Loading…
Salesforce Winter ’27: The Complete Guide to Agentforce, Platform Updates, Retirements, and More
TL;DR:
- What Changed: Salesforce Winter ’27 pushes AI deeper into sales, service, marketing, analytics, and development while expanding Data 360, Flow automation, Revenue Management, and developer tooling. At the same time, Salesforce is enforcing several security changes and putting legacy technologies on a firm retirement path.
- Why This Matters Now: The release moves Salesforce closer to a system where AI doesn’t just surface information; it executes work. But that only works when your data, permissions, integrations, and automation are ready for it.
- What This Means for You: Start with the foundations: audit authentication and integrations, clean up data and automation, review security permissions, and identify the workflows where Agentforce can safely take action.
Introduction
If Summer ’26 was about proving AI could be trusted, Salesforce Winter ‘27 is about putting it to work.
This release doesn’t reintroduce Agentforce. It operationalizes it.
Across sales, service, marketing, revenue management, analytics, and developer workflows, Salesforce is systematically closing the gap between “AI-assisted” and “AI-executed.”
And the shift isn’t limited to Agentforce.
Winter ’27 also brings deeper Data 360 access from Apex, more capable Flow automation, new developer tooling, expanded Marketing Cloud capabilities, stronger Experience Cloud controls, and new options for Revenue Management.
Then there’s the other side of the release: security enforcement and retirement deadlines.
Some changes are already being enforced. Others land in early, mid, or late 2027. Either way, the message is clear: Winter ’27 isn’t just a feature release. It’s an architectural checkpoint.
Here’s what matters, what it changes, and where your team should act.
1. Agentforce Is Becoming an Execution Layer
-
Agentforce Voice Gets Tighter Controls
Salesforce is enforcing limits on procured phone numbers, with production organizations limited by their license allocation, while sandbox and trial organizations are capped at 10 numbers. This helps prevent unintended overages and keeps deployments aligned with licensing.
The release also introduces flexible call recording control through the Salesforce Voice Toolkit API. Developers and admins can now pause and resume recordings without requiring the Control Call Recording user permission, making it possible to build custom recording controls with more granular permissions.
That reduces unnecessary access and strengthens security and compliance. One important distinction: the Control Call Recording permission is still required when managing recordings through the Salesforce UI.
Why This Matters: Organizations get tighter control over voice usage, licensing, and recording access.
What Changes for Your Team: Developers can build custom recording controls with fewer permissions, while UI-based recording management remains governed by the existing permission.
What Changes for End Users: Better usage and recording controls support more secure and compliant voice experiences.
Grazitti’s Take: Audit your Voice setup, phone-number usage, recording workflows, and permissions. Use API-based controls where custom solutions need more granular access.
-
New and Changed Standard Agent Actions and Subagents
Salesforce is also expanding its library of standard agent actions and subagents, letting admins add complex functionality to agents in minutes, not weeks.
The newly introduced Approval Agent comes with subagents for Approval Management, Search Approval Records, and Summarize Multiple Approval Work Items. Pre-built actions also cover common business processes out-of-the-box, reducing development effort and speeding up time-to-value for Agentforce implementations.
Why This Matters: Teams can move from agent experimentation to implementation faster, without building every capability from scratch.
What Changes for Your Team: Admins can combine pre-built actions and subagents to add complex functionality faster, reducing the development effort behind common Agentforce use cases.
What Changes for End Users: Faster agent implementation can translate into quicker automation of customer-facing processes, such as approvals and other routine requests.
Grazitti’s Take: Start with high-volume, repeatable processes where pre-built actions can deliver immediate value. Approval workflows are a strong starting point because the new Approval Agent already provides purpose-built capabilities.
-
Agentforce Observability
Agentforce Observability gives teams greater visibility into agent performance through session-level tracing and custom scorers, making it easier to evaluate how agents behave against defined standards.
Why This Matters: Once AI starts taking action, “it seems to work” isn’t enough. Teams need visibility into what an agent did, how it behaved, and whether it met expectations.
What Changes for Your Team: Admins and developers can define their own scoring criteria and trace individual agent sessions, making it easier to identify issues and measure performance.
What Changes for End Users: Better visibility into agent behavior can help organizations deliver more consistent and reliable AI-powered experiences.
Grazitti’s Take: Define your success criteria before scaling agents. Build observability into your Agentforce architecture from the start, not after something goes wrong.
2. Agentforce Sales Gets Smarter About Deals
-
Agentforce Sales Comes Directly to Slack
Agentforce Sales is now available directly in Slackbot, bringing AI agents into existing sales workflows. It can automate routine qualification and pipeline tasks, while a new Go page makes admin setup fast and straightforward.
Why It Matters: Reps no longer need to constantly switch between Slack and Salesforce for routine sales tasks. By taking care of repetitive qualification and pipeline work, Agentforce can give reps more time for high-value customer interactions.
What Changes for Your Team: Sales reps can access AI-powered sales assistance within Slack, while admins get a simpler setup experience through the new Go page.
What Changes for End Users: Less time spent on administrative work means reps can spend more time engaging with prospects and customers, potentially making interactions faster and more focused.
Grazitti’s Take: Start with high-frequency tasks such as pipeline updates, qualification, activity logging, and deal-status checks. Prove adoption and value first, then expand Agentforce Sales across more workflows.
-
Call Coaching
Salesforce introduces Call Coaching, allowing teams to rate reps against up to eight custom competencies from the Coaching tab on call records.
With Agentforce for Sales, an AI coach can automatically evaluate call recordings against those competencies. Without Agentforce for Sales, managers can still use a manual three-point scale.
Why It Matters: Manual call evaluation doesn’t scale. AI-assisted coaching creates a more consistent feedback loop, while the manual option ensures teams can still use the capability without Agentforce.
What Changes for Your Team: Managers can define the sales behaviors that matter most and evaluate reps against them. Agentforce automates the evaluation when enabled, reducing manual effort.
What Changes for End Users: More consistent coaching can help reps improve the quality and consistency of customer conversations.
Grazitti’s Take: Define a focused set of sales behaviors before rollout. Don’t turn eight competencies into eight vague scoring categories.
-
Pipeline Inspection Gets Cross-Cloud Risk Signals
Salesforce is enhancing Opportunity Scoring with Data 360 AI models that combine real-time conversation and engagement signals with standard CRM fields. AI can also automatically flag escalated support cases that pose a risk to active deals or customer relationships, bringing cross-cloud risk detection directly into Pipeline Inspection.
Why It Matters: A deal can be at risk because of something happening outside Sales. By connecting Service and Sales data, Salesforce can surface those risks before they directly impact the pipeline.
What Changes for Your Team: Sales teams get alerts about escalated support cases that could threaten active deals, giving reps an opportunity to intervene proactively instead of discovering the problem after pipeline impact.
What Changes for End Users: Service issues can be addressed before they damage an active relationship or derail a purchase, creating a more connected customer experience.
Grazitti’s Take: Make sure your Sales and Service data is connected and reliable before relying on cross-cloud risk signals. The AI can only protect deals if it can see the signals that put them at risk.
3. Agentforce Service Is Moving From Guidance to Execution
-
Case Management
Winter ’27 brings AI capabilities directly into the case workspace. Agents can translate Case Subject and Description fields on the fly and use AI-powered tone and clarity suggestions to refine case comments, all without leaving the case record.
Why It Matters: Agents can handle language barriers and improve the quality of their responses without relying on external tools or switching contexts.
What Changes for Your Team: Reps can translate case details and refine comments directly within the workspace, helping them resolve cases faster and with less manual effort.
What Changes for End Users: Clearer, more polished responses and faster resolution can create a smoother support experience, particularly for multilingual customers.
Grazitti’s Take: Enable these capabilities where you can remove repetitive work, but keep human review in the loop for customer-facing responses where tone and accuracy matter.
-
Knowledge Becomes the Foundation for AI-Powered Service
New Knowledge features enhance article creation, management, and delivery to both agents and customers, helping service teams work with accurate, up-to-date information and deliver more consistent, high-quality resolutions.
That’s more important than it sounds.
AI doesn’t magically create trustworthy service knowledge. It needs a trustworthy source to ground its responses in.
In Winter ’27, Knowledge increasingly serves as that foundation for AI-powered service replies, giving agents and AI a more reliable source of information to work from.
Why It Matters: Better knowledge means more consistent answers, fewer outdated responses, and stronger AI accuracy.
What Changes for Your Team: Service teams can create, manage, and deliver knowledge more effectively while giving AI a reliable source for customer-facing responses.
What Changes for End Users: Customers get more accurate, consistent, and relevant answers across service interactions.
Grazitti’s Take: Treat Knowledge as AI infrastructure, not just documentation. Audit outdated articles, close knowledge gaps, and establish clear ownership before expanding AI-powered service capabilities.
-
AI Solutions for Service Expands Language Support
AI Solutions for Service expands AI-generated summaries across more languages, making them more accessible for global service teams. Work Summaries now also extend to Voice calls and Enhanced Messaging sessions, while Service Replies for Email can be anchored in Enterprise Knowledge sources for more accurate, reliable responses.
Why This Matters: Broader language and channel support helps service teams use AI consistently across more customer interactions, while Knowledge grounding helps keep AI-generated responses factual.
What Changes for Your Team: Agents can work with AI-generated summaries across more languages and channels, while email replies can draw directly from trusted Enterprise Knowledge sources.
What Changes for End Users: Customers get more consistent, context-aware, and accurate support across Voice, Messaging, and Email.
Grazitti’s Take: Treat multilingual AI as a data and governance initiative, not simply a translation feature. Make sure your Knowledge sources are accurate, and your AI workflows are ready for the languages and channels you support.
-
Agentforce IT Service
Agentforce IT Service delivers a conversation-first, AI-driven approach to resolving IT tickets faster.
Instead of simply helping employees find information, it can deflect routine ticket volume, allowing IT teams to focus on more complex and strategic work. It also gives employees a more unified, proactive service experience across their IT interactions.
The Strategic Shift: Service automation is moving inward too. Agentforce isn’t only for customer-facing service teams; it’s becoming an operational layer for employee and IT support.
Why This Matters: Routine IT requests can consume significant team bandwidth. Deflecting those tickets gives IT teams more capacity for work that requires deeper expertise and judgment.
What Changes for Your Team: Employees get a conversational way to resolve routine IT needs, while IT teams can spend less time handling repetitive tickets and more time on strategic work.
What Changes for End Users: Although this is an employee-facing capability, faster internal IT support can reduce operational friction and help teams stay productive when serving customers.
Grazitti’s Take: Start with high-volume, repeatable IT requests that can be safely automated. Use those as the foundation before expanding Agentforce into more complex IT workflows.
4. Salesforce Flow Is Becoming a Much More Serious Automation Platform
-
Flow Tests Get Real Coverage
Salesforce can now help generate test scenarios for Flows, covering successful paths, edge cases, errors, required fields, and assertions.
But the more interesting addition in this release is real coverage visibility. Teams can build a test scenario per branch from a real record or an isolated data silo, then see actual coverage.
Why It Matters: Flow has become too important to rely on “we tested it once.” As automation becomes more complex, coverage becomes an operational requirement.
What Changes for Your Team: Instead of testing a few happy paths, admins and developers can validate individual branches and see which parts of a Flow are actually covered.
Grazitti’s Take: Build test scenarios around every meaningful branch, using real records where appropriate and isolated data silos when you need controlled testing. Don’t just prove that a Flow runs; prove that its logic is covered.
-
Bulk Screen Flows
Bulk Screen Flows allow teams to launch a screen flow across multiple records directly from a list view, without opening records one by one.
This is a deceptively practical improvement, and one that has clearly been wanted: a 650-vote Idea has finally made it to GA.
Instead of working through records individually, users can perform guided actions across an entire group of records.
Why It Matters: Repetitive, record-by-record workflows waste time. Bulk Screen Flows turn those tasks into a single guided action.
What Changes for Your Team: Users can select multiple records from a list view and launch the same screen flow across them, reducing repetitive admin work.
Grazitti’s Take: Look for high-volume tasks where teams repeatedly open records to perform the same guided action. Those are your quickest wins for Bulk Screen Flows.
-
Flow Builder Gets a UI Redesign
Flow Builder is getting a redesigned interface alongside new split elements, including Field Value and Date.
The update also improves how teams launch screen flows against multiple selected records from list views and related lists.
-
Flow Builder Does More
Salesforce is pushing more Agentforce capabilities directly into Flow Builder.
Admins can build and configure Agentforce agents inside Flow Builder, set custom batch sizes to avoid governor limits in scheduled flows, and edit screen flows using natural language.
The takeaway: The boundary between automation and AI is getting thinner.
Flow handles the process. Agentforce handles the reasoning and action.
And Salesforce is increasingly putting both in the same workspace, making it easier for admins to build automation that combines structured workflows with AI-driven decisions.
Why This Matters: Flow Builder is becoming more flexible for both designing automation and executing it at scale.
What Changes for Your Team: Admins get a cleaner interface, more ways to branch Flow logic, a dedicated Time screen component, and the ability to run screen flows across multiple selected records.
What Changes for Customers: More efficient Flow-driven processes can reduce repetitive steps and make guided experiences faster to complete.
Grazitti’s Take: Don’t treat the redesign as purely cosmetic. The new split elements, Time component, and bulk screen-flow capabilities expand what admins can build, and how efficiently users can execute it.
5. Data 360 Is Becoming More Usable for Developers and Operators
-
Data 360 SQL From Apex
Developers can now execute Data 360 SQL queries directly from Apex, bringing unified Data 360 data and custom logic together in the same transaction.
That means developers can work with Data 360 data without creating an unnecessary workaround layer, making it easier to incorporate unified data into custom Salesforce applications and AI-backed workflows.
Why This Matters: Data 360 becomes much easier to incorporate into custom Salesforce applications and AI-backed workflows.
What Changes for Your Team: Developers can combine Data 360 queries with Apex logic in a single transaction, reducing the need for additional integration or workaround layers.
What Changes for End Users: Custom applications can use unified data more directly, enabling more connected and context-aware experiences.
Grazitti’s Take: Look for custom workflows currently relying on workaround layers to access Data 360 data. Those are strong candidates for simplification with direct Apex access.
-
Governance Metadata Can Move With Your Deployment
Data 360 governance metadata can now move from sandbox to production, making environment promotions less dependent on manual setup.
Payoff: Teams spend less time reconfiguring governance metadata every time an environment is promoted.
For organizations treating Data 360 as production infrastructure rather than an analytics side project, this is a meaningful deployment improvement.
Why This Matters: Streamlined metadata deployment reduces friction when moving Data 360 configurations between environments.
What Changes for Your Team: Teams can promote governance metadata from sandbox to production with less manual reconfiguration, making deployments more consistent and efficient.
What Changes for End Users: More consistent Data 360 deployments can help teams deliver unified data experiences faster and with fewer environment-related delays.
Grazitti’s Take: Add governance metadata to your deployment checklist and standardize how it moves across environments. The less configuration that has to be rebuilt manually, the easier Data 360 becomes to operate at scale.
-
Problem Records Data Lake Object (DLOs)
Problem Records DLOs provide a structured place to identify records that failed ingestion, making data-quality issues easier to find and debug.
That matters because teams can catch bad records before they reach downstream segments, agents, and reports, rather than discovering data-quality problems after they have already affected business processes.
Why This Matters: Ingestion failures are inevitable at scale. Having a dedicated place to surface them makes data-quality management more proactive.
What Changes for Your Team: Data teams get a clearer way to identify and troubleshoot failed records before they affect downstream use cases.
What Changes for End Users: Cleaner data flows into segments, agents, and reports, leading to more accurate and consistent experiences.
Grazitti’s Take: Don’t measure Data 360 success only by how much data you ingest. Measure how much bad data you can identify and resolve.
-
Reports and Dashboards
Winter ’27 expands Reports and Dashboards with several practical improvements. Record previews now open details in a side panel while preserving your filters and position, so you can inspect records without losing your place.
Lightning report and dashboard components can now also work on LWR sites, while Rapid Publish activates segments to Data 360 targets faster.
Why This Matters: These updates reduce friction when reviewing records, embedding analytics, and activating segments across Salesforce experiences.
What Changes for Your Team: Users can preview records without navigating away from their report context, teams can bring Lightning report and dashboard components to LWR sites, and segments can reach Data 360 targets faster.
What Changes for End Users: Faster segment activation and more accessible reporting can help deliver more timely, data-driven experiences.
Grazitti’s Take: These may look like incremental improvements, but they remove small points of friction that add up at scale, especially for teams managing high-volume reporting and Data 360 activation workflows.
6. Revenue Management Gets More Enterprise-Ready
-
Advanced Approvals
Advanced Approvals now provides tighter control over approval delegation.
The important consideration: enabling it silently disables delegations configured on user records. It also requires Advanced Approvals plus Agentforce, and if those requirements aren’t met, the setup fails without an error message.
Payoff: More controlled approval routing, but with a critical governance dependency.
Why This Matters: A silent change to existing delegations can disrupt approval routing if teams don’t understand the impact before enabling the feature.
What Changes for Your Team: Admins need to verify existing user-level delegations and confirm the required Advanced Approvals and Agentforce capabilities are in place before switching it on.
What Changes for Customers: Better-controlled approval routing can reduce ambiguity, but misconfigured delegations could delay approvals or send them to the wrong place.
Grazitti’s Take: Audit existing delegations first. Don’t enable this blindly. Validate prerequisites, understand which user-record delegations will be disabled, and test approval routing before rolling it into production.
-
Large Transactions and Quote Processing
Revenue Management can now handle up to 15,000 line items per quote or order, helping eliminate timeouts and bottlenecks as quote complexity grows.
Asynchronous processing keeps reps and billing teams moving while calculations and document generation run in the background. It also supports nested bundles and multilevel grouping, making it better suited to enterprise-scale CPQ complexity.
Why This Matters: High-volume, complex quotes can strain CPQ performance. Higher line-item capacity and asynchronous processing help teams handle that complexity without slowing down users.
What Changes for Your Team: Reps and billing teams can keep working while calculations and document generation run in the background, and support for nested bundles and multilevel grouping accommodates more sophisticated product structures.
What Changes for End Users: Faster, more reliable quote and order processing can reduce delays in complex purchasing journeys.
Grazitti’s Take: If your CPQ environment is approaching performance limits, look beyond simply increasing capacity. Review line-item volume, bundle complexity, and synchronous processes to identify where asynchronous processing can remove bottlenecks.
-
Transaction Management
Revenue Management expands support for industry-specific procedure plans, backdated asset transactions, price amendments, and timezone accuracy across the asset lifecycle.
The release also makes transaction management more flexible. The Sales Transaction Line Editor gains enhanced filtering, auto-refresh, and configurable action buttons, while Dynamic Forms support enables more tailored transaction page layouts.
Why This Matters: Complex revenue and asset lifecycles often require more than standard transaction handling. These updates give teams greater control over how transactions are processed and presented.
What Changes for Your Team: Teams can manage backdated transactions and price amendments more effectively, while tailoring transaction pages and editor actions to their specific workflows.
What Changes for End Users: More accurate asset and pricing management can lead to smoother, more consistent transaction experiences.
Grazitti’s Take: Use these capabilities to align Revenue Management more closely with your actual business processes. Pay particular attention to backdated transactions, pricing changes, timezone handling, and industry-specific procedures where small inaccuracies can create downstream issues.
-
Agentforce for Revenue Management
The Approval Agent uses Agentforce AI to autonomously route and process revenue approval requests, bringing native AI-driven approvals directly into Revenue Cloud’s management process.
By automating approval routing and processing, it can eliminate manual bottlenecks in deal approval workflows and help accelerate deal closure.
Why This Matters: Approval delays can slow down otherwise ready-to-close deals. Automating the process helps move approvals forward without relying on every step being handled manually.
What Changes for Your Team: Revenue teams can reduce the manual effort involved in routing and processing approval requests, allowing teams to focus on higher-value deal work.
What Changes for End Users: Faster approval cycles can help customers move through purchasing decisions with fewer unnecessary delays.
Grazitti’s Take: Start with approval workflows that are high-volume, rules-driven, and prone to bottlenecks. Automate the routing first, then expand as confidence in AI-driven approvals grows.
7. Agentforce Marketing Is Moving Toward AI-Assisted Execution
-
Marketing Cloud Account Engagement
Marketing Cloud Account Engagement introduces new Agentforce capabilities, bringing autonomous AI directly into B2B campaign management. Enhancements span campaigns, audience management, and expanded channel support, extending AI-assisted marketing automation across the full Account Engagement workflow.
Why This Matters: Agentforce is moving beyond content assistance into the day-to-day execution of B2B marketing, helping teams manage campaigns and audiences with less manual effort.
What Changes for Your Team: Marketers can use AI across more stages of the Account Engagement workflow, from campaign management and audience operations to broader channel engagement.
What Changes for End Users: More connected and automated B2B engagement can help customers receive more timely and relevant interactions across channels.
Grazitti’s Take: Start with repetitive campaign and audience-management tasks where automation can deliver immediate value. Then expand Agentforce across the Account Engagement workflow as teams build confidence in autonomous marketing operations.
-
Marketing Intelligence
Marketing Intelligence expands data ingestion and adds historical backfill, pipeline monitoring, and data validation, helping teams build more reliable cross-channel ROI reporting.
It also provides AI-guided recommendations that surface actionable insights to improve campaign performance, while bringing marketing performance data across channels together with stronger controls for data accuracy and connection reliability.
Why This Matters: Reliable ROI reporting depends on reliable data. Better ingestion, validation, and monitoring help teams trust the numbers behind their marketing decisions.
What Changes for Your Team: Marketing teams get a more unified view of performance across channels, with AI-guided insights and stronger controls to identify data-quality or connection issues.
What Changes for End Users: Better-informed campaign decisions can lead to more relevant and effective marketing experiences across channels.
Grazitti’s Take: Don’t focus only on adding more data sources. Make data quality and connection reliability part of your Marketing Intelligence strategy. Then use AI-guided recommendations to turn that cleaner data into action.
-
Real-Time Offer Management
Real-Time Offer Management adds personalized targets, flexible promotion stacking rules, earlier coupon validation, and custom eligibility logic through tabular, rules-based promotions.
Agentforce can also turn a campaign brief into a fully drafted offer in minutes, accelerating offer creation at scale. Meanwhile, headless APIs let teams surface offers directly in their own UI, enabling seamless, contextually relevant customer experiences.
Why This Matters: Personalized offers only work when businesses can create, validate, and deliver them quickly. These updates make offer management more flexible while reducing the effort required to scale promotions.
What Changes for Your Team: Teams get more control over promotion eligibility, stacking, targeting, and coupon validation. Agentforce speeds up offer creation, while headless APIs give developers flexibility over how and where offers appear.
What Changes for End Users: Customers can receive more relevant offers surfaced directly within the experiences they already use, without unnecessary friction.
Grazitti’s Take: Treat offers as a real-time decision layer, not just static promotions. Use Agentforce to accelerate creation, rules-based logic to control eligibility, and headless APIs to deliver the right offer in the right customer experience.
8. Experience Cloud Gets More Control and One Important Migration Task
-
Aura and LWR Sites
Authenticated site visitors can now choose from multiple processes when submitting records for approval, giving them more flexibility within approval workflows.
Salesforce is also improving the Orchestration Work Guide, reducing time-on-task for users who need to complete multiple work items and making multi-step work easier to manage.
One important caveat: Salesforce discontinued the Experience Delivery beta feature in Winter ’27.
Why This Matters: Approval and orchestration workflows often create friction when users have limited options or have to work through multiple tasks manually. These updates make both experiences more flexible and efficient.
What Changes for Your Team: Teams can offer authenticated site visitors multiple approval processes, while users completing orchestrated work can move through multiple work items more efficiently.
What Changes for End Users: Customers and site users get a more flexible approval experience, with less friction when submitting records for review.
Grazitti’s Take: Map your approval and orchestration journeys end to end. Look for places where users are forced into a single process or spend too much time completing repetitive work items.
-
Security and Sharing
Admins can now specify the From address for Experience Cloud sites with unverified email domains and designate a secondary sender address, giving teams more control over site email configuration.
They can also hide sensitive fields from guest users while keeping them visible to authenticated portal members, creating a more precise approach to data access.
Why This Matters: Experience Cloud often serves different audiences with different privacy requirements. These controls give admins greater precision over who can see what and how site emails are sent, supporting GDPR, CCPA, and broader data privacy requirements.
What Changes for Your Team: Admins can configure sender addresses more flexibly and apply different field-visibility rules for guest and authenticated users.
What Changes for End Users: Authenticated portal members can access the information they need, while guest users are shielded from sensitive data—creating a more secure and tailored experience.
Grazitti’s Take: Review Experience Cloud permissions and email configuration with a least-privilege mindset. Don’t treat guest and authenticated users as the same audience; use the new controls to align visibility with actual access needs.
-
Stable Target Host Names
For organizations using Salesforce Edge Network with a third-party custom domain, Salesforce recommends updating the target host name to your My Domain login URL. This helps maintain stability through instance and Hyperforce cell changes.
The change also eliminates the need to coordinate domain updates during Salesforce maintenance, reducing downtime risk.
Why This Matters: Infrastructure migrations can create unexpected domain-management work and introduce downtime risks. Pointing the target host to the My Domain login URL makes the setup more resilient to Salesforce infrastructure changes.
What Changes for Your Team: Admins no longer need to coordinate target-host updates around every instance or Hyperforce cell change, simplifying ongoing domain management.
What Changes for End Users: A more stable domain configuration helps reduce service disruption during Salesforce maintenance and migrations.
Grazitti’s Take: Don’t wait for the next migration. If your org uses Salesforce Edge Network with a third-party custom domain, update the target host name to your My Domain login URL before the next org migration.
9. Headless 360 Gets Better Developer Tooling
-
Agentforce Vibes
Agentforce Vibes is an AI-powered developer tool available as both a VS Code extension and a standalone IDE, designed to accelerate agent creation and reduce the learning curve for developers.
Version 4.0+ is a full re-engineering built on the Salesforce Coding Agent Platform, with Claude (Anthropic’s AI) and Mastra (an open-source agent framework) serving as the underlying agent orchestration engines. It also brings AI-assisted agent development natively into VS Code, making Agentforce builds faster and more intuitive.
Why This Matters: Building AI agents can introduce a new learning curve for developers. Agentforce Vibes brings AI-assisted development into familiar developer environments, lowering that barrier.
What Changes for Your Team: Developers can create and configure Agentforce agents directly in VS Code, while the re-engineered platform provides the underlying orchestration capabilities.
What Changes for End Users: Faster agent development can help organizations bring AI-powered experiences and workflows to customers sooner.
Grazitti’s Take: Agentforce Vibes signals that Salesforce wants agent development to become part of the everyday developer workflow, not a separate skill set. Teams already building in VS Code should evaluate how much of their Agentforce development can move into this AI-assisted experience.
-
Salesforce CLI
Salesforce CLI now supports authoring Agentforce agents, creating development and test environments, and synchronizing source code directly from the command line.
Developers can also run tests and control the full application lifecycle programmatically, reducing context switching between development tools. Together, these capabilities enable end-to-end automation of the Salesforce development lifecycle through scripting.
Why This Matters: Agent development is becoming part of the broader software development lifecycle, not a separate workflow. CLI-based development makes that lifecycle easier to automate and standardize.
What Changes for Your Team: Developers can author agents, create environments, sync source, run tests, and manage application lifecycles from the command line, reducing manual steps across tools.
What Changes for End Users: More automated and consistent development workflows can help teams deliver and update Agentforce-powered experiences faster.
Grazitti’s Take: Treat Agentforce development like software engineering. Bring agents into your existing CI/CD and scripting workflows instead of managing them as a separate development process.
-
Salesforce Development Claude Code Plugin
The Salesforce Development Claude Code plugin brings AI-assisted development directly into the Headless 360 development workflow, helping developers build more sophisticated apps and agents faster using Claude’s AI capabilities.
The plugin also integrates natively with the existing Headless 360 toolchain, creating a more seamless development experience without requiring developers to step outside their established workflow.
Why This Matters: AI-assisted development can accelerate complex app and agent builds, especially as Salesforce development becomes increasingly agent-driven.
What Changes for Your Team: Developers can use Claude’s AI capabilities directly within the Headless 360 toolchain to build and iterate on apps and agents faster.
What Changes for End Users: Faster development cycles can help organizations deliver more sophisticated digital experiences and AI-powered capabilities sooner.
Grazitti’s Take: The real value is AI embedded into the development workflow, not AI as another standalone tool. Teams using Headless 360 should look at where Claude can reduce development effort without adding another disconnected layer to the toolchain.
10. Salesforce Is Also Tightening the Platform’s Foundations
-
Claudeforce Brings Claude Into Salesforce Workflows
Claudeforce embeds Claude within Salesforce and surfaces Salesforce capabilities within Claude, with 37 prebuilt skills that enable agentic workflows beyond traditional tab-based interfaces.
Why This Matters: Salesforce AI is no longer confined to a single Salesforce experience. By bringing Claude into Salesforce workflows and Salesforce capabilities into Claude, teams get a more flexible way to work across CRM data, applications, and AI.
The bigger shift is toward AI that works across the tools people already use, rather than forcing users to constantly switch between interfaces.
What Changes for Your Team: Teams can use AI capabilities within their existing workflows, while developers and admins get new ways to build agentic experiences without relying entirely on traditional Salesforce interfaces.
What Changes for End Users: More connected AI workflows can help teams work with CRM data and applications more efficiently, potentially reducing friction in customer-facing processes.
Grazitti’s Take: Identify workflows where users currently jump between Salesforce and external AI tools. Those are the strongest candidates for evaluating Claudeforce and seeing where embedded AI can eliminate unnecessary context switching.
-
Apex
Apex Guru now identifies duplicate and near-duplicate Apex code, helping developers reduce redundancy and keep their codebase cleaner. It’s also available directly from agentic IDEs such as Cursor, VS Code, and Agentforce Vibes.
Why This Matters: Duplicate code adds maintenance overhead and can make large Apex codebases harder to manage. Identifying redundancy earlier helps teams keep their code more efficient and maintainable.
What Changes for Your Team: Developers can spot duplicate or near-duplicate code while working in the IDEs they already use, rather than relying solely on manual code reviews.
Grazitti’s Take: Use Apex Guru as part of your code-quality workflow, especially when working across large or rapidly evolving Apex codebases.
-
Setup Audit Trail Gets Its Own Permission
A dedicated View Setup Audit Trail permission replaces broader View Setup access. Existing access is retained, so current users won’t lose access because of the change.
The update gives organizations more granular control over who can inspect administrative changes without relying on broader Setup permissions.
Why This Matters: Audit Trail access can now be managed more precisely, strengthening administrative visibility and access control.
What Changes for Your Team: Existing access remains intact, but admins can take a more targeted approach to determining who should be able to view Setup Audit Trail information.
What Changes for End Users: Tighter administrative access controls can help organizations better protect sensitive configuration and change information.
Grazitti’s Take: Review who currently has broad Setup access and whether they still need it. Use the new dedicated permission to align Audit Trail access with actual responsibilities.
11. Security Changes You Shouldn’t Push to the Backlog
-
Profile Filtering Is Now Enforced
Users without the View All Profiles permission can no longer see other profile names.
Review automation, custom UI, and processes that depend on profile visibility before the release reaches your instance.
-
Instanced URLs in API Traffic
Salesforce is enforcing the move away from instance-specific URLs in API traffic.
Any integration still calling na__.salesforce.com hostname needs to be reviewed and moved to My Domain URLs.
This is exactly the kind of change that can quietly break an integration.
-
Authorized Email Domains
Organizations need to configure DKIM or an authorized email domain before bulk email updates where Email Change Verification is involved.
-
Modify Transaction Security Policy
A dedicated permission is now required to create, edit, or delete transaction security policies.
Review your current policy owners and update permissions accordingly.
Features Retiring in Spring ’27 & Beyond
| Feature | Deadline | What to do |
|---|---|---|
| Salesforce Certifications | Feb. 1, 2027 | Map affected credentials to active certification paths |
| OAuth Username-Password Flow | Feb. 20, 2027 | Audit connected apps and migrate authentication |
| Lightning Sync | April 2027 | Move to Einstein Activity Capture |
| SOAP login() API 31.0–64.0 | Summer ’27 | Update legacy authentication and API calls |
| Standard-Volume Platform Events | June 2027 | Move to high-volume platform events |
| Salesforce for Outlook | Dec. 2027 | Transition to Outlook Integration with Einstein Activity Capture |
And Don’t Forget Connected Apps
Salesforce is also ending support for Connected Apps in Summer ’27.
Existing apps will continue to function, but Salesforce will no longer provide bug fixes or support for their authorization flows. The recommended migration path is External Client Apps.
Don’t wait for the deadline.
Inventory every Connected App, identify the business process behind it, and decide what the replacement architecture should be.
The Bottom Line: Winter ’27 Is a Readiness Test
Winter ’27 has a clear thesis:
AI in Salesforce is no longer an experiment. It’s becoming an operational layer.
The Approval Agent, prospect research, executable service actions, AI-powered marketing workflows, conversational experiences, agentic analytics, and AI-assisted development all point in the same direction.
But there’s a catch.
You can’t automate your way out of bad foundations.
Agents need trustworthy data. They need reliable integrations. They need well-designed automation. They need the right permissions.
And once they start making decisions or taking action, they need observability and governance.
That’s why the less flashy parts of Winter ’27 matter just as much.
Data 360 gets better tooling. Flow gets better testing and orchestration. Developers get stronger agent-development workflows. Security becomes more granular. Legacy authentication and integration patterns get deadlines.
Salesforce is effectively doing two things at once:
Giving organizations more power and removing the excuses for leaving old architecture untouched.
The organizations that get the most from Winter ’27 won’t simply activate the most features.
They’ll build the foundation that lets those features operate safely at scale.
Winter ’27 is both an opportunity and a deadline.
Salesforce Winter ’27 Implementation Checklist
For Admins
Before go-live
- Confirm your actual upgrade date on Trust.
- Go to Setup → Release Updates by enforcement date.
- Audit authentication, Connected Apps, and legacy integrations.
- Configure DKIM or an authorized email domain.
- Review transaction security policy ownership.
- Check profile visibility dependencies.
- Review Experience Cloud domains and target host names.
Worth doing next
- Pilot Bulk Screen Flows.
- Review Flow test coverage.
- Clean up unused Flow resources.
- Evaluate Agentforce use cases with clear business guardrails.
- Review Data 360 ingestion failures and governance metadata.
For Developers
Before go-live
- Audit integrations using grant_type=password.
- Search Apex, Flow, and validation logic for profile visibility dependencies.
- Replace instance-specific Salesforce hostnames with My Domain URLs.
- Review SOAP API authentication dependencies.
- Check integrations using Salesforce-to-Salesforce or Connected Apps.
Worth doing next
- Explore Data 360 SQL from Apex.
- Test SOQL FORMULA() in a scratch org.
- Evaluate Agentforce Vibes and Salesforce CLI for agent development.
- Add Agentforce observability and testing to new agent projects.
Still figuring out what the Winter ‘27 release means for your org? Drop a line at [email protected], and our Salesforce experts can help you assess your readiness, prioritize migrations, and build an activation roadmap for the features that matter most to your business.
Frequently Asked Questions
1. What is new in Salesforce Winter ’27?
Salesforce Winter ’27 expands Agentforce across sales, service, marketing, revenue operations, analytics, and development. Major additions include Agentforce Sales in Slack, Voice controls, new standard agent actions, Agentforce Observability, Data 360 improvements, Bulk Screen Flows, expanded Revenue Management, and new developer tooling.
2. How does Salesforce Winter ’27 change Agentforce?
3. What Salesforce features need immediate attention after Winter ’27?
4. What Salesforce features are retiring in 2027?
5. How should businesses prepare for Salesforce Winter ’27?
6. What does Salesforce Winter ’27 mean for the future of AI?
What do you think?
0Like
1Love
0Wow
0Insightful
0Good Stuff
0Curious
0Dislike
0Boring
RELATED LINKS