Cloud Contact Center Migration Guide

Moving from on-premise contact center infrastructure — Avaya, Cisco, Genesys Engage, Mitel — to a cloud CCaaS platform is one of the highest-risk IT projects an operations team will run. The risk is not technical. The risk is that customer-facing queues go dark during the cutover, or that the new platform cannot access the CRM data agents need on day one. This guide is a practical playbook that addresses both.

Average migration timeline for 50–150 agents: 6–8 weeks when all pre-work is complete before week one begins. Migrations that skip pre-work stretch to 12–16 weeks.

Before You Start: Pre-Migration Checklist

Do not set a go-live date until you have completed the following.

Inventory your current stack. List every integration connected to the on-premise system: CRM (Salesforce, HubSpot, ServiceNow, Zendesk), helpdesk, WFM tool, dialer, quality management platform, call recording storage, custom reports. Each integration needs to be replicated or replaced on the new platform.

Document your routing logic. On-premise IVR and ACD routing rules are often undocumented and have accumulated years of one-off changes. Interview your current platform admin to map every queue, skill group, and routing condition. Export whatever configuration exports the platform supports. This is your blueprint for rebuilding routing on the new cloud platform.

Baseline your KPIs. Pull 90-day rolling averages for AHT, FCR, ASA, Service Level (calls answered within your target threshold), CSAT, agent occupancy, and abandonment rate. Do this per queue, not just overall. These numbers are your recovery target: if post-migration KPIs are significantly worse after 30 days, you can identify exactly where the deviation started.

Confirm compliance requirements. If you handle PHI (HIPAA) or payment card data (PCI-DSS): confirm the new platform signs a Business Associate Agreement covering AI features, and verify that pause-and-resume recording during payment capture is supported. Do this before vendor selection, not after contract signature.

Get number porting in motion early. Phone number porting from the current carrier takes 2–4 weeks. Submit your Loa (Letter of Authorization) and porting request as early as possible — porting delay is the most common cause of migration schedule slip.

The 6-Week Migration Plan

Week 1: Platform provisioning and routing rebuild

• Provision the new cloud CCaaS tenant. Set up SSO, user directory sync, and admin roles.
• Import agents, supervisor structure, and skill groups.
• Rebuild IVR and ACD routing logic using your pre-migration documentation.
• Configure voicemail, call recording, and quality management settings.
• Do not connect to production telephony yet. This week is configuration-only in a test tenant.

Week 2: CRM and helpdesk integration

This is the highest-risk week. The CRM integration must be production-ready — not in a sandbox with test data — before agents train on the platform.

• Install the native CRM connector (Salesforce, HubSpot, ServiceNow, Zendesk are native on DialPhone and Genesys; Talkdesk and NICE also have certified connectors for these four).
• Map CRM fields to the agent desktop screen-pop. Test with real customer records, not dummy data.
• Configure case creation and disposition sync. When an agent wraps a call, does a case auto-create in the CRM? Does the disposition code write back correctly?
• Run end-to-end test scenarios: inbound call from known customer, screen-pop fires, agent handles call, wrap-up creates CRM case. Do this 50 times with varied scenarios before moving to week 3.

Week 3: WFM and quality management handoff

If your team uses a workforce management tool — Verint, Calabrio, Aspect, NICE WFM, or a built-in platform WFM — this week handles the handoff.

• Export historical volume data from the old WFM tool (at least 13 months to capture seasonality).
• Import into the new WFM system or configure the cloud platform’s built-in WFM.
• Rebuild schedules, shift configurations, and time-off policies.
• Configure quality management: call scoring forms, sampling rules, supervisor queues.
• For platforms with built-in WFM (DialPhone, Genesys mid-tier), this handoff is straightforward. For platforms that require a third-party WFM integration (Five9, RingCX, Talkdesk), plan an additional 1–2 weeks for the WFM vendor integration.

Week 4: Agent training

Run training in cohorts of 10–15. Do not attempt to train all agents in one session.

Training structure (8 hours per agent):

• Module 1 (2 hours): Softphone basics, status management, call handling controls.
• Module 2 (2 hours): CRM screen-pop, case creation, wrap-up disposition, notes.
• Module 3 (2 hours): Digital channel handling (email, chat, SMS queues if applicable).
• Module 4 (2 hours): Live shadowing with a trained agent before going solo.

Supervisor and admin training (16–24 hours):

• Routing rule management: how to modify queues, skills, and thresholds without IT involvement.
• Real-time dashboard and wallboard configuration.
• WFM: forecast review, schedule publication, intraday adjustment.
• Quality management: scorecard configuration, calibration process.

Training completion must reach 90% before setting a go-live date. Track completion per agent.

Week 5: Parallel operation and final testing

• Activate phone number porting for your primary DID blocks (porting should have been submitted in week 1).
• Run the cloud platform in parallel with the on-premise system: route a subset of inbound volume (typically 10–20%) to the new platform using a split number or a test DID.
• Monitor KPIs on the new platform: ASA, AHT, CSAT, escalation rate. Compare against your baseline.
• Identify and fix any routing gaps or CRM integration issues surfaced by live traffic.
• Conduct a go/no-go review at the end of week 5. Criteria: porting complete, CRM integration passing QA, training above 90%, KPIs within acceptable variance of baseline.

Week 6: Cutover and stabilization

Cutover timing: Choose a Tuesday or Wednesday morning at your lowest weekly volume point. Avoid Mondays (volume spikes), Fridays (reduced staffing), and month-end periods.

Cutover sequence:

1. Final backup of on-premise configuration and call recordings.
2. Route 100% of new inbound contacts to the cloud platform.
3. On-premise system remains in “drain” mode: handles only contacts already in queue, no new routing.
4. Once on-premise queue is empty (typically 2–4 hours), freeze the on-premise system.
5. On-premise system remains powered but in standby for 2 weeks (rollback window).

Stabilization (days 7–14 post-cutover):

• Daily KPI review versus baseline. Expect AHT to be 5–15% above baseline for the first week as agents adjust.
• Supervisor check-ins with each agent cohort daily.
• Routing adjustment as real-traffic patterns emerge (lunch peaks, after-hours patterns may differ from historical data).

Rollback Plan

Define your rollback trigger before go-live: specific KPI thresholds that, if breached on day 1–3, trigger a reversion to the on-premise system. Common thresholds: service level below 60% for more than 2 consecutive hours, CSAT below 3.0 per day, CRM integration failure rate above 5%.

Rollback procedure:

1. Re-route 100% of inbound traffic back to on-premise DIDs (requires porting coordination with your telephony carrier — confirm this is possible before cutover).
2. Notify agents to switch back to old desktop.
3. Document what broke and why before re-attempting migration.

Rollback is rarely needed when weeks 1–5 are executed correctly. It is almost always needed when teams skip the parallel operation week or cut corners on CRM integration testing.

Integration Cutover Detail: CRM and Helpdesk

The three most common CRM integrations and their migration considerations:

Salesforce. Native CTI connectors exist for DialPhone, Genesys, Talkdesk, NICE, and Five9. The migration risk is Salesforce custom objects that were mapped to the old CTI connector — these need to be remapped. Budget 2–5 days for a Salesforce admin to remap custom fields. Test case creation, screen-pop, and activity logging end-to-end.

HubSpot. HubSpot’s native calling integrations are generally lighter than Salesforce. The migration risk is losing historical call activity from the old platform’s HubSpot sync. Export call logs before migration and retain them for 12 months as backup.

Zendesk. Zendesk’s CTI integration supports ticket creation, screen-pop, and status sync. Migration risk: ticket routing rules configured in the old CCaaS that mapped to Zendesk groups need to be rebuilt as cloud platform routing logic. Map these during week 1, not week 4.

Cost Categories for a 100-Agent Migration

Category	Low end	High end
Professional services (vendor)	$0 (self-serve)	$45,000
Number porting (100 DIDs × $20)	$2,000	$2,000
CRM integration custom dev	$0 (native connector)	$15,000
WFM data migration and rebuild	$2,000	$10,000
Agent training (8 hrs × 100 agents × $25/hr)	$20,000	$20,000
Parallel operation overlap (2 weeks, both platforms)	$3,000	$8,000
Total one-time migration cost	$27,000	$100,000

The low end is achievable when: the new platform supports self-serve onboarding (no mandatory professional services), CRM integrations are native connectors (no custom dev), and WFM can import data from the old tool in a standard format. DialPhone’s mid-market deployments under 150 agents typically land in the $30,000–$50,000 one-time range.

KPI Baselines and Recovery Targets

KPI	What to baseline	Expected post-migration dip	Recovery window
AHT	90-day per-queue average	+5–15%	2–3 weeks
FCR	90-day rate	-3–8%	3–4 weeks
Service Level	% calls answered in target time	-5–15 percentage points	1–2 weeks
CSAT	90-day average	-0.2–0.5 points	3–4 weeks
Abandonment Rate	90-day rate	+2–5%	1–2 weeks

Post-migration KPIs that do not recover to baseline within 4 weeks indicate a systemic issue — usually routing misconfiguration or CRM integration gaps — not normal adjustment.

Why 40% of migrations fail before go-live: stakeholder alignment

Technical execution accounts for roughly 60% of migration risk. The other 40% is people — specifically, agents, supervisors, and business stakeholders who were not aligned before kickoff.

The most common failure modes before go-live:

Routing logic undocumented. The person who built the IVR routing in 2018 has left. No one can explain why certain calls route to a specific queue. When the migration team goes to rebuild routing in the new platform, they discover undocumented edge cases that only surface in production. Prevention: allocate two weeks before kickoff to interview every supervisor about current routing logic, record escalation paths, and map every queue to its business purpose.

Agent resistance to desktop change. Agents who have used the same phone system for five years develop muscle memory. A new interface — even a simpler one — triggers anxiety and resistance that shows up as prolonged onboarding and increased escalation rates in the first two weeks. Prevention: involve three to five high-influence agents as “migration champions” early in the process. Their peer endorsement reduces resistance more than any management communication.

Stakeholder misalignment on go-live criteria. The CTO defines success as “the system works.” The VP of Customer Experience defines success as “CSAT does not drop.” The finance team defines success as “we stop paying for the old system.” Without a single written go/no-go criteria document signed before week 5, the go-live decision becomes a political negotiation. Prevention: define numeric go-live criteria in week 1 and attach them to the project plan.

Compliance sign-off not on the critical path. Healthcare contact centers that need HIPAA compliance review, or financial contact centers that need PCI-DSS scope confirmation, often treat compliance as a post-go-live activity. If the compliance team discovers a gap on go-live day, the migration halts regardless of technical readiness. Prevention: schedule compliance sign-off for week 4 at the latest — before the parallel operation begins.

On-premise vs cloud: 6-category comparison

Buyers mid-contract on Avaya, Cisco, or Genesys Engage often ask whether the cost of staying justifies deferring migration. This framework answers that:

Category	On-premise CCaaS	Cloud CCaaS
Cost structure	CapEx (hardware) + OpEx (maintenance, license)	OpEx only (per-seat SaaS)
Scalability	Capacity planning 6–12 months in advance	Add seats in under 1 minute
Reliability	Depends on your data center; redundancy is CapEx	99.99%–99.999% SLA, geo-redundant
AI features	Middleware bolt-on, 12–24 month lag	Native, continuous deployment
Upgrade cycle	Major version every 2–4 years, disruptive	Continuous, no scheduled downtime
Compliance tooling	Manual audit log review, custom builds	Native audit logging, automated reporting

The hidden cost of staying on-premise that most TCO analyses miss: the IT labor cost of maintaining and troubleshooting the on-premise system. A contact center running 100 agents on on-premise Avaya typically requires one to two full-time IT staff dedicated to platform support. At $80,000–$120,000 per FTE, that is $80,000–$240,000/year in maintenance labor that disappears when you move to a self-serve cloud platform. Factor this into your migration ROI calculation.

Migration day, hour by hour

The go-live day is the highest-anxiety moment in the project. A detailed timeline removes uncertainty.

Time	Action	Owner	Rollback trigger
6:00 AM	Final configuration snapshot of old platform	Platform admin	—
7:00 AM	Confirm all agents logged into new platform desktop	Supervisor	>10% login failures → delay
7:30 AM	Route 10% of inbound volume to new platform	Telephony admin	AHT >150% of baseline → pause
8:30 AM	Review 1-hour KPIs: ASA, AHT, CSAT, errors	Migration lead	Service level below 50% → rollback
9:30 AM	Route 50% of volume to new platform	Telephony admin	—
11:00 AM	Review 2-hour KPIs	Migration lead	CSAT below 3.0 → rollback
12:00 PM	Route 100% of new contacts to new platform	Telephony admin	—
2:00 PM	On-premise system enters drain mode (no new contacts)	Platform admin	—
4:00 PM	On-premise queue empty — freeze old system	Platform admin	Keep powered in standby
5:00 PM	End-of-day KPI review vs baseline	Migration lead	Document all variances

Rollback procedure (if triggered): re-route 100% of inbound traffic back to on-premise DIDs via the telephony carrier, notify agents to return to the old desktop, document what triggered rollback, and convene a post-mortem within 24 hours before scheduling a retry date.

Top 5 migration failure causes and mitigations

Failure cause	Frequency	Mitigation
CRM integration breaks on go-live day	Very common	Test 50 real-case scenarios in week 2; retest in week 5 with production data
Routing logic gaps discovered in production	Common	Full routing documentation before week 1; agent UAT in week 4
Agent training below 90% completion	Common	Track completion per agent; block go-live if below threshold
Number porting delay	Common	Submit LOA in week 1; confirm FOC date before scheduling go-live
Compliance gap discovered post-go-live	Occasional	Complete compliance sign-off checklist in week 4, not week 6

How We Tested

DialPhone re-verifies every comparison in this guide every 90 days. We pull pricing directly from each vendor’s public pricing page on the dates listed in the frontmatter (lastVerifiedAt or updatedAt). Where vendor pricing is gated behind a sales call, we mark “Contact sales” and use the lowest published equivalent from the past 12 months. Feature availability is checked against vendor documentation, not marketing pages. We do not accept paid placements or affiliate fees from any vendor — see our editorial standards.

What We Don’t Like

No platform is perfect, including DialPhone. Honest drawbacks based on user feedback and our own testing:

• Smaller integration catalog than RingCentral (~40 vs 200+). Niche vertical CRM integrations may require API work.
• Newer brand awareness. RingCentral and 8x8 have 15+ years of analyst coverage. Enterprise procurement reviews may take longer.
• Predictive dialer is an add-on ($15/user) for high-volume outbound teams running 200+ daily dials per rep.
• HIPAA BAA starts on Advanced tier ($34/user), not the $24 Core plan. Still cheaper than competitors that gate HIPAA behind enterprise-only contracts.

Related Resources

• Best AI contact center platforms 2026
• DialPhone AI Contact Center overview
• AI contact center pricing by tier
• Workforce management for contact centers
• VoIP pricing research hub
• Contact center pricing calculator
• What is UCaaS

FAQ: cloud contact center migration

How long does a cloud contact center migration take? +

Most 50–150 agent migrations complete in 6–10 weeks when the project scope is defined before kickoff. Common causes of delay: CRM integration requiring custom API work (adds 2–3 weeks), phone number porting windows set by the losing carrier (typically 2–4 weeks from submission), and agent training completion below 90% on the go-live target date. Migrations for 200–500 agents typically run 12–20 weeks. Genesys publishes an average enterprise deployment window of 90 days; DialPhone's average mid-market deployment is 35 days.

What does a cloud contact center migration cost beyond the per-seat license? +

Plan for four cost categories outside the recurring seat price: (1) Professional services — $25,000–$80,000 for enterprise deployments, lower or zero for self-serve mid-market platforms; (2) Number porting — carriers typically charge $15–$25 per DID ported, plus a 2–4 week porting window; (3) Integration development — CRM integrations that require custom API mapping add $5,000–$20,000 depending on complexity; (4) Training — budget 8–12 hours per agent for platform training, plus supervisor and admin training.

Total one-time migration cost for a 100-agent team typically runs $40,000–$120,000.

Can we run the old system and new cloud platform in parallel during migration? +

Yes — running parallel systems for 2–4 weeks is the standard low-risk approach. All new contacts route to the cloud platform; the on-premise system handles only active open cases transferred from the old queue. Parallel operation adds some cost (you pay for both platforms briefly) but eliminates the risk of a hard cutover failing during a high-volume period. The cutover window should be a low-volume period, typically a Tuesday or Wednesday morning.

What KPIs should we baseline before migration? +

Establish 90-day rolling averages for: Average Handle Time (AHT), First Contact Resolution (FCR), Average Speed of Answer (ASA), Service Level (calls answered within target threshold), CSAT score, Agent Occupancy, and Abandonment Rate. Capture these at the queue level, not just overall, so you can detect if migration affected specific queues. Give the new platform 30 days post-cutover before comparing KPIs — the first two weeks often show temporary dips as agents adjust to the new desktop.

What is the biggest risk in a contact center cloud migration? +

The most common failure point is the CRM integration cutover. If the new cloud platform cannot read the full customer history from the CRM on the first day, agents work blind, AHT spikes, and CSAT drops. Run the CRM integration in parallel with agent training, not after. Test the full customer journey — from IVR to agent desktop to CRM record — with real production-representative test cases before go-live.

Do agents need extensive retraining after a cloud migration? +

Cloud CCaaS platforms are generally simpler to use than on-premise systems — the agent desktop is browser-based, softphone controls are standardized, and the CRM data appears in a side panel rather than a separate screen. Most agents need 4–8 hours of product training plus 1–2 live shadowing sessions. The bigger training requirement is for supervisors and WFM staff who need to reconfigure routing rules, forecasts, and schedules in the new system. Budget 16–24 hours for supervisor/admin training.

How should we handle HIPAA or PCI compliance during migration? +

Compliance continuity is a parallel track, not an afterthought. Before migration kickoff: (1) Confirm the target platform signs a Business Associate Agreement (if HIPAA-covered) and covers all AI features — transcription, agent assist, and summarization — within the BAA scope; (2) Verify PCI-DSS pause-and-resume recording is supported on the new platform; (3) Establish a compliance sign-off checklist that must be completed before go-live. DialPhone signs BAA at all CCaaS tiers without requiring an enterprise contract.