Salesforce has unveiled a new series of AI agents designed to boost productivity for sales, technical support and customer service teams.

According to a recent report by SiliconANGLE, Salesforce introduced these agents along with an update to Agentforce Co-Worker, an automation tool embedded in multiple cloud services.

Among them, Hunter helps B2B sales staff find potential customers and write sales emails. It then automates deal stages. Salesforce explained that Hunter also automates tasks such as preparing for presentations and drafting proposals.

B2B deals typically take a considerable amount of time to close. Salesforce said Hunter can also cover such long-term tasks based on the Long Horizon Runtime module. Long Horizon Runtime helps agents make long-term plans and collect data across multiple conversation sessions. Long Horizon Runtime currently applies only to Hunter, and Salesforce plans to roll it out to other agents in stages.

Along with Hunter, Salesforce also unveiled sales-focused agents Piper and Carter. Companies can embed these agents into their websites. Piper greets B2B leads and connects them to the appropriate sales representative. Carter helps online shopping malls respond to consumer product inquiries and also provides an in-chat payment widget.

Salesforce also introduced two new agents for customer support teams, Casey and Fin. Casey automates processing of product return requests. Fin, based on technology from Fin that Salesforce acquired this year for $3.6 billion, is used for similar tasks.

Salesforce also released automation tools for internal help desks. Page handles employee HR inquiries and technical support requests. Marshall automates repetitive tasks for supply chain teams.

Companies can customise agents using the Agent Script programming language. Developers can use Agent Script to specify that agents perform critical tasks in prescribed ways, reducing the risk of hallucination.

The updated Agentforce Co-Worker supports corporate users in using business data such as customer records and managing agents created by colleagues.

The core of the update is a multi-agent orchestration function. It supports Agentforce Co-Worker in assigning complex tasks by splitting them among multiple specialised sub-agents. In addition, an AI Skills function makes it easier for users to customise how Co-Worker operates.