Salesforce has launched Agentforce 360, an AI-driven platform designed to embed virtual agents across its ecosystem, including Slack, to automate routine work and unify data, people, and AI.

Industry experts emphasize the importance of clear business problem definition and realistic expectations to ensure successful AI deployment and avoid project failures.

Real-world deployments of Agentforce have demonstrated significant ROI, with companies like DirecTV, Under Armour, and Heathrow Airport reporting substantial savings, increased efficiency, and improved customer experiences.

Salesforce reports that over 12,000 customers, including Reddit and Adecco, are already using Agentforce, with plans to expand through its AppExchange for industry-specific modules.

Market analysts warn of high adoption risks, with Gartner predicting that 40% of agentic AI projects may be abandoned by 2027 due to unclear ROI and rising costs, despite the technological advancements.

Salesforce is addressing industry shifts, including layoffs of 4,000 customer service agents replaced by AI, emphasizing that AI will augment human productivity rather than replace jobs.

At the upcoming Dreamforce conference, Salesforce is expected to showcase enterprise customer adoption of Agentforce, highlighting broader deployment and measurable business value.

The platform connects users, data, and AI agents across the enterprise, with upcoming features like enterprise search, a context-aware Slackbot, and integrations via Model Context Protocol (MCP).

Salesforce is also integrating reasoning capabilities into its AI, including hybrid reasoning that combines predetermined workflows with flexible language model reasoning, monitored through dashboards for accuracy and compliance.

Agentforce 360 includes controls such as Agent Script for defining agent behavior, reasoning models from major AI providers, and a no-code Agentforce Builder for designing and testing agents to enhance reliability and trust.

Key features of Agentforce 360 include a conversational builder for natural language agent creation, Intelligent Context for unstructured data access, voice capabilities, and multi-agent orchestration, all aimed at balancing predictability and creativity.

Upcoming features like the beta launch of Agent Script in November will enable AI agents to respond to complex scenarios using flexible prompting and a unified platform for building and deploying agents.

The company envisions an agentic enterprise as a new work paradigm where AI amplifies human potential, supports 24/7 operations, and drives productivity and customer engagement.