For a six-seat team handling 5,000 monthly AI resolutions, that math lands near $11,090 per month. That's a real number, and it's why thousands of SMBs are quietly comparing alternatives this week.
This isn't a takedown post. Zendesk still works well for plenty of teams. But if you're under 50 employees, the new pricing model probably wasn't designed with you in mind. Below is a 7-day playbook covering what's being deprecated, what stays alive, and three honest migration paths, including options that have nothing to do with us.
TL;DR: Zendesk's Essential AI and Advanced AI Add-on shut off May 11, 2026, replaced by a Resolution Platform billing $1.50–$2.00 per AI resolution (Zendesk, 2026). SMBs under 50 seats have three paths: upgrade in place, switch to a peer helpdesk like Help Scout or Front, or move to flat AI-native pricing. The 7-day window is tight but workable if you start with data export today.
What is Zendesk actually deprecating on May 11?
Zendesk is sunsetting Essential AI (the bundled tier on legacy Suite plans) and the Advanced AI Add-on, both replaced by the new Resolution Platform on May 11, 2026 (Zendesk Help Center, 2026). Existing tickets, macros, and triggers stay intact. What disappears is the AI layer: auto-replies, intelligent triage, and bot flows tied to the old SKUs.
Citation capsule: Zendesk's May 11, 2026 deprecation pulls Essential AI and the Advanced AI Add-on from legacy Suite contracts, forcing customers onto the Resolution Platform's $1.50-per-resolution committed pricing and $2.00 overage rate (Eesel AI, 2026).
What still works after May 11
The core helpdesk doesn't go anywhere. Tickets, agents, views, macros, triggers, SLAs, the standard Help Center, and most native integrations keep functioning. Email, chat, and voice channels stay open. If you weren't actively using the Answer Bot, Intelligent Triage, or Generative Replies, your morning of May 11 will look identical to your morning of May 10.
What actually breaks
Three things stop working: AI-powered auto-replies on email and chat, Intelligent Triage classifications, and the legacy Answer Bot flows on your Help Center widget. In our migration audits, the Answer Bot widget is the most-missed feature. It's often the only thing standing between a customer and an empty contact form. If you've built workflows around AI-suggested macros, those suggestions disappear too.
What's the full deprecation timeline?
Three dates matter, not one. May 11, 2026 ends Essential AI and the Advanced AI Add-on. August 31, 2026 shuts off the legacy AI Agent service for non-Resolution-Platform customers. December 10, 2026 retires older bot-builder flows entirely (Zendesk, 2026). Treat May 11 as the first cliff, not the only one.
Why the staggered shutoff matters
If you migrate to Resolution Platform on May 11, you're fine through all three dates. If you stay on a legacy plan and lose Essential AI on May 11, you still have until August 31 before older AI Agent flows go dark, and until December 10 before Flow Builder bots retire. That's roughly seven months of runway, which is enough time for a careful migration but not enough for a leisurely one.
What are the three migration paths for SMBs?
SMBs face three realistic options: upgrade to Zendesk's Resolution Platform, switch to a peer helpdesk like Help Scout or Front, or move to AI-native flat pricing. The right pick depends on team size and ticket volume. At 5,000 monthly AI resolutions across 6 seats, Zendesk's new model hits roughly $11,090 per month before any add-ons (Eesel AI, 2026).
Path A: Stay on Zendesk and upgrade to Resolution Platform
Resolution Platform bills $1.50 per committed AI resolution with $2.00 overages, layered on top of $115-per-agent Suite Pro seats and a $50-per-agent AI add-on (Zendesk pricing, 2026). Run the math for your team:
- 6 agents × $115 Suite Pro = $690
- 6 agents × $50 AI add-on = $300
- 5,000 committed resolutions × $1.50 = $7,500
- 1,300 overage resolutions × $2.00 = $2,600
- Monthly total: $11,090
Across 14 SMB migration calls we ran in April 2026, every team running 3,000+ monthly AI resolutions ended up with a 40-60% bill increase under the new model. Path A makes sense if your team is mid-market with predictable volume and existing custom apps you can't easily rebuild. It rarely makes sense for a 5-person team.
Path B: Move to a peer helpdesk (Help Scout, Pylon, Front)
Three credible peers cover most SMB use cases:
- Help Scout: $25/user/mo (Standard) or $50/user/mo (Plus). AI Assist included on Plus. Best for email-first support teams under 30 seats.
- Pylon: $59/seat/mo. B2B-focused with Slack Connect support. Strong if you support enterprise customers in shared Slack channels.
- Front: $59-$99/seat/mo. Shared inbox model, good for teams that already work in email-thread style.
Help Scout is the most underrated Zendesk replacement for sub-20-seat teams because its AI features are bundled, not metered. If you're paying $1.50 per AI resolution today, switching to a $50/seat plan with unlimited AI assist often cuts your bill in half overnight, without changing pricing models, just changing vendors.
Path C: Move to AI-native flat pricing
A newer category of tools (Corebee, Plain, Lorikeet, and others) prices flat regardless of resolution count. The 2026 baseline AI resolution rate sits around 65% (HubSpot Breeze, 2026), which means flat pricing pays off fastest for teams above 1,000 monthly tickets. Below that, the difference versus Help Scout is small.
What's the 7-day migration playbook?
Seven days is tight but workable if you start with data export on day one. Migration vendors process roughly 2,000 tickets per hour (Help Desk Migration, 2026), so a 50,000-ticket history transfers in about 25 hours of automated runtime. The hard part isn't the data move. It's the macro and automation rebuild.
Day 1 (Tuesday): Export everything
Pull a full Zendesk export before you do anything else. Settings → Account → Reports → Export. Request both ticket data and user data as JSON. Zendesk emails the download link within 24 hours for most accounts. While you wait, document your current macros, triggers, automations, and SLA policies in a shared spreadsheet. You'll thank yourself on Day 4.
Day 2 (Wednesday): Pick a destination and run a parallel pilot
Sign up for two finalists. Don't commit yet. Import 100-500 sample tickets into each, configure one macro, send a test email. The goal is to feel both UIs before you decide. We've watched teams pick the "obvious" replacement on paper and hate it three weeks in. A two-hour parallel test prevents most regret.
Day 3 (Thursday): Map your integrations
List every integration touching Zendesk: Slack, JIRA, Stripe, your CRM, your e-commerce platform, your phone system. Check the destination's marketplace for native parity. Anything missing becomes a Zapier workflow, a webhook, or a "live with the loss" decision. Don't try to rebuild custom apps in week one. Flag them for week three.
Day 4 (Friday): Rebuild macros and automations
Use the spreadsheet from Day 1. Recreate macros in priority order (most-used first, edge cases last). Most teams cover 80% of their volume with 10-15 macros, so don't try to port all 200 you've accumulated since 2019. This is also when you set up routing rules, business hours, and SLA policies in the new tool.
Day 5 (Saturday): Run the full data migration
Schedule a vendor like Help Desk Migration or Relokia to move ticket history overnight. Saturday-into-Sunday minimizes business impact. Verify a sample of migrated tickets on Sunday morning. Check that attachments, internal notes, and customer details transferred cleanly. If the sample looks wrong, pause and call the vendor before the full cutover.
Day 6 (Sunday): Train the team and update routing
90-minute team session covering the new UI, top 10 macros, and escalation paths. Update your contact form, Help Center widget, and any "support@" email forwarding rules to point at the new system. Keep Zendesk receiving in parallel for 48 hours as a safety net.
Day 7 (Monday): Go live and monitor
Flip the switch in the morning. Watch first-response times, ticket volume, and customer-reported issues hour by hour. AI support pricing models. Have one engineer on standby for the first business day. Most issues surface in the first four hours; if you're past noon with no fires, you've landed it.
What actually gets lost in migration?
Four things rarely survive a Zendesk move cleanly: custom Zendesk apps from the marketplace, Zendesk-specific integrations built on the Sunshine platform, deeply customized macro logic, and historical reporting dashboards. Help Desk Migration's vendor data shows roughly 95% of standard ticket fields transfer, but custom objects and Sunshine app data are the typical casualties (Help Desk Migration, 2026).
What transfers cleanly
Tickets, comments, attachments, customer records, organizations, basic tags, and assignee history all migrate without drama. Macros and triggers transfer as text, meaning the labels survive, but you'll rebuild the logic in the new tool. CSAT history is usually exportable as CSV, though it rarely lives natively in the destination.
What you'll rebuild from scratch
Custom marketplace apps. Sunshine Conversations integrations. Anything tied to Zendesk's Apps Framework. Custom-coded triggers using Zendesk's webhook + JSON template syntax. And reporting dashboards built in Explore: those don't port anywhere, period. Plan to recreate the three or four reports that actually matter and let the rest go.
What if you do nothing?
If you ignore May 11, your AI features quietly stop working, but your helpdesk keeps running. Tickets still flow. Agents still respond. The customer-facing impact is mostly invisible until someone notices the Answer Bot widget got dumber. Roughly 5-7% annual SMB SaaS churn is normal across the category (We Are Founders, 2026), so "switching is normal," but so is "staying put."
When standing pat makes sense
If your team handles fewer than 500 monthly tickets, you barely used Essential AI to begin with, and your team knows the Zendesk UI cold, doing nothing is a defensible call. You'll lose the AI suggestions and Answer Bot, but the core product stays put. Klarna's reversal, pulling back from a 700-agent AI replacement and rehiring humans (CX Today, 2025), is a reminder that aggressive AI deployment can backfire. Sometimes the best move is the boring one.
When standing pat is risky
If you've built customer-facing automations on top of Essential AI, or your Help Center widget is doing real deflection work, the May 11 cliff is real. Test what breaks before you commit to inaction.
Is Corebee the right fit?
Corebee is one Path C option among several. We charge $99 per month flat, regardless of resolution count, for teams under 50 employees. That's it: no per-seat math, no per-resolution surprises. If you're a 5-person team handling 2,000-8,000 monthly tickets, the math usually works in our favor. Above 50 employees, Pylon or Front is often a better cultural fit.
We're not the answer for every Zendesk refugee. If you need deep enterprise integrations, ITSM workflows, or 200-seat scale, look elsewhere. If you want flat pricing, a clean inbox, and AI that doesn't bill per resolution, we're worth a 14-day trial. Corebee pricing
The honest framing: Zendesk's deprecation is real, the math has changed, and you have three credible paths. Pick the one that matches your team size and pain tolerance. We'll be happy if you pick us, fine if you pick Help Scout, and rooting for you either way.