CCaaS migration to Salesforce without the go-live weekend nobody sleeps through
Moving a contact center from Genesys, Five9, NICE CXone, Avaya or a standalone Amazon Connect onto Salesforce is mostly a routing, IVR and change problem, not a software problem. I plan it as waves by queue, keep reporting whole, and cut over when the numbers say so.
What a CCaaS migration to Salesforce involves
A CCaaS migration to Salesforce replaces or integrates your contact center platform so that calls, messaging and cases are routed, handled and reported in Service Cloud. The telephony lands on Service Cloud Voice, either bundled Amazon Connect or your existing platform through partner telephony, or on Agentforce Contact Center, Salesforce's newer native option. The agent desktop becomes the Service Console. IVRs, queues, skills, recordings and reports are redesigned rather than copied, and menu trees are increasingly replaced by Agentforce Voice agents.
The two big architectural choices are whether to keep the existing CCaaS as the telephony layer under Salesforce or to retire it, and whether to rebuild the IVR as a shorter, intent-first flow or replace it with an Agentforce Voice agent that resolves calls instead of routing them. Everything else, from number porting to agent training, follows from those two.
Platforms I migrate from, or keep underneath
| Coming from | Typical telephony decision | What usually drives the timeline |
|---|---|---|
| Genesys Cloud / Engage | Keep as partner telephony first, consolidate later | Contract dates, IVR depth, WFM integration |
| Five9 | Depends on outbound dialer needs | Campaign and dialer parity |
| NICE CXone | Partner telephony first; Amazon Connect or Agentforce Contact Center to consolidate | Workforce management and quality parity |
| Avaya / Cisco on-premise | Amazon Connect or Agentforce Contact Center | Carrier porting, recording retention, compliance |
| Standalone Amazon Connect | Attach Service Cloud Voice to the instance you own | Routing redesign; smoothest path of the five |
The decision column is the usual starting point, not a rule. Contract dates and what the current platform does well decide it in practice.
When a migration is the wrong first move
If the service team is not on Salesforce yet, or the current CCaaS contract has years left and the platform works, a full migration is not where to start. Fixing routing, IVR length and the agent desktop in place usually pays back sooner, and it makes the eventual move smaller. I will say which situation you are in during the first call.
What I deliver
- Assessment. Inventory of queues, skills, IVR branches, integrations, recordings, reports, WFM and compliance rules, with volumes behind each.
- Target architecture. Service Cloud Voice model, Omni-Channel routing design, data and identity flow, reporting parity, and what gets retired when.
- IVR replacement and self-service redesign. Intent-first flows that identify the customer and the reason, and an Agentforce Voice agent in place of the menu tree wherever it pays back.
- Build and integration. Routing, console, CTI, recording, WFM and quality integrations, dashboards.
- Wave cutover. Migration by queue, parallel running, number porting coordinated with carriers, training, hypercare and a measured handoff.
- Decommission, when consolidating. Contract exits and shutting down the old platform without losing recordings or history you are required to keep. When the platform stays underneath as partner telephony, this step does not apply.
How the engagement runs
Typical results
| Metric | Before: on-premise | After: cloud + AI |
|---|---|---|
| Customer satisfaction | 65% | 85%+ |
| Average handle time | 9.2 min | 5.5 min |
| First-contact resolution | 58% | 75%+ |
| Systems on the agent desktop | 7+ | One console + AI |
Typical results across 30+ enterprise programs, comparing operations before transformation with cloud plus AI after. Migrations usually show the desktop effect first, one console instead of seven systems, then handle time as routing and screen pop settle in. Every engagement sets its own targets up front.
Where migrations go wrong
- One big-bang cutover. Waves by queue let you fix routing on a small group before the whole floor is on the new desktop.
- Reporting discovered after go-live. Supervisors need the same numbers on day one. Parity is a workstream, not a cleanup.
- Call history nobody can find. Decide early what is migrated, what is archived and what is retained for compliance.
- Training on screenshots. Agents need the real console with real queues in a sandbox weeks before their wave.
Frequently asked questions
Should we keep our CCaaS as the telephony layer or replace it?
If your contract has time left and the platform's telephony is solid, keeping it as partner telephony under Service Cloud Voice gets agents into one console sooner. If the contract is ending or you want one vendor, Service Cloud Voice with Amazon Connect removes a platform. I model both with your contract dates and volumes before recommending.
How long does a CCaaS migration to Salesforce take?
A single-site migration with a modest IVR typically runs three to six months from assessment to final wave. Multi-site, multi-carrier programs with compliance recording run longer, and the wave plan is built around porting lead times and seasonal peaks.
Do we lose call recordings and reports from the old system?
Not if it is planned. Recordings are archived or retained per your compliance rules, historical reports are preserved as exports or in a warehouse, and new reporting is built to match the metrics supervisors already use before the first wave.
Can we migrate in phases?
Yes, and you should. I migrate by queue or business unit, run the old and new platforms in parallel for each wave, and move the next group only when the agreed KPIs hold.
What happens to our IVR?
It is redesigned or replaced, not copied. Most legacy IVRs are longer than they need to be. Where a voice agent pays back, the Agentforce Voice agent takes the call, identifies the caller, answers questions, takes action in your systems and resolves most calls without a human, routing the rest with context. Where it does not yet, a short intent-first flow does the identification and routing.
Go deeper
Let's talk for 30 minutes.
Tell me where your contact center or Salesforce is stuck. If I can help, we'll map out what's possible. If I can't, I'll tell you that too. It's free either way.
Book a call