Top Innovations in the Winter ’27 Release: AI Agents Run the Work, Freeing the Enterprise to Work Smarter and Faster

The Winter '27 Release marks a fundamental shift in how AI agents operate within the enterprise. Instead of assisting employees one task at a time, agents now run entire workflows - qualifying pipeline, resolving service cases, booking appointments, underwriting risk. They operate from start to finish on every channel companies already use, from Slack and Teams to voice and legacy enterprise systems. None of that works without trust. Every agent acts on governed CRM data, making it completely safe to hand off entire workflows rather than just single questions.

This combination of AI flexibility and CRM reliability turns autonomous action into concrete business results.

The proof is already in production. Autonomous Scheduling with Voice cuts 15-minute booking calls down to seconds. Agentic Commerce Search drives a 13% conversion lift for shopper agents live today. Four customers, including PowerSchool with more than 550 users, are already running Adaptive Experience in production. The result: fewer handoffs, faster resolutions, and AI that drives outcomes instead of just generating answers.

Most importantly, this shift empowers your teams to do what humans do best: build relationships; make judgment calls; and solve complex, out-of-the-box problems. For the business, this translates to faster-closing revenue, first-contact service resolutions, and reduced operating costs as manual tasks disappear. The Winter '27 Release makes the enterprise faster, leaner, and far more responsive.

Top 10 updates in the Winter '27 Release:

• Slack Code: Turn agentic coding into a multiplayer experience
  Slack Code brings software development out of isolated private tabs and into Slack, giving teams and AI agents dedicated code channels to write, review, and ship code together in the open. Organizations can now close the gap between solo AI development and team collaboration by connecting code execution directly to conversational context in one shared workspace, empowering developers, product managers, designers, and nontechnical teammates to build software together, from idea to shipped product, in one place in Slack.

• Adaptive Experiences & Dynamic Plans: Resolve complex cases with real-time AI guidance
  Service teams can now resolve complex, evolving cases faster with Adaptive Experiences & Dynamic Plans, a capability of the Service Rep Assistant. An ambient AI agent listens to case and messaging conversations in real time, reasoning over live context to generate and update resolution plans as issues change. Four customers are already live in production, including PowerSchool with more than 550 users, as the broader Service Rep Assistant program surpasses 100 customers.

• Agentforce Contact Center: Expands globally to over 30 countries
  Agentforce Contact Center is now available in over 30 countries, giving global contact centers the ability to bring CRM, channels, and AI together in a single solution for the first time. With all channels - including voice - built directly into the CRM, human and AI agents can access real-time customer data to deliver proactive, personalized service without third-party contact center integrations.

• Tableau Knowledge: Ground every AI agent in governed business context
  Tableau Knowledge unlocks trapped enterprise data by automatically ingesting structured and unstructured data, generating living knowledge graphs, and delivering curated context to agents via open standards. Business experts stay in control, agents stay accurate, and AI investments deliver true ROI.

• Informatica Headless: Access and trust data, power AI agents faster
  Enterprises building AI agents can now access governed enterprise data without leaving their tools. Informatica Headless decouples IDMC's backend from its UI, so developers call enterprise-grade data management directly inside VS Code, Cursor, and Slack or via universal MCP servers. Teams stay in their native workflow; cloud compute costs drop by eliminating hallucination loops; and AI agents get clean, governed data they can rely on.

• Segmentation and Activation: Build and activate audiences in plain language Marketers lose weeks to data modeling and multichannel campaign configuration. Agentic Segmentation and Activation removes both bottlenecks. Users describe target audiences in plain language, and the Segmentation agent builds a complete audience plan while verifying the data. The Activation agent then sets up campaigns conversationally across an expanded set of ad platforms, flagging errors and suggesting optimizations before reaching external partners.

• Autonomous Scheduling with Agentforce Voice: Scale dispatch operations, not headcount Cut 15-minute appointment bookings down to seconds. Autonomous Scheduling with Agentforce Voice lets customers book appointments 24/7 using agents that check real-time availability and find the best technician, in the right territory, the moment they call. No human handoff required.

• Agent Skills and Plugins: Governed agent skills and plug-ins across every surface
  Every new AI agent starts with zero context of how an organization works, forcing it to guess at best practices agent after agent - guesses that produce inconsistent results no matter how capable the underlying model is. Agent Skills and Plugins closes that gap with a unified strategy for discovering and reusing skills across every surface. Salesforce has launched with over 100 prebuilt skills covering common platform workflows, and companies can author and publish their own skills into the same registry. Plug-ins bundle skills together with the MCP connectors, hooks, and configuration an agent needs, so a single install brings a complete, governed capability to any surface - starting with the Salesforce Development plug-in, live today in the Claude Code marketplace. Developers build reusable capabilities once; agents execute them consistently and securely across every surface an enterprise runs.

• Third-Party Agent Orchestration: Connect Agentforce to any enterprise agent
  Enterprises running multi-vendor AI agent environments no longer face fragmented handoffs and expensive point-to-point integrations. Agentforce's native Third-Party Orchestration extends Agentforce across any A2A-compliant ecosystem - enabling Agentforce to orchestrate agents from AWS, Azure, Google, and more or to participate as a subagent in architectures where external orchestrators already exist.

• Agentic Commerce Search: Boost conversion with intent-driven product search Shoppers using Agentforce Commerce now find products through natural conversation instead of exact keyword matches. Agentic Commerce Search trains a dedicated small language model on each merchant's catalog and shopper behavior, interpreting natural-language queries to surface relevant products in real time. Early results show a 13% conversion lift and a 17% increase in add-to-cart rates, and the capability is now standard for all Shopper Agent customers.

5 more updates worth noting

While those are incredible updates, the pattern holds across the rest of the release. Five more updates extend agents further into commerce, marketing, revenue, insurance, life sciences, and the front line - freeing every part of the business to move faster and focus on higher-value work.

• Slack Frontline: Unify frontline and corporate teams in one work OS
  Slack Frontline brings a purpose-built Slack experience to shift-based workers, giving them a simple way to stay connected with the rest of the enterprise. Customers can now close the gap between HQ and the floor by centralizing all conversation, company news, tasks, and support in one governed work operating system, providing both halves with the access and information they need to work effectively on a day-to-day basis.

• Agentforce Marketing Goals Agent: Shift campaign execution from reactive to goal-driven autonomy
  Marketing teams have long struggled to coordinate campaigns across channels without a system that arbitrates conflicting goals, prevents over-messaging, and ties activity to business outcomes. Agentforce Marketing Goals Agent changes that equation. Marketers set goals, budgets, and guardrails, while the agent orchestrates and optimizes every campaign in real time. With a pilot program that has organizations already live, early customers report a fundamental shift from managing campaigns to owning strategy.

• Revenue Management Agent Enhancements: Expands agent quoting across buyers, partners, and account executives
  Agentforce Revenue Management Agent Enhancements extends quoting and renewal capabilities across three motions: self-service quoting for customer buyers, integrated quoting for partners, and automated renewal package assembly for account executives (AEs). Customer buyers now order without a rep, partners transact without leaving their workspace, and AEs arrive at renewals prepared with a drafted quote and risk signals - accelerating time to cash across the revenue lifecycle.

• Claims Service Customer Assistance: Automate FNOL intake with 24/7 voice-driven claims access
  Insurance carriers now have a voice-enabled agent template that lets policyholders file First Notice of Loss (FNOL) claims by phone and other channels at any time of day, no customer service rep (CSR) required. Claims Service Customer Assistance captures incident details, validates policy data, and creates claim and case records automatically across six business lines, freeing customer service teams from manual intake and reducing operational costs while improving the policyholder experience.

• Voice Based Visit Logging: Reclaim selling time lost to manual HCP visit documentation
  Simplify field reporting and keep teams in sync with a voice-activated, hands-free reporting experience to quickly log key insights, objections, and questions from healthcare provider (HCP) visits. Pharmaceutical and medical technology field reps reclaim time previously lost to data entry, while managers gain faster access to field activity insights.

The Winter '27 Release is generally available October 12, 2026. Visit the release page to learn more.

Explore more:

• Discover what else is new on the release page
• Take the Winter '27 Release module on Trailhead
• Read the Winter '27 Release Notes