---
title: "Zendesk Contact Center GA (2026): Talk vs Un-CCaaS & Migration Paths"
slug: zendesk-contact-center-ga-2026-talk-vs-un-ccaas-migration-paths
date: 2026-07-10
author: Abdul
categories: [Zendesk]
excerpt: "Zendesk Contact Center hit GA on June 9, 2026. Here's how it differs from Zendesk Talk, what Un-CCaaS means, and what migrating actually involves."
tldr: "Zendesk Contact Center (Local Measure + Amazon Connect) is now GA — not Talk with a new name. Migration requires IVR rebuilds, recording extraction, and routing reconfiguration."
canonical: https://clonepartner.com/blog/zendesk-contact-center-ga-2026-talk-vs-un-ccaas-migration-paths/
---

# Zendesk Contact Center GA (2026): Talk vs Un-CCaaS & Migration Paths


Zendesk Contact Center — the company's full CCaaS platform built on the Local Measure acquisition — went generally available on June 9, 2026. This is not an incremental update to Zendesk Talk. It is a full Contact Center as a Service (CCaaS) solution natively embedded in the Zendesk Agent Workspace, built on an Amazon Connect backend.

If you're evaluating whether to upgrade from Zendesk Talk, migrate from a third-party CCaaS like Genesys or NICE, or just trying to understand what "Un-CCaaS" actually means in practice, this guide covers the technical architecture, the real differences from Talk, the pricing structure, and what migration actually looks like.

## What Is the Zendesk Contact Center GA Release?

**Zendesk Contact Center** is the productized version of Local Measure's Engage platform. <cite index="29-1,29-2">Zendesk completed its acquisition of Local Measure — a CCaaS and advanced voice solutions provider and long-standing AWS partner — in May 2025, expanding its platform into larger, high-volume service environments through Local Measure's integration with Amazon Connect.</cite>

<cite index="2-1,2-2,2-3">The general availability rollout was announced on June 9, 2026, with the rollout completing by June 19, 2026. Contact Center is now built directly into the Zendesk Agent Workspace.</cite> <cite index="4-1">The product was first introduced at Zendesk Relate 2026.</cite>

The GA release includes several key changes:

- **Unified experience** — <cite index="2-4">Call controls appear in the workspace without a separate Marketplace app.</cite> The legacy Local Measure ZAF sidebar app is gone.
- **Single status management** — <cite index="2-5">One status picker controls availability across both Zendesk and Amazon Connect.</cite>
- **Simpler authentication** — <cite index="2-6">Agents sign in once to Zendesk with no additional authentication required.</cite>
- **Native settings** — <cite index="2-7">Admins access Contact Center settings from the Product menu.</cite>
- **30+ language support** — <cite index="2-8">Now available in 30+ languages.</cite>
- **Outbound call restriction** — <cite index="1-17">Admins can now restrict outbound calling permissions by agent group to reduce misrouting and costs.</cite>

<cite index="2-9">Your Amazon Connect configuration, phone numbers, call recordings, and workflows remain unchanged.</cite> That last point matters for existing Local Measure customers: this isn't a rip-and-replace of the underlying telephony. It's a UI and integration layer change.

### The "Un-CCaaS" Positioning

<cite index="11-2,11-3,11-4,11-5">Zendesk isn't marketing the solution as "CCaaS." Instead, it's positioning the platform as "Un-CCaaS" — a deliberate echo of T-Mobile's 2013 "Un-carrier" campaign to frame itself as a disruptive market alternative.</cite>

The strategic logic: <cite index="11-6,11-7,11-8">many large contact centers have experimented with several CCaaS solutions, having rushed into "COVID-contracts" that didn't deliver. Zendesk hopes to help enterprises cut through this experimentation and implement a solution adjacent to their existing Zendesk CRM.</cite>

<cite index="11-10,11-11">CEO Tom Eggemeier hopes to secure between 10,000 and 30,000 CCaaS customers. Jonathan Barouch, GM of Contact Center at Zendesk (formerly Founder and CEO of Local Measure), has reaffirmed this objective.</cite>

Strip away the branding and the pitch is straightforward: buy your contact center from the same vendor that already holds your ticketing data, knowledge base, and AI agents. Whether the "Un-CCaaS" label sticks is an open question — <cite index="17-2">analysts are unsure about the staying power of the name but enthusiastic about Zendesk's resolution-focused differentiation.</cite>

## Zendesk Contact Center vs. Zendesk Talk: What's the Difference?

This is the question most existing Zendesk customers are asking. The short answer: **Zendesk Talk** (also called Zendesk Voice) is the basic voice capability bundled into Suite plans. **Zendesk Contact Center** is the enterprise-grade CCaaS add-on built on Amazon Connect.

| Capability | Zendesk Talk (Voice) | Zendesk Contact Center |
| --- | --- | --- |
| Target customer | SMB and mid-market | Mid-market to enterprise (250–4,000+ agents) |
| Infrastructure | Twilio | Amazon Connect (AWS) |
| IVR | Basic, single-level | Multi-level, AI-enhanced IVR with conversational capabilities |
| Routing | Round-robin, group-based (no skills-based routing) | Skills-based, multi-queue, priority routing with Amazon Connect routing profiles |
| Status management | Zendesk-only | Unified across Zendesk + Amazon Connect |
| WFM integration | Limited | Native WFM and QA via Zendesk WEM |
| Call recording | Basic (no PII/PCI redaction in recording file) | Enterprise-grade with pause/resume, post-call redaction, and KMS encryption |
| Outbound dialing | Basic click-to-call | Outbound campaigns, group-based permissions, and click-to-call |
| AI agent integration | Limited to Talk-level triggers | Zendesk AI agents can handle voice IVR deflection and intent detection pre-routing |
| Pricing model | Bundled in Suite + per-minute usage | Add-on starting at $83/agent/month + Minutes Blocks + AWS usage |
| Admin experience | Zendesk admin panel | Dedicated Contact Center settings in Product menu |

<cite index="6-5,6-6">Zendesk itself has acknowledged that Talk "covered it well in the mid-market and SMB segments, but large enterprises with complex contact centers, multiple teams, languages, and workflows exposed our limitations."</cite>

Both products coexist. There is no forced migration from Talk to Contact Center. But the signal is clear: <cite index="6-4">Zendesk hopes to grow its CCaaS install base from hundreds to thousands, something it couldn't achieve with its lighter-weight Zendesk Voice offering.</cite> Future feature investment will prioritize Contact Center — Zendesk shipped native call monitor and barge capabilities for Contact Center on June 29, 2026, which signals where development effort is going.

> [!NOTE]
> **Naming confusion:** You'll see "Zendesk Talk," "Zendesk Voice," and "Zendesk Contact Center" used in overlapping ways across Zendesk's own documentation. Talk and Voice refer to the same legacy product. Contact Center is the new, separate enterprise offering built on Amazon Connect.

### Pricing Structure

<cite index="31-2,31-3">Zendesk Contact Center adds enterprise voice, IVR, and Amazon Connect telephony to any Suite plan, starting from $83/agent/month.</cite> On top of that base:

- <cite index="31-6,31-7">A Minutes Block is a pre-purchased bundle of minutes used for features such as conversational IVR, outbound campaigns, real-time transcription, call recording, screen sharing, and video calling. One Minutes Block per user per month is included with the purchase of Contact Center.</cite>
- <cite index="31-5">AWS Added Services and Amazon Connect Telephony usage are billed based on consumption.</cite>

**TCO example at different scales:**

| Team size | Suite Professional | Contact Center add-on | Annual seat cost (before minutes/AWS) |
| --- | --- | --- | --- |
| 50 agents | $115/agent/month | $83/agent/month | **$118,800/year** |
| 250 agents | $115/agent/month | $83/agent/month | **$594,000/year** |
| 1,000 agents | $115/agent/month | $83/agent/month | **$2,376,000/year** |

AWS telephony consumption adds to these baseline costs. Amazon Connect charges $0.018/minute for inbound voice (US) and $0.018/minute for outbound voice (US), plus per-second charges for telephony minutes at carrier rates that vary by country. For a 250-agent operation handling 500,000 inbound minutes per month, AWS telephony alone could add $9,000–$15,000/month depending on region mix and outbound volume.

When comparing against a standalone CCaaS like Genesys Cloud CX (which lists at $75–$155/user/month depending on tier) or NICE CXone (typically $71–$209/user/month), include the cost of maintaining a Zendesk-to-CCaaS integration layer — typically $15,000–$50,000/year in middleware, CTI connector licensing, and engineering time — in your current-state TCO.

### Where Talk Still Fits

If you have a small voice footprint — a few lines, straightforward IVR, and no need for Amazon Connect flows or tighter governance — Talk is the right answer. Zendesk is still documenting, selling, and updating Talk as of July 2026. A decision tree:

- **Stay on Talk** if: <50 agents, single-level IVR, no skills-based routing requirement, no WFM integration need, no PCI/HIPAA call recording redaction requirement.
- **Evaluate Contact Center** if: 50–250 agents with growing complexity, multi-language queues, or WFM/QA needs.
- **Contact Center is likely required** if: 250+ agents, skills-based routing, multi-queue structures, compliance-sensitive recording, or AI-driven IVR deflection.

### Where Contact Center Fits Better

If voice is central to your operation, the value isn't just more features. It's the operating model: one workspace, one login, one status layer, Amazon Connect as the routing engine, and a clearer path into Zendesk QA and WFM add-ons.

> [!WARNING]
> **Do not treat Talk to Contact Center as a toggle.** Talk is a Twilio-based voice product. Contact Center is an Amazon Connect add-on with different billing, routing, status mapping, network rules, and recording behavior. They share the Zendesk workspace, but they share almost nothing at the infrastructure layer.

## The Technical Architecture Behind Zendesk Contact Center

Zendesk did not build a proprietary telephony engine. They built a deep integration layer on top of Amazon Connect. When you purchase Zendesk Contact Center, you are provisioning an Amazon Connect instance managed through the Zendesk UI.

Here's what that means in practice:

**Telephony layer.** Amazon Connect provides the global telephony network — phone numbers, call routing, recording infrastructure, and SIP trunking. You're running on AWS infrastructure whether you realize it or not. Amazon Connect currently operates in 14 AWS regions globally (including US East, US West, EU Frankfurt, EU London, Asia Pacific Sydney, Singapore, Tokyo, and Seoul), which determines available DID number countries and call quality routing paths.

**Identity and SSO.** Zendesk documents a one-to-one link between each Contact Center environment and a specific Amazon Connect instance. Zendesk acts as a SAML 2.0 identity provider to the AWS services behind Contact Center — agents keep their normal Zendesk login and are federated onward to Amazon Connect, Amazon Cognito, and the AWS console without a separate SSO project. <cite index="14-22">In the first post-acquisition year, Zendesk customers who bought Contact Center were required to have both a Zendesk and an AWS account.</cite> The GA release eliminates that friction.

**Agent workspace.** <cite index="4-15,4-16">Zendesk cites customer feedback as the primary reason for moving call controls into the workspace — many customers felt the sidebar app was disconnected and managing two separate status pickers was confusing.</cite>

**AI layer.** <cite index="5-16,5-17,5-18,5-19">The Contact Center overlays the core Zendesk offering, providing a unified CRM-CCaaS experience with a brand-new call console. It allows agents to transfer calls, conference with available representatives, place calls on hold, and pause or resume recordings — all within a unified workspace alongside customer history and intent data.</cite>

Zendesk AI agents interact with Contact Center at the IVR layer: AI agents can handle pre-routing intent detection, deflect calls to self-service resolution (e.g., order status lookups, password resets) before an agent is involved, and pass structured intent and customer context data to the agent workspace when escalation occurs. This means the Amazon Connect contact flow can invoke a Zendesk AI agent as a first touchpoint, with the handoff populating the agent's screen with the full conversation history and detected intent — no re-asking the customer's question.

<cite index="52-11,52-12">To simplify procurement, Zendesk offers a bundled commercial solution with AWS, bringing enterprise telephony, platform minutes, and AI features into a single arrangement billed by Zendesk.</cite>

### Amazon Connect Quotas That Affect Migration Planning

These AWS defaults become planning constraints for any high-volume deployment:

| Quota | Default limit | Adjustable? | Planning impact |
| --- | --- | --- | --- |
| Concurrent active calls | 10 per instance | Yes (request increase) | Must increase before any production deployment |
| Contact flows | 100 per instance | Yes | Complex IVRs hit this fast — audit flow count early |
| Lambda functions | 50 per instance | Yes | Relevant if using external data lookups in flows |
| Agent statuses | 50 per instance | **No — hard limit** | Must rationalize statuses before migration |
| `SearchContacts` API | 0.5 requests/second | Yes | Affects reporting and analytics integrations |
| Routing profiles | 500 per instance | Yes | Rarely a bottleneck but check at scale |
| Quick connects | 100 per instance | Yes | Impacts transfer/conference button configurations |

Quota increases are possible for most limits, but AWS documentation states larger increases can take up to three weeks and very large global increases can take months. If you're moving a high-volume operation, request quotas early — not during UAT.

### Recording and Compliance

AWS stores native recordings in an S3 bucket created for the instance, with KMS server-side encryption enabled by default. Recording only exists if you enable it in the contact flow. Post-call sensitive-data redaction is available through Amazon Connect Contact Lens, but AWS documentation explicitly notes it may miss some instances of sensitive data — particularly non-standard formats of credit card numbers, custom account identifiers, or sensitive data spoken in unsupported languages.

In Zendesk, admins can choose whether tickets receive redacted, unredacted, or no recording/transcript links. This is a significant upgrade over Talk, which cannot redact PII/PCI in recording files. But it still requires explicit policy decisions before rollout. Specifically, you must define: (1) which contact flows trigger recording, (2) whether agents can pause/resume recording for sensitive data capture, (3) the S3 retention policy and lifecycle rules, (4) who has access to unredacted recordings, and (5) whether transcripts are written back to Zendesk tickets.

<cite index="51-8">Zendesk Contact Center has earned official support for HIPAA-enabled accounts.</cite> AWS will sign a Business Associate Agreement (BAA) for Amazon Connect, which is required for PHI handling. If healthcare compliance is a factor, see our [HIPAA Compliance for Healthcare Help Desks](https://clonepartner.com/blog/blog/hipaa-compliance-for-healthcare-help-desks/) guide.

### Network Requirements

Zendesk Contact Center's network requirements are AWS-oriented. Specific requirements include:

- **Domain allowlisting:** Zendesk Contact Center domains, Amazon Connect API endpoints, and CloudFront CDN domains must be reachable from agent workstations.
- **Port requirements:** HTTPS (443) for signaling and API traffic; UDP ports for WebRTC media (typically a range within 1024–65535 depending on region).
- **Proxy exclusions:** Real-time audio (RTP/WebRTC) should bypass web proxies, SSL inspection, and packet-filtering firewalls where possible. Proxy-induced latency above 150ms one-way will produce audible call quality degradation.
- **Bandwidth:** Amazon Connect recommends a minimum of 100 kbps per concurrent call for audio quality.

This is a different network profile than Talk — plan for firewall and proxy changes before agent rollout, and conduct call quality testing from every distinct network path (office, VPN, home) agents will use.

## Feature Parity: Zendesk Contact Center vs. Genesys Cloud CX vs. NICE CXone

If you're evaluating a migration from a mature CCaaS, this feature comparison shows where Zendesk Contact Center has reached parity and where gaps remain as of July 2026:

| Capability | Zendesk Contact Center | Genesys Cloud CX | NICE CXone |
| --- | --- | --- | --- |
| Skills-based routing | ✅ Available | ✅ Available (predictive routing add-on) | ✅ Available |
| Predictive routing (AI-based) | ❌ Not available | ✅ Available | ✅ Available |
| Outbound dialer (preview) | ✅ Available (campaigns) | ✅ Available (preview, progressive, predictive) | ✅ Available (preview, progressive, predictive, agentless) |
| Outbound dialer (predictive) | ❌ Not available | ✅ Available | ✅ Available |
| Workforce management | ✅ Available (Zendesk WFM add-on) | ✅ Available (native) | ✅ Available (native) |
| Quality assurance / QA | ✅ Available (Zendesk QA add-on) | ✅ Available (native) | ✅ Available (native) |
| Screen recording | ❌ Not available | ✅ Available | ✅ Available |
| Speech analytics | ✅ Partial (via Contact Lens) | ✅ Available | ✅ Available |
| Real-time transcription | ✅ Available | ✅ Available | ✅ Available |
| Post-call transcription | ✅ Available | ✅ Available | ✅ Available |
| PCI/PII redaction | ✅ Available (post-call, with limitations) | ✅ Available (real-time and post-call) | ✅ Available (real-time and post-call) |
| Digital channels (chat, email, social) | ✅ Native in Zendesk Suite | ✅ Available | ✅ Available |
| CRM / ticketing | ✅ Native in Zendesk Suite | ❌ Requires integration | ❌ Requires integration |
| Native case management | ✅ Native | ⚠️ Recently added | ⚠️ Recently added |
| Video calling | ✅ Available | ✅ Available | ✅ Available |
| Co-browse / screen share | ✅ Available | ✅ Available | ✅ Available |
| Monitor / barge / whisper | ✅ Available (shipped June 29, 2026) | ✅ Available | ✅ Available |
| Journey management / orchestration | ❌ Not available | ✅ Available | ⚠️ Partial |
| Sentiment analysis (real-time) | ⚠️ Partial (via Contact Lens) | ✅ Available | ✅ Available |

**Key gaps to validate before committing:** Zendesk Contact Center does not currently offer predictive outbound dialing, screen recording, real-time PCI masking (only post-call redaction), or journey orchestration. If your current operation depends on any of these, migration requires either accepting the gap, finding a third-party supplement, or waiting for Zendesk's roadmap to close it.

## Migrating from Third-Party CCaaS to Zendesk Contact Center

<cite index="11-9">Zendesk is wagering that many of its 110,000+ customers will deploy a contact center offering that sits inside the broader Zendesk Resolution Platform.</cite> The pitch to companies running Genesys, NICE CXone, Five9, or Talkdesk alongside Zendesk is consolidation: stop maintaining two ecosystems, stop paying for integration middleware, and stop forcing agents to context-switch.

<cite index="50-10,50-11">Many existing Zendesk customers have implemented a third-party CCaaS solution, with customer support teams operating across two separate ecosystems.</cite> That creates real technical overhead — and it's where migration demand is emerging.

But consolidation is not the same thing as feature parity. Run a feature-by-feature gap review using the comparison table above before you commit to a date.

### Migration Timeline Benchmarks

Based on the complexity tiers observed across CCaaS migrations, these are realistic timeline ranges:

| Migration scenario | Agent count | Typical duration | Key timeline drivers |
| --- | --- | --- | --- |
| Simple (Talk → Contact Center) | <100 | 4–6 weeks | IVR rebuild, number porting, agent training |
| Moderate (single CCaaS → Contact Center) | 100–500 | 8–14 weeks | Flow reconstruction, status rationalization, compliance review, integration redirect |
| Complex (multi-site CCaaS → Contact Center) | 500–2,000 | 14–26 weeks | Multi-region porting, historical data migration, WFM/QA tool cutover, phased rollout |
| Enterprise (global CCaaS consolidation) | 2,000+ | 6–12 months | Multi-instance Connect architecture, regulatory review per jurisdiction, union/works council requirements |

The single largest variable is IVR and contact flow reconstruction. A 50-flow Genesys Architect environment typically requires 3–5 weeks of dedicated flow engineering to translate into Amazon Connect. There is no automated import tool.

### What Actually Needs to Move

Here's the migration scope most teams underestimate:

**1. IVR trees and call flows.** Your multi-level IVR logic needs to be rebuilt in Amazon Connect's contact flow editor. There is no automated import from Genesys Architect, NICE Studio, or Five9 IVR Designer. This is manual reconstruction work — audit every queue, skill assignment, and schedule condition, then translate complex logic (e.g., "if queue time exceeds 5 minutes, check external database for VIP status, route to overflow group") into Amazon Connect flows. The default 100-flow-per-instance limit matters faster than most teams expect.

**2. Agent status rationalization.** If your current CCaaS has dozens of AUX states, approval states, or regional reason codes, you need to simplify before go-live. The AWS 50-status-per-instance cap is a hard design input, not a footnote. Map every existing status to one of the 50 target states, merge duplicates, and get operations sign-off before you touch Amazon Connect.

**3. Routing rules and queue configurations.** Skills-based routing definitions, priority queues, agent groups, and overflow rules all need to be mapped to Amazon Connect's routing profiles. The data model is different between vendors — Genesys uses skill expressions with boolean logic; Amazon Connect uses routing profiles with queue-priority-delay tuples.

**4. Number porting.** Phone number porting timelines vary significantly by carrier and country:
- **US/Canada:** Typically 7–15 business days for standard ports; complex ports (toll-free, multi-carrier) can take 4–6 weeks.
- **UK/EU:** 5–20 business days depending on the losing carrier; some EU countries require notarized authorization letters.
- **APAC:** Highly variable — Australia typically 5–10 business days; Japan and South Korea can take 4–8 weeks with additional regulatory documentation.
- **Toll-free numbers:** Generally take longer than DID numbers across all regions.

Submit port requests as early as possible, and maintain parallel routing (forwarding from old numbers to new) during the transition window. Number porting failures are the most common cause of voice migration go-live delays.

**5. Historical call recordings.** If you need to retain recordings for compliance (PCI, HIPAA, GDPR), those files need to be extracted from your existing CCaaS, catalogued with metadata (timestamp, agent, customer, duration, disposition codes), and stored accessibly. AWS records native recordings as S3 objects tied to Amazon Connect contacts — there is no official path to turn foreign WAV or MP3 archives into native Contact Center contact records. Old recordings require an archive or linkback strategy.

> [!CAUTION]
> **Do not attach years of raw audio files directly to standard Zendesk tickets.** This will trigger storage limits and severely degrade your Zendesk instance performance. Enterprise migrations should route historical audio files into a secure AWS S3 bucket, mapping pre-signed URLs into custom Zendesk ticket fields. A typical approach: create a custom ticket field "Historical Recording URL," store recordings in S3 with a defined lifecycle policy, and generate time-limited pre-signed URLs on access.

**6. Compliance and redaction policy.** Decide early whether agents should see redacted or unredacted recordings, whether transcripts should write back into Zendesk tickets, and how post-call redaction limits affect your risk model. If your current vendor does live masking (Genesys and NICE both support real-time PCI masking; Zendesk Contact Center currently supports post-call redaction only), validate the exact replacement behavior and whether the gap is acceptable to your compliance team.

**7. Integrations.** Any CTI connectors, CRM screen-pops, WFM feeds, or QA tool integrations pointing at your current CCaaS will break. Each needs to be redirected to the Zendesk/Amazon Connect stack. Common examples: Verint/Calabrio WFM feeds need to point to Zendesk WFM or to Amazon Connect's real-time metrics API; Salesforce screen-pop connectors need to be replaced by Zendesk's native ticket pop; and custom reporting dashboards need to consume Amazon Connect's contact records from S3 or Kinesis streams.

**8. Reporting and analytics baselines.** Your historical KPIs (AHT, ASA, service level, abandonment rate) won't migrate. Export at least 12 months of historical data into a separate BI tool (Looker, Power BI, Tableau) before cutover so you can maintain trend visibility across the transition. Plan for a 2–4 week reporting normalization period post-cutover where metrics may reflect the transition rather than steady-state performance.

> [!NOTE]
> **Edge case:** If you're running custom Lambda functions or Lex bots in an existing Amazon Connect instance, those assets may be reusable when moving to Zendesk Contact Center — since Contact Center runs on Connect. Confirm with your CSM whether your existing Connect instance can be linked or if a new instance is provisioned. If a new instance is required, Lambda functions can be shared across instances within the same AWS account, but Lex bots must be re-associated.

For a broader framework on planning this kind of move, our [Zendesk Migration Checklist](https://clonepartner.com/blog/blog/zendesk-migration-checklist/) covers the general data migration steps. For teams consolidating multiple help desks alongside a CCaaS migration, see [How to Migrate Multiple Help Desks Into One](https://clonepartner.com/blog/blog/how-to-migrate-multiple-help-desks-into-one/).

## Upgrading from Zendesk Talk or Legacy ZAF Apps

The internal upgrade path is simpler but still requires planning.

### If You're on the Legacy ZAF App or Engage

<cite index="2-13,2-14">Existing Contact Center Zendesk Application Framework app and Engage legacy customers: your current setup continues working without disruption. Contact your Customer Success Manager or Zendesk Advocacy team to plan the migration.</cite>

<cite index="2-11,2-12">New Contact Center customers get the native experience enabled by default — no action required.</cite>

Because both the legacy app and the new GA release share the Amazon Connect backend, this is largely a UI and configuration transition. The migration requires CloudFormation version `20260106 / 1.1.2` or later plus omnichannel routing enabled. The practical steps:

- Switch from the Marketplace app to the built-in call controls
- Move from dual status pickers to the unified status picker
- Update admin workflows from the separate Engage admin panel to the native Zendesk Product menu
- Re-test any automations or triggers that referenced the old app's ticket fields or tags
- Verify that any custom ZAF app event listeners (e.g., `voice.dialout`, `voice.status`) are updated to the new event model

### If You're on Zendesk Talk

There is no forced migration. Talk continues to function. But if you're upgrading, understand this: because Talk runs on Twilio and Contact Center runs on Amazon Connect, you should budget for a **voice replatform**, not assume backend portability.

**Reusable assets:** Zendesk-side constructs like brands, user identity, ticket workflows, macros, triggers, and automations that don't reference Talk-specific fields.

**Non-reusable assets:** Phone numbers (porting depends on carrier and region — see porting timelines above), IVR greetings (need to be re-recorded or re-uploaded in WAV/MP3 format to Amazon Connect), routing logic (must be reconstructed in Connect's flow editor), status models, recording and redaction policy, network allowlists, and cost controls.

Historical Talk tickets containing voicemail recordings and call metrics should be preserved so that reporting consistency is maintained across the cutover date. Talk-originated recordings are stored in Zendesk's infrastructure, not AWS — they will remain accessible on existing tickets but will not appear in Amazon Connect's reporting or Contact Lens analytics.

For context on the broader Zendesk ecosystem these changes affect, see our [Ultimate Zendesk Guide for 2026](https://clonepartner.com/blog/blog/ultimate-guide-zendesk-2026/).

## How to Execute a Zero-Downtime Voice Migration

Voice migrations are fundamentally different from standard ticketing migrations. If a ticketing sync is delayed by an hour, agents can catch up. If a voice cutover fails, customers hear dead air and your business stops functioning.

Use this sequence alongside our [Zendesk Migration Checklist](https://clonepartner.com/blog/blog/zendesk-migration-checklist/):

**1. Freeze the live inventory.** Export numbers, queues, IVR trees, agent statuses, recording policies, and every voice-linked integration before anyone starts cleaning things up.

```yaml
voice_migration_inventory:
  numbers: []          # DID numbers, toll-free, international
  queues: []           # Name, skills, priority, overflow target
  ivr_flows: []        # Flow name, entry point number, logic branches
  agent_statuses: []   # Status name, category (available/unavailable/offline), mapping to target
  recording_policy: [] # Which flows record, retention period, redaction rules
  outbound_permissions: [] # Which groups can dial out, number restrictions
  external_integrations: [] # WFM feeds, QA tools, CRM connectors, reporting pipelines
  historical_data: []  # Recording archive location, metadata format, retention requirements
```

If you cannot fill out that inventory from your current environment in one sitting, you are not ready to cut over.

**2. Choose the target operating model.** Decide whether you are standardizing on Contact Center, staying on Talk, or running a temporary hybrid stack. If hybrid, define the coexistence rules: which agent groups use which system, how transfers between systems work (or don't), and the hybrid end date.

**3. Build a mapping matrix.** Queue to queue, line to brand, status to status, recording rule to recording rule. If it's not in the mapping sheet, it doesn't exist in the target system.

**4. Rebuild routing logic.** Translate your IVR and routing rules into Amazon Connect contact flows. This is manual work and the single biggest source of migration delays.

**5. Sync historical data.** Extract historical call metadata and recordings, mapping them securely to Zendesk tickets or S3-backed custom fields without violating API limits. The Zendesk API rate limit (typically 700 requests/minute for Suite plans) constrains bulk ticket updates — plan for batch processing with exponential backoff.

**6. Validate in staging.** Test inbound, outbound, transfers, callbacks, redaction, SSO, reporting, and supervisor workflows with real scenarios — not happy-path demos. Specific test cases that catch issues: transfers between queues, conference calls with 3+ parties, agent disconnect during active call, recording pause/resume with PCI data, supervisor monitor/barge, and concurrent call limits under load.

**7. Cut over in waves.** Start with low-risk numbers or regions, then move high-volume lines once reporting and agent behavior are stable. A typical wave plan for a 500-agent operation:
- Wave 1 (week 1): Internal test group, 10–20 agents, low-volume queue
- Wave 2 (week 2): Single region or product line, 50–100 agents
- Wave 3 (week 3–4): Remaining queues, full agent population
- Wave 4 (week 4–5): Toll-free numbers and high-visibility lines

**8. Define rollback triggers and procedure.** Before cutover, document the conditions that trigger a rollback (e.g., call failure rate exceeds 5%, average hold time increases by more than 3x, recording failures detected). Maintain the ability to redirect numbers back to the legacy system for a minimum of 2 weeks post-cutover. This typically means keeping the old CCaaS instance active (but idle) during the stabilization period.

**9. Run post-go-live QA.** Check routing outcomes, missed-call behavior, timestamps, transcripts, recording access, and outbound permissions in the first 24 to 72 hours. Compare call volume and abandonment rates against pre-migration baselines.

## Market Traction and Competitive Context

<cite index="50-2,50-3,50-4">Zendesk has landed a 4,000-seat deployment of its CCaaS platform with a global brand in the sports, media, and entertainment industry. The Contact Center has notched more than 100 wins since its May 2025 launch, spanning every continent (except Antarctica).</cite>

<cite index="50-8,50-9">Momentum is growing particularly among organizations in the 250–750-agent range, with Zendesk also securing wins in the strategic 600–1,000-agent tier.</cite>

<cite index="53-4">Zendesk research shows voice still represents 40% of contact center volume and is growing, while 75% of contact center leaders say legacy technology prevents them from delivering a true omnichannel experience.</cite>

The competitive dynamics are worth watching. <cite index="6-1,6-2,6-3">Salesforce recently followed Zendesk's lead by launching its Agentforce Contact Center. CCaaS providers such as Genesys and NICE are moving in the opposite direction, offering native case management. Zendesk believes the breadth of its platform — analytics, WFM, and QA alongside CRM — gives it an edge.</cite>

This isn't a one-way street. <cite index="11-24">Zendesk will still offer close integrations with NICE, Genesys, Five9, or other prominent alternatives.</cite> The question isn't whether you *can* keep your existing CCaaS — it's whether the operational overhead of maintaining two platforms is worth it. For a 250-agent operation, the integration maintenance cost (CTI connector updates, dual-system status sync, separate WFM feeds, fragmented reporting) typically runs $30,000–$75,000/year in direct costs plus ongoing engineering time — a concrete number to weigh against migration cost and risk.

## What to Evaluate Before You Move

Before committing to a migration, pressure-test these areas:

**1. Agent count and complexity threshold.** If you're running fewer than 50 agents with straightforward call flows, Talk is sufficient. Contact Center's value scales with complexity — multi-queue, multi-language, WFM-dependent operations. The 250-agent mark is where the consolidation economics typically become compelling.

**2. Existing Amazon Connect investment.** If you already run Amazon Connect, Zendesk Contact Center may reuse parts of your infrastructure (Lambda functions, Lex bots, existing contact flows). If you're on Genesys or NICE, expect a ground-up rebuild of call flows and routing logic.

**3. Feature gap tolerance.** Review the feature parity table above. If your operation depends on predictive outbound dialing, screen recording, real-time PCI masking, or journey orchestration, Zendesk Contact Center does not currently offer these. Determine whether the gap is acceptable, can be supplemented, or is a deal-breaker.

**4. Compliance requirements.** Verify that call recording storage, encryption, and retention policies meet your specific regulatory obligations (PCI DSS, HIPAA, GDPR, MiFID II for financial services). AWS post-call redaction has documented limitations — validate the exact behavior against your requirements with your compliance team, not just your IT team.

**5. Total cost of ownership.** Map out the full cost: Suite plan + Contact Center add-on ($83/agent/month) + Minutes Blocks + AWS consumption + WFM/QA add-ons if needed. Compare this against your current CCaaS contract *including* the cost of maintaining the Zendesk integration layer, CTI connector licensing, and dual-system engineering overhead. Factor in the [hidden costs of a platform change](https://clonepartner.com/blog/blog/7-hidden-costs-of-switching-from-zendesk-in-2026/) as well.

**6. Migration timeline and risk.** IVR reconstruction, number porting, agent retraining, and reporting gaps all need to be scoped. Use the timeline benchmarks above to set realistic expectations. A 500-agent operation should plan for 14–26 weeks, not 4.

### Should You Move Now?

**Move now** if you're already on legacy Engage, if your current CCaaS creates real data and workflow separation from Zendesk, or if Talk is too small for your routing and governance needs.

**Wait or pilot** if Talk already fits your scale, if you depend on features Zendesk Contact Center hasn't shipped yet (predictive dialing, screen recording, real-time masking), or if you haven't rationalized your statuses, flows, and recording policies.

**Watch and re-evaluate in 6 months** if you're mid-contract with a competitive CCaaS and the integration overhead is manageable. Zendesk's feature roadmap will continue closing gaps, and being six months later on a stable migration beats being six months early on a disrupted one.

The honest read: Zendesk Contact Center is a serious product launch, not a rename. It is also a different architecture from Talk. Teams that treat that difference seriously will have a much cleaner migration.

## When to Bring in a Migration Partner

The technical surface area of a CCaaS migration is larger than a typical help desk data move. You're dealing with real-time voice infrastructure, carrier-level number porting, IVR logic that took years to tune, and compliance-sensitive call recordings that can't be lost.

If you're planning a move to Zendesk Contact Center — whether from Talk, from a third-party CCaaS, or from the legacy ZAF app — and you need engineers who understand both the Zendesk platform and the Amazon Connect infrastructure underneath it, [schedule a 30-minute scoping call](https://cal.com/clonepartner/meet?duration=30).

## Frequently asked questions

### What is the difference between Zendesk Talk and Zendesk Contact Center?

Zendesk Talk (Voice) is basic call handling bundled into Suite plans, backed by Twilio — suitable for SMB teams with simple routing. Zendesk Contact Center is a separate enterprise-grade CCaaS add-on built on Amazon Connect, offering multi-queue routing, advanced IVR, WFM integration, PII/PCI redaction, and enterprise call recording. Contact Center starts at $83/agent/month on top of your Suite plan.

### Is there a forced migration from Zendesk Talk to Zendesk Contact Center?

No. Zendesk Talk continues to function and there is no mandatory migration. However, Zendesk has signaled that future feature development — including native monitor and barge — will prioritize Contact Center over Talk, so teams with growing voice complexity should evaluate the upgrade path.

### What does Zendesk Un-CCaaS mean?

Un-CCaaS is Zendesk's marketing positioning for its Contact Center product, modeled after T-Mobile's Un-carrier branding. It frames Zendesk as a simplified, CRM-native alternative to traditional CCaaS platforms like Genesys and NICE, emphasizing that your contact center should live where your customer data already exists.

### How much does Zendesk Contact Center cost per agent?

Zendesk Contact Center starts at $83/agent/month as an add-on to any Suite plan. It includes one Minutes Block per agent per month, with additional blocks and AWS telephony usage billed on consumption. A team on Suite Professional pays at least $198/agent/month ($115 + $83) before AI or WFM add-ons.

### What needs to be migrated when moving from a third-party CCaaS to Zendesk Contact Center?

You need to rebuild IVR trees and call flows in Amazon Connect's flow editor, rationalize agent statuses to fit the 50-status AWS cap, recreate routing rules and queue configurations, migrate historical call recordings with metadata for compliance, set redaction and compliance policy, and redirect any CTI connectors, WFM feeds, or QA integrations to the new stack.