentforce, alongside FedRAMP High and Moderate and ISO 27001 categories. Sourcecompliance.salesforce.com/enread 2026-09-18 |
|
| Observability & Auditability | Full |
|
Every step of an agent interaction is recorded. The Agentforce Session Tracing data model and the Generative AI Audit and Feedback data model capture turns, messages, LLM calls, actions, metric scores and feedback, and the Session Trace OTel API exports a full session as OpenTelemetry JSON for any observability tool or OTLP collector. Trace files can also be managed from the command line during development, and the Einstein Trust Layer keeps an audit trail of generations. Sourcedeveloper.salesforce.com/docs/ai/agentforce/guide/otel-api.mdread 2026-09-18 |
|
| Memory & State Persistence | Full |
|
Agent Script variables carry context through a session, shared by every subagent, and a session ID carries it across invocations. Beyond the session, Data 360's Agentic Memory extracts a summary of each conversation, profile facts and preferences from a registered agent's conversations. Memory is scoped to a data space and a unified individual, and agent level access control decides which agents read which memories. Conversation summaries and profile facts persist until deleted, while preferences fade unless they are reinforced, and an expired entry can be recovered by updating its validTo field. Memory Management APIs create, retrieve, update and delete memory entries, which is how a product lets people view, edit and delete their own facts and preferences. Deletion is permanent. Sourcedeveloper.salesforce.com/docs/data/agentic-memory-context memory overview and memory managementread 2026-10-25 |
|
| Deployment & Data Residency | Full |
|
Salesforce runs on Hyperforce, its public cloud infrastructure, where customer data is generally stored in the country where the customer's org is located, provided the services in use are available in that Hyperforce country, and a secure backup of an org can be held in a secondary Hyperforce region. Sourcesalesforce.com/platform/public-cloud-infrastructureread 2026-09-18 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Agentforce Builder starts agents from out of the box subagents, actions and templates that cover common jobs such as order management, reservation management and general FAQs, which the customer then extends and refines with natural language instructions. Sourcesalesforce.com/agentforce/agent-builderread 2026-09-18 |
|
| Triggers & Channel Coverage | Full |
|
Headless agents are called without a human trigger from a user interface. An Apex trigger or a scheduled job calls the agent as part of a background process, and a standard AI Agent action lets any flow, including one started by a record change, invoke an active agent with a prompt. Sourcedeveloper.salesforce.com/blogs/2025/04/build-headless-agents-with-the-agent-api.htmlread 2026-09-18 |
|
| Model Flexibility & Routing | Full |
|
An administrator sets the org level model for Agentforce on the Agentforce Agents page in Setup, and Agent Script overrides it for a specific agent or subagent with a model configuration naming any supported model. The supported list spans Salesforce managed models from OpenAI, Anthropic on Amazon Bedrock, Google Gemini on Vertex AI and Amazon Nova, most running inside the Salesforce trust boundary, and customers can also bring their own model. Sourcedeveloper.salesforce.com/docs/ai/agentforce/guide/ascript-model.mdread 2026-09-18 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
The Agent API opens a conversation with an agent, sends prompts synchronously or as a stream of server sent events marking progress, text chunks and the end of each turn, and closes the session, authenticated through a connected app with OAuth scopes, with a public Postman collection and Node.js client. Inside the platform, an invocable action runs any active agent from Apex or Flow, and Salesforce Hosted MCP Servers expose Salesforce data and automation to external AI agents and LLM applications over MCP. Sourcedeveloper.salesforce.com/blogs/2025/04/build-headless-agents-with-the-agent-api.htmlread 2026-09-18 |
|
| Testing, Debugging & Optimization | Full |
|
Agents are tested before release with scored test cases. A test spec sets out test cases as utterances with the expected agent behavior, can set context variables and include ready made metrics in the output, and is created and run in a development org from the command line or the Agentforce Testing Center, with pass and fail results. Sourcedeveloper.salesforce.com/docs/ai/agentforce/guide/agent-dx-test.mdread 2026-09-18 |
|
| Browser & Computer Use | Not documented |
|
Operating a browser or desktop interface is not part of Agentforce. Agents act through actions, flows, Apex, APIs, prompt templates and MCP tools, reached through the agent router and subagents in Agent Script. Sourcedeveloper.salesforce.com/docs/llms-genai.txtread 2026-09-18 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
Agent Kill Switch is now generally available in Agent Fabric, the Salesforce control plane for governing agents across platforms, so a team can pull an agent offline from one place. Model Wallet, also generally available, sets hard spending limits that stop requests before costs pass the budget, and the FinOps Dashboard puts AI spending in one view. Cost Optimization Policies reduce token use through semantic caching, payload compression and response trimming.
Bears on: Human approval / guardrails
View sourceAgentforce agents can now authenticate to external MCP servers with per user credentials, listed under Agentforce Interoperability in the Winter '27 release notes and rolling out to upgraded orgs from the week of October 5. Calls to an external MCP server can carry the individual user's identity instead of one shared integration login.
Bears on: Security / enterprise
View sourceSalesforce announced beta availability of Agent Optimizer, which connects agent configuration, conversational testing, and production session analysis. It can identify failure patterns, propose improvements, build regression tests, and stage changes with review points chosen by the user. The help guide documents enablement for eligible Enterprise, Performance, and Unlimited Edition customers with qualifying Einstein subscriptions.
Bears on: Observability / auditability
View sourcePricing
$500 per 100,000 Flex Credits, with a standard action consuming 20 credits ($0.10), or $2 per conversation. Per user licenses start at $5 per user a month, and the Agentforce Coworker User License is $50.
Flex Credits drawn down per agent action, a per conversation rate or per user licensing, all on top of Salesforce edition subscriptions.
Included quota
200,000 Flex Credits at no cost through Salesforce Foundations for Enterprise Edition and above.
What is public
The Flex Credit rate ($500 per 100,000), the consumption per action (20 credits, $0.10, and 30 for a voice action), Conversations at $2 each, per user prices (the Agentforce User License at $5, the Agentforce Coworker User License at $50, add ons at $125 and $150 a month, Max Edition from $550 with 2.75 million Flex Credits per org a year), worked examples of credit use by scenario, and 200,000 included credits for Enterprise Edition and above through Salesforce Foundations. Data 360 publishes Flex Credits at $500 per 100,000, Profiles at $240 per 1,000 a year and Enterprise Profiles at $420 per 1,000 a year, with ingestion free on all three.
Billing mechanics
Flex Credits are bought in advance in packs of 100,000, or under PayGo and Pre-Commit models, and drawn down per agent action, tracked in the Salesforce Digital Wallet. A Flex Agreement lets customers reallocate budget between licenses and credits. Conversations charge $2 per customer conversation, and per user licensing covers employee use. Data 360 is licensed by profiles per 1,000 a year, each carrying included Flex Credits, or by Flex Credits alone.
Cost watchouts
Credits sit on top of the Salesforce edition subscriptions the agents run in. Credit consumption varies by action type. A standard action consumes 20 credits and a voice action 30, and Salesforce's worked examples show multi step requests using more, so a chattier or more autonomous agent costs more per outcome. Flex Credits and conversation pricing are separate meters. Data 360, the customer data layer agents ground in, is priced on its own, by profiles or by Flex Credits drawn per data action.
Variable cost rationale
Flex Credits are drawn down per agent action, and the charge varies by action type. A standard action consumes 20 credits ($0.10) and a voice action 30, while multi step requests use more, so the bill grows with the number of conversations and with how many actions each one takes. Conversations are a separate meter at $2 each. Credits are bought in advance in packs of 100,000 or under PayGo and Pre-Commit models, and are tracked in the Salesforce Digital Wallet, and a Flex Agreement lets customers move budget between licenses and credits. Enterprise Edition and above include 200,000 credits at no cost through Salesforce Foundations. Add ons at $125 or $150 per user a month give employees unmetered use, while the $5 Agentforce User License still draws on Flex Credits. Data 360 adds its own meter, Flex Credits drawn per data action or profile licenses bought per 1,000 a year.
Additional watchouts
Credit burn scales with action complexity and volume, so cost follows the number of actions in each conversation, not just the agent count. Agents also need the Salesforce platform subscriptions and clouds they run in, which are priced separately.
Sales call required
Yes, required for paid access
Free / trial
Developer Edition for building, and 200,000 Flex Credits at no cost through Salesforce Foundations on Enterprise Edition and above
Lowest paid plan
Flex Credits at $500 per 100,000 credits
Key ambiguities
Credit, conversation, per user add on and Data 360 profile prices are public, while the Salesforce edition subscriptions underneath, and the terms of each buying model, come through sales.
Missing data
Prices for the Salesforce edition subscriptions the agents run in, the full rate card for usage types beyond standard and voice actions, and the storage and credits that come with Data 360's free provisioning.
Related vendors
- 11th Estate — Agentic platform whose AI engine scans markets, matches an…
- Abacum — FP&A platform for growing companies whose AI Modeler builds, edits…
- Adopt AI — AI native accounting and bookkeeping firm whose agents run…
- Aera Technology — Decision intelligence platform where an always on agent executes…
- Akro AI — On premise operational intelligence that automates document heavy…
- alfred_ — AI executive assistant that triages the inbox overnight and scores…
Alternatives to Salesforce
The closest documented capability profiles to Salesforce among enterprise operations agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Atlassian Rovo13.5 / 14Adds documented Browser & Computer Use
- Boomi12.5 / 14A lighter documented profile than Salesforce
- Dust12.5 / 14A lighter documented profile than Salesforce
- Harvey12.5 / 14A lighter documented profile than Salesforce
- IBM watsonx Orchestrate12.5 / 14A lighter documented profile than Salesforce
- Atomicwork12.0 / 14A lighter documented profile than Salesforce
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded