Zendesk is sunsetting its legacy AI agent functionality, including AI agents Essential and the AI agent framework (bot builder, answers, and intents). It is also retiring default messaging responses, which are not AI-powered but have become largely redundant due to changes in the platform’s AI agent capabilities.
The goal is to give customers access to more advanced AI functionality, but the timeline is relatively tight. Zendesk announced the shutdown on March 30, 2026, with technical development ending on August 31 and complete removal scheduled for December 2026. This leaves customers with a limited window to migrate to Zendesk's new AI agents experience.
While all Zendesk customers using legacy AI agent functionality are affected, enterprises face the most complex migration path. Years of accumulated bot logic, intents, integrations, and multi-channel deployments can make migration a substantial effort rather than a regular upgrade.
This guide breaks down the Zendesk legacy AI functionalities being shut down, the features that replace them, and how enterprise teams can prepare for a smooth transition.
What Zendesk Is Shutting Down (and Exactly When)
Before planning your migration, it's important to understand exactly which Zendesk features are being discontinued. Knowing what needs to be replaced is the first step toward estimating the required effort.
The three features being removed
Zendesk is retiring AI agents Essential, legacy AI agent functionality (bot builder, answers, and intents), and default messaging responses. Each feature serves a different purpose within Zendesk's legacy customer support automation framework:
- AI agents Essential — self-service generative bots that handle customer queries by searching the help center, identifying relevant content, and generating responses.
- Legacy bot builder, answers, and intents — the framework for building and managing automated support conversations. Bot builder is used to create workflows, intents identify customer requests, and answers define how the system responds to them.
- Default messaging responses — the baseline automated responses displayed when customers interact with a messaging channel. Common use cases include welcome messages, business-hours notifications, and basic conversation routing.
Many organizations use these features together as part of the same customer support workflow.
The two critical dates that matter most
Although the affected features require different levels of migration effort, they all follow the same deprecation timeline.
| Feature | Announced | Development ends | Removal starts |
| AI agents Essential | March 30, 2026 | August 31, 2026 | December 10, 2026 |
| Legacy AI agent functionality (bot builder, answers, and intents) | March 30, 2026 | August 31, 2026 | December 10, 2026 |
| Default messaging responses* | March 30, 2026 | August 31, 2026 | December 10, 2026 |
While the timeline itself is straightforward, Zendesk has outlined what happens at each stage so organizations can align their migration plans accordingly:
- August 31, 2026 — this is the point at which AI agents Essential and legacy AI agent functionality (bot builder, answers, and intents) transition into the maintenance-only phase. After this date, Zendesk will limit support to critical issues required to keep the functionality operational until its retirement.
- December 10, 2026 — Zendesk will fully remove the affected features. Any workflows, automations, or customer support experiences that still depend on them will stop functioning.
Many organizations focus on December 2026 as the final deadline. In practice, however, August is the more important date, as customers who have not migrated by then will continue relying on legacy functionality with only limited support from Zendesk.
What Zendesk’s New AI Agent Experience Looks Like
The good news is that Zendesk is replacing legacy AI agent functionality with a more capable AI agent experience. Beginning May 11, 2026, Zendesk initiated a phased rollout of the new AI agent packaging that introduced both new functionality and changes to the way AI capabilities are packaged and delivered.
How the new AI agent packaging model works
The major shift is that Zendesk removed the distinction between “Essential” and “Advanced” AI agent plans. Previously, Essential AI agents provided baseline AI capabilities, while Advanced AI agents offered access to more sophisticated functionality, such as:
- Agentic reasoning — determining the next action dynamically based on context and customer responses rather than following a fixed decision tree.
- Multi-step procedures — executing a sequence of actions to complete a task instead of simply answering a single question.
- External API integrations — retrieving data from or triggering actions in third-party systems rather than relying solely on Zendesk knowledge-base content.
In the new AI agent packaging, these capabilities are included in the unified Zendesk AI experience available across Suite and Support plans.
The change also affects how AI capabilities are packaged and billed, shifting the focus away from feature-based tiers and toward AI agent outcomes, such as the number of customer requests resolved by AI agents.
Key vocabulary changes
As Zendesk moves away from intent-based workflows and toward more autonomous, agentic AI, many familiar concepts are being restructured. As a result, the terminology used across the platform is changing as well.
| Legacy term | New term | Why Zendesk changed it |
| Bot builder | Dialogue builder | AI conversations are now built around dialogues rather than predefined bot flows. |
| Answer | Dialogue | Responses become part of broader AI-driven conversations. |
| Intent | Use case | AI agents are organized around customer scenarios rather than intent classification. |
The channel separation rule that enterprise teams must know
Each AI agent in the new Zendesk AI experience works on one channel type only: messaging or email. This marks a significant change from the legacy AI agent model: under AI agents Essential, a single AI agent could be used across multiple customer communication channels.
As a result, organizations that support customers through both messaging and email need to create separate AI agents for each channel. This is an important consideration during migration planning, as it can significantly increase migration scope, particularly in enterprise environments with complex cross-channel support operations.
How to Audit Your Legacy Setup Before You Migrate
Migrating to Zendesk's new AI agent experience is not a simple lift-and-shift exercise. Attempting to rebuild legacy configurations without fully understanding how they work can lead to broken workflows and unnecessary rework.
The risk applies to all customers, but it is particularly high for enterprise teams with undocumented bot flows. To mitigate it, it's important to audit your current setup before making any changes in the new environment.
Step 1. Inventory your legacy AI agents
Zendesk's migration guide recommends reviewing all legacy deployments currently active before beginning migration. The process differs for legacy AI agents and default messaging responses.
To inventory legacy AI agents:
- In the Admin Center sidebar, click AI → AI agents.
- Under the Essential heading, click AI agents one by one to open their settings.
- Review all available tabs for each AI agent and count answers and intents where applicable.
To inventory default messaging responses:
- In the Admin Center sidebar, click Channels → Messaging and social → Messaging.
- Click brands one by one to open their settings.
- Click the Responses section. If it contains configured messages, review and count them.
Once the review is complete, consolidate the findings into a single inventory. This will serve as the foundation for estimating migration effort and planning the next steps.
Step 2. Score complexity
Migration complexity varies from one implementation to another. It depends on factors such as the size of the deployment, workflow complexity, integrations, channels, and the overall implementation architecture. Assessing it realistically is important for choosing the right migration approach.
The table below provides a general guideline. Use it as a starting point, but consider your actual setup and available resources when estimating effort.
| Complexity level | Typical characteristics | Recommended approach |
| Low | Fewer than 10 intents, simple answer flows, no API integrations, single-channel deployment. | Self-migration is usually feasible. |
| Medium | 10–50 intents, multiple answer flows, basic branching logic, and 1–3 API integrations. | Plan and resource a 2–4 week migration project. |
| High | More than 50 intents, many answer flows, extensive branching, multiple API integrations, and/or other highly customized configurations. | Budget 6–8 weeks and consider engaging a Zendesk AI Expert* or third-party migration partner. |
*The recommendation to engage a Zendesk AI Expert for high-complexity migrations comes from Zendesk's official migration guidance.
Step 3. Document before you rebuild
The new AI agent experience is not a one-to-one replacement for legacy AI agents. Although Zendesk has announced an automated migration tool, details about how it will work are not yet available.
Whether you decide to wait for the tool or migrate earlier, maintaining a complete record of your current configuration is a sensible precaution.
- All answers and intents associated with each AI agent, along with the relationships between them
- All branching paths, conditions, and decision points within answer flows
- Any API integrations, third-party systems, or custom actions used by each workflow
- The channels and brands where each AI agent is currently deployed
- Any default messaging responses used alongside the Zendesk legacy AI functionality
For complex deployments, screenshots or screen recordings are often the fastest way to preserve workflow logic.
What If This Migration Prompts a Bigger Question?
For many organizations, migration to Zendesk's new AI agents experience will be a straightforward upgrade. For some enterprise teams, however, a mandatory rebuild of support automation may be an opportunity to step back and evaluate whether Zendesk remains the right long-term fit.
It may be worth exploring alternative solutions if:
- AI-related costs are continuing to rise
- Your team requires capabilities beyond Zendesk's current offering
- Your contract renewal aligns with the Q4 2026 migration timeline
- Freshdesk is often chosen by small and mid-sized organizations for its ease of use, straightforward implementation, and competitive pricing. It offers Freddy AI, which includes Freddy AI Agent (autonomous customer-facing AI), Freddy AI Agent Studio, and Copilot capabilities.
- Salesforce Service Cloud is typically a better fit for large enterprises that need extensive customization, advanced automation, and deep integration with other Salesforce products. It offers Agentforce, Salesforce's enterprise agentic AI platform, which enables AI agents to reason, use tools, execute workflows, and complete tasks autonomously across the Salesforce ecosystem.
Considering a platform change? Help Desk Migration lets you safely transfer your data to a new platform through a guided Migration Wizard.
Frequently Asked Questions about Zendesk Legacy AI Shutdown
Zendesk removes AI agents, Essential, legacy bot builder/answers/intents, and default messaging responses. Removal starts December 10, 2026. Any workflow, automation, or customer support experience that still relies on these features stops working immediately. Development already ended on August 31, so removal is the final, hard cutoff, but not a gradual phase-out.
After December 10, 2026, legacy AI agents, bot builder flows, and default messaging responses stop functioning entirely. Customer conversations relying on these workflows break. Between August 31 and December 10, Zendesk only offers maintenance-level support, so waiting until the deadline leaves no time to fix issues.
Yes. In the new AI agent experience, every AI agent works on one channel only, messaging or email, regardless of plan. This applies across Suite and Support. Teams supporting both channels need separate AI agents for each channel, which increases the scope of migration compared to the old multi-channel bot builder setup.
It depends on complexity. Zendesk's own guidance points to 2-4 weeks for medium setups (10-50 intents, a few integrations) and 6-8 weeks for high-complexity deployments with 50+ intents, extensive branching, or multiple API integrations. Enterprise teams with undocumented bot flows should budget on the higher end.
Technically, yes, but only in maintenance mode. Zendesk stops development on August 31, 2026, and limits support to critical issues, keeping the functionality operational. Bot builder still works until December 10, but there are no updates or new fixes beyond critical ones. Treat this window as borrowed time, not stability.
