What Is Agentforce?
Agentforce is Salesforce's platform for building AI agents that do real work inside your CRM. An agent can pick up a service case or qualify a lead, then plan the steps and act on its own, staying inside the guardrails you set. Salesforce calls the technology behind it the Atlas Reasoning Engine.
Salesforce now calls it Agentforce 360. It builds autonomous agents that work across Sales, Service, and its other clouds. You define each agent by five things: its role, the data it sees, the actions it can take, its guardrails, and where it runs. Agents reason with the Atlas engine and pull data from Data 360, formerly Data Cloud.
How does Agentforce work?
Every agent runs on the Atlas Reasoning Engine, which Salesforce calls the platform's brain. It doesn't follow a fixed script. Instead it loops: reason about the next step, take an action, look at what happened, then go again until the request is done. Salesforce says Atlas leans on deliberative System 2 reasoning and retrieval to keep answers tied to real data, and it uses reinforcement learning and feedback loops to cut down on hallucinations.
The agent repeats this loop until the request is done.
How do you build an Agentforce agent?
Salesforce says an agent comes down to five attributes. You set them up in Salesforce's low-code tools, with no separate stack to stand up.
The agent's job or purpose.
The knowledge and information the agent can access.
The tasks and capabilities the agent can carry out.
The boundaries and restrictions on what the agent may do.
Where the agent operates, such as chat or a specific surface.
Source: Salesforce's own descriptions of the Atlas Reasoning Engine and Agentforce agents, verified July 2026.
What can Agentforce agents do?
Salesforce ships more than 200 prebuilt templates, topics, and actions, plus agents you can drop in and adjust. Two get named most often:
- Service Agent. Handles a wide range of service issues without pre-scripted flows, in place of a traditional chatbot.
- SDR (Sales Development Representative) Agent. Works prospects around the clock, answering questions, handling objections, and booking meetings from CRM and outside data.
Does Agentforce use Data Cloud?
Salesforce points to Data 360 for this. Data 360 is the old Data Cloud, renamed on October 14, 2025. It is the data engine that unifies company data and feeds Agentforce agents through Zero Copy, and Salesforce recommends it as the way to get reliable data into an agent.
Where does GPTfy fit?
GPTfy takes a different route. It is a Salesforce-native managed package that installs in the org you already run. The idea is bring-your-own-model: connect the AI you already pay for, whether that is OpenAI, Anthropic's Claude, or Google Gemini, and point it at your Salesforce data. It runs inside your org and needs no Data Cloud. If you would rather pick your own model and keep the AI close to your data, that is the case GPTfy is built for.
Agentforce FAQ
What is Agentforce?
Agentforce is Salesforce's platform for building autonomous AI agents that handle work inside the CRM, like resolving a service case or qualifying a lead. Salesforce now brands the platform Agentforce 360. Each agent stays inside the guardrails you set and runs on the Atlas Reasoning Engine.
How does Agentforce work?
Agents run on the Atlas Reasoning Engine, which Salesforce calls the platform's brain. The agent reasons about a step, takes an action, checks the result, and goes again until the request is done. It uses retrieval to keep its answers tied to your data.
What is the Atlas Reasoning Engine?
It is the reasoning system behind every Agentforce agent. Salesforce describes it as deliberative System 2 reasoning paired with retrieval-augmented generation, plus reinforcement learning and feedback loops meant to improve accuracy and cut hallucinations.
What can Agentforce agents do?
Salesforce ships more than 200 prebuilt templates, topics, and actions, along with ready-made agents. The Service Agent handles service issues without pre-scripted flows, in place of a chatbot. The SDR Agent works prospects around the clock, answering questions, handling objections, and booking meetings from CRM and external data.
How do you build an Agentforce agent?
Salesforce says an agent comes down to five attributes: a role (its purpose), the data it can reach, the actions it can take, the guardrails on those actions, and the channel where it runs. You configure it in Salesforce's low-code tools.
Does Agentforce use Data Cloud?
Salesforce points to Data 360, the former Data Cloud, renamed on October 14, 2025. It is the native engine that unifies company data and grounds Agentforce agents through Zero Copy. Salesforce recommends it as the way to bring reliable data to an agent.
How is Agentforce different from Einstein?
Einstein is Salesforce's wider set of AI features built into its clouds, like predictions and generative assists. Agentforce is the platform for autonomous agents that reason and act on their own. They are related parts of Salesforce's AI stack, not the same product.
Run the numbers on AI agents for your Salesforce data
Model the cost and payback for your own team, then see how a bring-your-own-model setup on the org you already run stacks up for your use case.