Salesforce Is Killing These 5 Features in 2026 - Is Your Team Ready?
![]() |
There was no announcement email. No countdown banner in your org. Just one morning in February, a team in Sydney opened Salesforce, clicked on their Live Agent queue — and nothing happened. Access was gone. Salesforce had retired it, right on schedule, and the team had missed the memo.
If that story makes you slightly uncomfortable, it should. Because there are four more changes either already live or coming before the end of 2026 — and most Salesforce orgs aren't ready for a single one of them.
Why Salesforce Retirements Catch Teams Off Guard
Salesforce releases three major updates every year — Spring, Summer, and Winter. Buried inside every set of release notes are deprecation notices: features being sunset, behaviours being changed, integrations being retired. Most admins scan the headlines. Very few read the retirement section.
That's not laziness. It's capacity. A Salesforce admin managing a mid-sized org is already handling requests, running reports, maintaining automations, and supporting a sales team that treats "Salesforce is broken" as a complete IT ticket. Deprecation notices don't make it to the top of the pile until something stops working.
This blog exists to move them to the top of your pile — before something stops working.
Here are the five Salesforce features being retired in 2026, what each one means for your org, and what your team needs to do before the deadline arrives.
Feature #1: Salesforce Live Agent — Already Gone
Retirement date: February 14, 2026 Status: Retired. Access removed.
This one is done. If your org was still using Salesforce's original Live Agent — the embedded real-time chat solution built into Service Cloud — access was removed on Valentine's Day 2026. No grace period. No extension. Gone.
Live Agent was Salesforce's first native chat product. For years, it powered the embedded chat windows on thousands of support portals and service sites. It worked. It was familiar. And for many teams, it was simply never replaced because there was no urgent reason to replace something that wasn't broken.
Until it was retired.
The migration path Salesforce has been pushing is Messaging for In-App and Web — the modern, asynchronous messaging framework that supports WhatsApp, SMS, and web chat within a single unified interface. For teams wanting AI-powered support, Einstein Bots combined with Agentforce is the recommended destination.
What to do if you're affected: If your service portal still shows a broken chat widget or your team is fielding calls that used to come through Live Agent, this is an urgent fix, not a roadmap item. The channel is down. The migration needs to happen now, not in the next planning cycle.
What this means for your wider platform: Moving to Messaging for In-App and Web isn't just a like-for-like replacement. It's an opportunity to rebuild your support channel properly — with AI triage, intelligent routing, and omnichannel visibility that Live Agent never had. Teams that treat this retirement as a forced upgrade tend to end up with a significantly better support operation within 60 days.
Feature #2: Connected Apps — Creation Disabled, Migration Required
Restriction date: Spring '26 (March 2026) Status: New Connected Apps can no longer be created by default.
This one caused more confusion than any other Salesforce change in the first quarter of 2026. Developers across the ecosystem opened their orgs in March and discovered the button to create a new Connected App simply wasn't there anymore. Support cases flooded in. Most of them were false alarms — but the underlying change is real and the implications are significant.
Here is exactly what happened: Salesforce blocked the creation of new Connected Apps by default across all orgs, including sandboxes, as part of Spring '26. Existing Connected Apps — the ones already running your integrations, your Dataloader configurations, your Postman collections — continue to work without interruption. Nothing broke. But any new integration from this point forward must use External Client Apps (ECAs) instead.
ECAs are Salesforce's modern OAuth-based integration framework. They offer better security isolation, cleaner governance, and proper metadata compliance. In most technical respects, they are a genuine improvement over Connected Apps. The problem is that most teams don't know they exist, and every developer who tries to create a Connected App the way they've always done it is now hitting a wall.
What to do now: Audit your current integration stack. Identify every Connected App in your org and understand which ones are still active versus legacy. For any new integrations being scoped or built, ECA must be your starting point — not an eventual migration target. If your team has an ISV or managed package still shipping Connected Apps, that architecture is now officially unsupported for new builds.
Re-enabling Connected App creation is currently possible by raising a Salesforce Support ticket, but Salesforce has signalled this exception process will be removed in future releases. Do not build a new integration strategy around an exception process that is explicitly being phased out.
Feature #3: Salesforce Functions — Officially Retired
Retirement date: January 2026 Status: Removed from AppExchange. Retired.
Salesforce Functions was the platform's attempt at serverless compute — a way for developers to run compute-intensive logic outside the Salesforce governor limits by invoking functions hosted on Salesforce's own infrastructure. It launched to genuine excitement in 2022. By January 2026, it was gone.
The retirement isn't catastrophic for most orgs. Salesforce Functions never achieved widespread adoption — it was predominantly used by development teams exploring the edges of the platform rather than by business users or admins. But for the teams that built real production logic on Functions, the retirement created an urgent gap that needed filling before their order term expired.
The most natural replacement paths are AWS Lambda and Google Cloud Functions for teams already in cloud infrastructure, or Heroku for teams wanting to stay close to the Salesforce ecosystem. However — and this matters — Heroku itself is now a complicated recommendation.
What to do if your team used Functions: Identify every Functions deployment in your org and classify it by business criticality. Map the replacement architecture — Lambda, Azure Functions, or native Apex refactoring where the logic is small enough. Do not assume Heroku is a safe long-term home for this logic without reading Feature #5 below first.
Feature #4: Document Generation 1.0 — Sunset Coming July 2026
Sunset date: July 31, 2026 Status: Retiring mid-year. Migration required before deadline.
This is the retirement most likely to catch teams completely off guard — because Document Generation 1.0 is embedded quietly inside processes that nobody thinks about until they stop working.
Document Generation 1.0 is the tool many orgs use to produce templated documents directly from Salesforce records: proposal PDFs, contracts, service agreements, client reports, renewal letters. It plugs into flows, process builders, and button-triggered actions. It does its job invisibly. And for that exact reason, the teams using it often have no idea how deeply embedded it is until they start mapping dependencies.
Salesforce is sunsetting Document Generation 1.0 on July 31, 2026. The replacement is Document Generation 2.0 — a rebuilt version with better template management, improved formatting control, and compatibility with the modern Salesforce platform.
The migration itself is not technically complex, but the discovery work before the migration often is. Most orgs cannot tell you off the top of their head every place Document Generation 1.0 is being used. The answer is almost always more than expected: a renewal flow here, a proposal button there, an automated email attachment buried inside a legacy process.
What to do before July 31: Run a full dependency audit on Document Generation 1.0 across your org. Identify every flow, process, and trigger that calls it. Map those to the Document Generation 2.0 equivalent and rebuild in order of business criticality — highest-volume, client-facing documents first. Do not leave this until June. Teams that start in June will be cutting it close. Teams that start in Q4 2025 or Q1 2026 are in the strongest position.
If your team doesn't have a Salesforce admin with capacity to run this audit, this is exactly the kind of engagement where a certified Salesforce partner can complete the discovery in a single focused sprint and give you a clear migration roadmap before the pressure builds.
Feature #5: Heroku — No New Features. Future Uncertain.
Announcement date: February 2026 Status: Sustaining engineering only. No new feature development.
This one is different from the others. Heroku hasn't been retired. There is no shutdown date. Existing customers can continue to use it, renew their subscriptions, and run their applications without any immediate disruption. But what Salesforce announced in February 2026 is significant enough to warrant serious attention in any organisation that has built architecture decisions around Heroku's future.
Salesforce SVP and GM Nitin T. Bhat confirmed in February 2026 that Heroku is moving to "sustaining engineering" mode. This means the platform will receive stability updates, security patches, and support — but no new features. No new integrations. No investment in the platform's capabilities. Heroku is, in the language of enterprise software, being maintained rather than developed.
For teams that chose Heroku because of its deep connection to the Salesforce ecosystem — native data connectivity, shared identity, co-marketing positioning — the signal is clear: Salesforce is not Heroku's future. The platform will be supported until it isn't, and the timeline for "until it isn't" has not been communicated.
What to do now: If Heroku is a peripheral part of your architecture — a small microservice, a prototype environment — the calculus is simple: maintain what you have, but build anything new elsewhere. If Heroku is a core part of your Salesforce integration architecture — housing critical business logic, managing significant data flows, or running production applications your business depends on — this announcement is a signal to begin an architectural review.
The practical alternatives depend on your workload. AWS is the most common migration destination for teams moving Heroku applications to a longer-term platform. Azure and Google Cloud Platform are also strong options depending on your existing cloud relationships. For teams that used Heroku specifically for its Salesforce connectivity, native Salesforce Platform solutions — Apex, Flow, and increasingly Agentforce — are worth re-evaluating as the platform itself has matured significantly.
The Common Thread — And What It Means for Your Org
Look across these five retirements and a clear pattern emerges. Salesforce is consolidating. The platform is moving away from legacy tools (Live Agent, Connected Apps, Functions, Heroku) and towards a smaller set of modern, AI-native capabilities — Agentforce, ECAs, Messaging for In-App and Web, and the Salesforce Platform itself as the primary compute environment.
This is not a criticism of Salesforce. Every mature platform goes through this cycle. Legacy features accumulate technical debt, slow down the core product, and eventually need to be retired in favour of better alternatives. What Salesforce is doing in 2026 is a reasonable modernisation programme.
The challenge — and the risk — is for the orgs that have not kept pace. Salesforce's installed base includes organisations running architecture decisions made in 2016, 2018, and 2020, on features that are now being retired or deprioritised. The teams that understood this change was coming are already migrated or in the process of migrating. The teams that didn't are discovering it at the worst possible moment: when something stops working.
The goal of this blog is to make sure your team is in the first group, not the second.
Quick Reference: 2026 Salesforce Retirement Summary
Is Your Org Affected? Here's How to Find Out.
The honest answer for most Salesforce orgs is: probably, at least partially. The features being retired in 2026 are not niche or experimental. Live Agent was a mainstream Service Cloud feature. Connected Apps are in virtually every integrated org. Document Generation is used across sales and service operations. Heroku is part of the architecture of thousands of organisations that built on the Salesforce ecosystem over the past decade.
The question is not whether your org needs to respond to these changes — it's how urgently and in what order.
That starts with a clear-eyed audit of your current Salesforce configuration: what features are in use, where dependencies exist, and which retirements create immediate risk versus manageable future planning. Most orgs do not have a current, accurate picture of their own Salesforce architecture at this level of detail. That's not a failure — it's an honest reflection of how platforms evolve over years of incremental change.
At Amroar, we run free 30-minute Salesforce audits for exactly this kind of situation. In a single session, we can walk through your org, identify which of these retirement changes affect you, and give you a clear, prioritised action list — no sales pitch, no obligation, just a specific answer to the question "what does our team actually need to do?"
If any of these five retirements have landed on your desk, or if you're not sure whether they should have, that conversation is worth having before July arrives.
Book Your Free 30-Minute Salesforce Audit →
Amroar Technologies is a certified Salesforce Consulting Partner with 12+ Salesforce certifications and 200+ enterprise implementations across the US, UK, and Australia. We specialise in Salesforce implementation, migration, and AI integration — helping B2B teams get more from their CRM without the implementation horror stories.
Questions about this post? Reach us at amroar.com/contact


Comments
Post a Comment