Salesforce is introducing a new approach to workplace AI that moves beyond general-purpose assistants toward specialized “digital teammates” that operate directly within Slack conversations, the company announced Monday.
The new product, called Agentforce in Slack, allows companies to build and deploy multiple task-specific AI agents that can search across workplace conversations, access company data, and perform actions within the messaging platform where many employees already spend their workday.
“Similar to how employees with specialized skills and knowledge collaborate to solve problems, our customers need AI agents that collaborate to solve problems for both their own customers and their employees,” said Rob Seaman, Chief Product Officer of Slack at Salesforce, in an exclusive interview with VentureBeat.
The launch represents a significant shift in enterprise AI strategy. While most companies began their AI journeys with broad “do-everything” assistants, many organizations found these generalists struggled with complex or specialized tasks. Salesforce’s move signals an emerging consensus that purpose-built AI may deliver more reliable results than jack-of-all-trades approaches.
“Conventional AI assistants are limited to having conversations — such as the customer service chatbots you might find online,” explained John Kucera, SVP of Salesforce AI in an interview with VentureBeat. “However, with these ‘digital teammates’ in Agentforce and Slack, AI agents can reason and take multi-step actions like orchestrating workflows, interacting with multiple systems across the business, and adapting their approach to the specific task at hand.”
Ready-made AI templates promise faster time-to-value for enterprise teams
Salesforce is also introducing Slack Employee Agent Templates starting June 12, which will provide pre-configured starting points for common use cases. These include Customer Insights (accessing and summarizing CRM data), Onboarding (guiding new employees), and Employee Help (answering common HR, IT, and operational questions).
The templates aim to address a common complaint about AI implementations: the extensive customization typically required before seeing tangible business value.
“Companies need to make great AI agents that can accurately handle a wide range of tasks and questions, and they need the flexibility of LLMs to answer questions with repeatable logic and policy adherence,” Kucera said.
To support these agents, Salesforce is introducing two key capabilities: Slack Enterprise Search, which allows agents to search across messages, files, and apps in real-time, and General Slack Topic, which enables agents to understand and respond within the flow of team conversations.
Enterprise-grade guardrails: How Salesforce balances AI power with data protection
The company emphasized that Agentforce respects existing user permissions in both Salesforce and Slack, addressing a common concern about AI systems potentially exposing sensitive information.
“Agentforce in Slack operates within the same security frameworks as Salesforce and Slack. This ensures that customer data is handled with the utmost care and that Agentforce’s actions and responses are generated only using data that the user is also authorized to access,” Seaman said.
In a move to broaden adoption, Salesforce is also enabling companies to grant non-Salesforce users access to Agentforce through no-cost Salesforce Identity licenses, allowing any Slack user to interact with the AI agents.
ROI reality check: Customer implementations show measurable productivity gains
Several companies have already deployed early versions of Agentforce with promising results. Paper tablet maker reMarkable reported that the technology provided faster support across IT and internal knowledge needs. Formula 1 saw fan questions answered 80% faster with first-call resolution rates exceeding 95%.
Perhaps most telling are Salesforce’s internal results with its Engineering Agent in Slack. “In just six months, the Engineering Agent has handled over 18,000 conversations across 3,500 users. On average, it answers 130 questions per Slack channel per month, saving about 30 hours per channel per month,” Seaman said. “Projecting forward based on its current volume, it could deflect over 50,000 conversations annually, saving 17,000 hours, equivalent to more than 8 full time employees and $1.4M!”
Other customers have reported similarly impressive metrics. OpenTable handled 73% of restaurant web queries in just three weeks with Agentforce, while payment processor Engine reduced average handle time by 15% and projects $2 million in annual cost savings.
Slack’s conversational edge: How Salesforce plans to outmaneuver Microsoft and Google
The announcement positions Salesforce against Microsoft’s Copilot and Google’s Gemini in the increasingly competitive enterprise AI assistant market. Each company is leveraging its platform strengths — Microsoft with Microsoft 365, Google with Google Workspace, and now Salesforce with Slack and its CRM ecosystem.
“Agentforce supports customer-facing, employee-facing, and background AI agents all on the same platform. And those agents can be deployed wherever customers and employees engage – from websites, to Salesforce applications, to Slack,” Kucera said. “And with Data Cloud’s deep integration with Agentforce, customers have low-code tools to unlock trapped data from other systems for these AI agents to use.”
Seaman highlighted Slack’s unique position in the collaboration landscape as a key advantage. “Slack also consolidates conversational data in a way that our competitors can’t, making it a unique knowledge source for these agents and providing more robust context for them to take action on. Think of all the conversations, video transcripts, canvases, apps, and workflows shared in Slack on a daily basis!”
Digital teammates: Reimagining workplace collaboration between humans and AI
The introduction of specialized AI agents raises questions about how work will evolve as companies delegate increasingly complex tasks to digital assistants. Both Salesforce executives stressed that the goal is to augment rather than replace knowledge workers.
“Companies are seeing the biggest impact right now with agents tackling lower-level, time-consuming tasks to free humans to refocus on work that requires creativity and critical thinking, like fostering deeper customer relationships, innovating new solutions, making nuanced judgments in complex scenarios, and driving business growth,” Kucera said.
Seaman envisions a future where employees spend less time switching between different applications. “Instead of swivel chairing between tools, employees will be able to simply tell an agent what they need, and an agent can take care of the rest. The traditional app that powers your workflow in Slack or holds data from another system doesn’t go away in this instance. Rather, the agent becomes the front-end and the go-between with the apps.”
From jack-of-all-trades to specialists: The evolution of enterprise AI strategy
Salesforce’s move reflects a broader realization that general AI assistants often struggle with the complex, context-dependent nature of enterprise workflows. By narrowing an agent’s focus, companies can reduce what Salesforce calls “semantic noise” — the distractions that occur when an AI tries to reason across too many possible actions.
“We’re in the very early days of working with AI agents in the everyday course of business, with most companies just starting to think about how they can augment their human workforces,” said Kucera. “I think that over the coming years, we’ll see the relationship transform into a more robust partnership.”
The shift toward specialized agents also mirrors how human work is organized in most enterprises, with specialists handling specific domains rather than generalists trying to master everything.
“Every role and team has a business process or need that can be fulfilled with Agentforce in Slack,” Seaman said. “We have to start preparing employees for an agentic future, and Slack is the natural home for this to occur because it’s the carrier of an organization’s rich knowledge, culture, and, with Salesforce, customer data.”
As companies continue experimenting with AI in the workplace, Salesforce’s specialized approach may signal a new phase in enterprise AI adoption — one that focuses less on headline-grabbing general intelligence and more on reliable, domain-specific assistance that integrates directly into existing workflows.
The post Salesforce just unveiled AI ‘digital teammates’ in Slack — and they’re coming for Microsoft Copilot appeared first on Venture Beat.
