On Tuesday, Salesforce introduced a fully rebuilt version of Slackbot, its workplace AI assistant, marking a major evolution from a simple notification tool into a powerful AI agent. This new Slackbot is designed to search enterprise data, draft documents, and perform actions on behalf of employees, positioning Salesforce at the forefront of the emerging agentic AI movement where software agents collaborate with humans to complete complex tasks.
Now available to Business+ and Enterprise+ customers, the upgraded Slackbot embodies Salesforce’s strategic push to make Slack the central platform for AI-enhanced enterprise productivity. The launch also serves to reassure investors that AI will strengthen Salesforce’s offerings rather than disrupt them.
From Basic Helper to Intelligent Agent: The Slackbot Transformation
Parker Harris, Salesforce co-founder and Slack CTO, likened the transformation by saying, “The old Slackbot was like a tricycle, and the new Slackbot is like a Porsche.” The original Slackbot offered simple, algorithm-based functions such as reminders and notifications. In contrast, the new Slackbot operates on a robust architecture powered by a large language model (LLM) and advanced search capabilities that connect to Salesforce records, Google Drive files, calendar data, and historical Slack conversations.
Despite this overhaul, Salesforce retained the Slackbot name to maintain brand familiarity among users.
Powered by Anthropic’s Claude with Plans for Multi-Model Support
The new Slackbot is currently powered by Claude, Anthropic’s large language model, chosen primarily for compliance reasons since Slack holds FedRAMP Moderate certification to serve U.S. federal clients. Harris stated that Anthropic was “the only provider that could give us a compliant LLM” at the start of development.
However, Salesforce plans to expand AI model support within the year, including Google’s Gemini and potentially OpenAI models. Harris emphasized the commoditization of LLMs, comparing them to CPUs in their democratization and widespread availability.
On data privacy, Salesforce firmly stated it does not train AI models on customer data to prevent security risks associated with exposing confidential information.
Internal Testing Yields Rapid Adoption and Significant Productivity Gains
Salesforce conducted extensive internal testing, deploying the new Slackbot to its 80,000 employees, resulting in what Ryan Gavin, Slack’s Chief Marketing Officer, described as “the fastest adopted product in Salesforce history.” Two-thirds of employees tried Slackbot, with 80% continuing regular use and satisfaction rates hitting 96%, the highest for any AI feature by Slack.
Employees reported saving between two and 20 hours weekly, with adoption driven largely through organic social sharing rather than top-down mandates. Kate Crotty, Salesforce’s UX researcher, noted that 73% of adoption came from peer-to-peer knowledge exchange.
Delivering Actionable Insights by Synthesizing Enterprise Data
Slackbot can integrate and analyze data from multiple sources. In demonstrations, it synthesized customer feedback and usage dashboards, correlating qualitative and quantitative data to generate actionable insights. It can identify potential enterprise accounts for pilot programs and convert findings into Slack’s collaborative Canvas documents, scheduling meetings by checking stakeholder calendars.
Rob Seaman, Slack’s Chief Product Officer, highlighted that Slackbot’s ability to create shared documents and integrate with third-party tools points to its future direction of expanded AI interactions within Slack.
Pilot Customers Report Significant Time Savings and Easy Deployment
Among early adopters is Beast Industries, owned by YouTube personality MrBeast. CIO Luis Madrigal praised the straightforward implementation and quick security approvals due to Slackbot’s design that restricts data access to user permissions.
Employees at Beast Industries report saving at least 90 minutes daily, with others describing Slackbot as a vigilant assistant. Additional pilot customers such as Slalom and Engine also report time savings and improved workflow by reducing context switching.
Competing with Microsoft Copilot and Google Gemini in Enterprise AI
Slackbot enters a competitive market against Microsoft’s Copilot integrated into Teams and Google’s Gemini within Workspace. Slack executives emphasize Slackbot’s advantage lies in its seamless integration and contextual understanding of users’ work without requiring setup or training.
As Parker Harris put it, Slackbot aims to be an “employee super agent,” delivering a delightful experience akin to ChatGPT but focused on enterprise needs and workplace context.
Amy Bauer, Slack’s product experience designer, noted Slackbot’s frictionless operation, improving continuously as it learns from ongoing Slack activity without user intervention.
Vision for Slackbot as the Central AI ‘Super Agent’ in the Workplace
Salesforce envisions Slackbot as a “super agent” that can coordinate with other AI agents across an organization. The platform is already seeing third-party AI agents launch within Slack, including Anthropic’s Claude Code for coding assistance and agents from OpenAI, Google, and others.
Harris anticipates Slackbot evolving into a Model Context Protocol (MCP) client, integrating tools from across the software ecosystem to serve as the central AI hub. However, he cautions that widespread multi-agent coordination is still emerging, with significant advancements expected around fiscal year 2026.
Pricing and Potential Impact on Enterprise Costs
Slackbot is included at no extra charge for Business+ and Enterprise+ customers. However, some enterprises may experience increased costs tied to Salesforce’s broader data access pricing policies, which could affect third-party app integrations and data workflows, as noted by industry experts.
Current Features, Upcoming Enhancements, and Roadmap
The rollout to all eligible customers will complete by the end of February, with mobile support following by early March. While calendar reading and availability checking are available at launch, meeting booking capabilities will arrive shortly after. Image generation is under consideration for future updates.
Regarding integrations with competing CRM platforms like HubSpot and Microsoft Dynamics, Salesforce has not disclosed specific plans but acknowledges the competitive landscape influences development priorities.
The Future of Work: Conversational AI at the Center
Salesforce’s Slackbot launch reflects a broader industry trend toward conversational AI interfaces that prioritize natural language interaction over traditional software navigation. Harris highlighted the immense value of applying LLMs to unstructured enterprise information, unlocking insights from communication and documents inaccessible to humans at scale.
Looking forward, he expects AI agents to develop interfaces tailored to user intent rather than relying solely on conversational interactions, signaling an evolution in how workers engage with AI tools.
Microsoft, Google, and numerous AI startups share this vision, competing to embed AI seamlessly into existing workplace tools, aiming to become the invisible intelligence layer that enhances productivity.
For Salesforce, the stakes are high following a challenging year on Wall Street and concerns about AI’s impact on its core business. The company is betting that Slackbot will demonstrate AI’s potential to enhance, not threaten, its ecosystem by leveraging Slack’s vast user base.
Salesforce account executive Haley Gault summarized the transformation succinctly: “I honestly can’t imagine working for another company not having access to these types of tools. This is just how I work now.” This sentiment encapsulates Salesforce’s confidence in Slackbot’s role in the future of work.
Fonte: ver artigo original
Why Traditional Human-Centric IAM Fails for Agentic AI and What Comes Next
Harvard Dropouts Launch Smart Glasses with Always-On AI Listening Feature
Major AI Model Releases of 2025 Mark a Turning Point in Artificial Intelligence
Arm’s Vision for AI at the Edge Signals a Shift in Computing Paradigms
