(Video = Salesforce)
Salesforce is introducing a brand new artificial intelligence (AI) agent platform that connects existing business apps to automate customer support tasks. The platform is described as going beyond traditional chatbots to handle tasks that require multiple steps and decision-making.
VentureBeat reported on the twelfth (local time) that Salesforce has launched a brand new AI autonomous agent platform called ‘Agentforce’ designed to enhance the work of employees and handle tasks across service, sales, marketing, and commerce.
AgentForce allows businesses to construct custom AI agents that may scale with their workforce needs, analyze data, make decisions, understand customer needs, and take motion on their behalf. Salesforce says, “Any business can easily construct and configure these agents, and deploy them for any use case.”
It is available in the shape of a pre-built agent that users can easily customize with just just a few clicks. It could actually be arrange without writing any code, and the agent is designed to be arrange and deployed in minutes.
Available roles include a “service agent” that may replace traditional chatbots and handle a wide range of service issues without pre-programmed scenarios; a “sales coach” for sales teams to practice prospecting and fascinating with potential customers; and e-commerce managers can create and deploy “store agents” to assist with site setup, goal setting, personalized promotions, product descriptions, and day-to-day operations.
Typically, agents have five properties that enable them to act autonomously: ▲the tasks to be performed (role), ▲the knowledge to access (data), ▲the functions that might be performed (actions), ▲the foundations and guidelines (guardrails), and ▲the realm during which to work (channel).
Salesforce cited the Data Cloud as providing functionality, security, filters and management capabilities through data, APIs and partner tools, and integrating all of those elements into the platform.
These elements work along with a brand new inference engine called Atlas, which evaluates possible actions and data and generates a series of automated tasks. Salesforce explains that Atlas can simulate the way in which humans think and plan, enabling them to make decisions based on data and complete business tasks.
For instance, a service agent can provide personalized assistance to a customer, resembling completing an order, returning a product, or other customer support tasks. This might be done via each text and voice. If the agent is unable to handle the user’s request, the request might be passed on to a human operator at any time.
‘Agent Builder’ makes customization and deployment easy. Users can define roles and assign tasks and functions through a low-code interface without having to grasp AI prompts.
For advanced users, the ‘Model Builder’ is provided to pick a Large Language Model (LLM), test it, and activate it. The ‘Prompt Builder’ can be used to check specialized tasks.
To expand the capabilities of the platform, Salesforce has partnered with Amazon Web Services (AWS), Box, Certian, Copado, Google, Honeywell International, and IBM, who’ve already deployed greater than 20 agents and agent operations through AgentForce.
Salesforce emphasized that “AgentForce is the one tool that brings together humans, AI, data, and behaviors,” and that “it should prepare you for the following step because the world’s #1 CRM for AI.”
Reporter Park Chan cpark@aitimes.com
