Salesforce Inc. today unveiled Agentforce, a new artificial intelligence autonomous agent platform designed to augment work for employees and handle tasks across service, sales, marketing and commerce.

Using Agentforce, companies can build customizable AI agents that can scale according to workforce demands which can analyze data, make decisions, understand customer needs and take actions on their behalf. According to Salesforce, any organization can easily build, set up and deploy these agents for any use case right now.

“Agentforce represents the Third Wave of AI — advancing beyond copilots to a new era of highly accurate, low-hallucination intelligent agents that actively drive customer success,” said Marc Benioff, chair and chief executive of Salesforce. “Unlike other platforms, Agentforce is a revolutionary and trusted solution that seamlessly integrates AI across every workflow, embedding itself deeply into the heart of the customer journey.”

The platform launched with out-of-the-box agents that customers can easily customize for their own needs in a few clicks and without the need to write any code. The agents are designed to be set up and deployed in minutes.

The different available roles include a service agent that can replace a traditional chatbot and can handle a large range of service issues without preprogrammed rote scenarios, a sales development representative that can engage prospects and a sales coach that can perform role-play for sales teams. E-commerce merchandisers can also spin up and launch a merchant agent that can assist with site setup, goals, personalized promotions, product descriptions and daily tasks.

Agents have five attributes that allow them to act autonomously: role, data, actions, guardrails and channels. This means that they have a job that they do, what knowledge they access, their capabilities, their rules and guidelines, and the areas that they’re set to work within. Salesforce already brings all of this together in its platform through its platform with Data Cloud for data, application programming interfaces and partner tools for capabilities, as well as built-in security, filters and management.

All of this comes together with a new reasoning engine called Atlas, which uses the role to evaluate what actions and data it has available to generate an automated series of tasks. The company said Atlas can simulate how humans think and plan, allowing it to make decisions and complete business tasks based on data and software tools.

For example, a service agent could provide personalized support for a customer by reasoning around things such as completing orders, returns or other customer service operations. These actions can take place over text and voice – and on voice, they feel almost as natural as talking to a human with the AI mimicking normal conversation. And, if the agent cannot handle the user needs it can always escalate to a human operator.

As mentioned above, customization and building agents is made easy using the Agent Builder, where users don’t need to worry about understanding AI prompts but can define the role and then assign actions and capabilities with a low-code interface. For advanced users, there is a Model Builder that allows customers to assign large language models of their choice to agents, test them in a playground and activate them and a Prompt Builder to test out specialized actions.

“As AI continues to evolve, I’m seeing a fundamental shift in how work is done, with the emergence of agents playing a central role in driving efficiency and scale,” said Ritu Jyoti, group vice president for AI and data market research and advisory at International Data Corp. “Salesforce’s new Agentforce stands out as a promising suite of tools, offering businesses a comprehensive platform of intelligent agents designed to automate and streamline tasks across a wide range of industries.”

To add to the platform’s skills, Salesforce partnered with Amazon Web Services Inc., Box Inc., Certinia, Copado Inc., Google LLC, Honeywell International Inc., IBM Corp. and more. These partners have made their services available through Agentforce to build more than 20 agents and agent actions that are available through AppExchange. Customers can use them to customize the AI agents for specialized tasks and to take unique actions, even outside of Salesforce.

Image: SiliconANGLE/Microsoft Designer