Customer service has moved far beyond phone calls. Customers now use email, messaging, self-service portals and other digital channels, but voice still matters when an issue is urgent, complicated or simply easier to discuss with another person.
For service teams, the challenge is not only answering calls. They also need those conversations to connect with the customer information, cases and processes already managed in Salesforce.
This is where Salesforce Voice, formerly known as Salesforce Service Cloud Voice, comes in. It brings telephony into the Salesforce service environment so agents can handle calls together with customer data, routing and service workflows.
But Salesforce Voice is more than a phone inside Salesforce. The way it works depends heavily on the telephony model, routing setup and level of automation an organisation chooses.
How Salesforce Voice differs from traditional CTI
Before Service Cloud Voice, many organisations connected their phone systems to Salesforce through CTI integrations.
This allowed agents to make and receive calls from Salesforce, but the CRM and telephony platform often remained largely separate. The phone controls appeared inside Salesforce, while much of the routing, reporting and call handling still happened elsewhere.
Salesforce Voice brings these areas closer together. Calls can become part of the Salesforce service process and work alongside customer records, cases, routing, transcription and automation.
This can also change how teams think about their contact centre. Moving from a traditional CTI integration to Salesforce Voice is not only about replacing the softphone. It may also involve reviewing call flows, routing, data, agent workflows and the way voice fits alongside other service channels.
If you are looking at the broader platform rather than Voice alone, our guide to Salesforce implementation explains how architecture, integrations, data and user processes fit together in a wider Salesforce project.
Choosing the right telephony model
One of the first decisions is what to do with the existing telephony setup.
Some organisations are building a new contact centre. Others already use Amazon Connect or another provider and want to keep it.
Salesforce supports several approaches.
Salesforce Voice provides a Salesforce-native telephony option. This can suit organisations that want Salesforce to manage more of the voice environment and keep the setup closely connected with the rest of the platform.
Salesforce Voice with Amazon Connect uses Amazon Connect for telephony while Salesforce handles much of the provisioning and setup.
Organisations that already manage their own AWS environment can use Partner Telephony from Amazon Connect. This gives them more control over Amazon Connect and the underlying AWS configuration.
There is also Partner Telephony, which allows companies to connect supported third-party telephony providers instead of moving their contact centre to Salesforce or Amazon Connect.
There is no single option that works best for everyone. The right model depends on the existing technology, required features and who should own the telephony environment after launch.
Routing calls with Salesforce Omni-Channel
Connecting telephony is only part of the setup. The next question is how calls reach the right person.
Salesforce Omni-Channel can coordinate voice with other service work, including messaging and cases. This gives the contact centre a shared view of agent availability and capacity instead of treating phone calls as a completely separate workload.
Omni-Channel flows can also help control routing. For example, calls can move to particular queues or agents based on skills and other business rules.
The system can also open relevant Salesforce records when an agent accepts a call. Instead of starting the conversation and then searching for the customer or case, the agent can already have the right information in front of them.
The exact routing model still depends on the telephony provider. Some organisations manage most routing in Salesforce, while others keep part of it within their contact-centre platform.
Where AI fits into Salesforce Voice
Voice is also becoming more closely connected with Salesforce AI.
Real-time transcription can turn a conversation into text while the call is happening. This gives agents another way to follow the conversation and creates information that can support later analysis.
Einstein Conversation Insights can analyse recorded conversations and help supervisors identify important moments, recurring topics and coaching opportunities.
Agentforce Voice takes this further. Voice-enabled agents can understand spoken requests, work with Salesforce data and carry out supported actions. Depending on the use case, they can handle suitable parts of a customer conversation and transfer the interaction to a human agent when needed.
These capabilities solve different problems.
Transcription captures what was said. Conversation intelligence helps teams analyse calls. Agentforce can take action during the conversation.
The important part is deciding where AI genuinely helps. Not every customer interaction needs automation, especially when the issue is complex or sensitive.
What to plan before implementation
A Salesforce Voice project touches more than telephony, so planning should start with the wider service process.
First, review the existing contact centre. Look at phone numbers, IVR, queues, routing, integrations and reporting. Understanding what already works makes it easier to decide what actually needs to change.
Then look at real customer journeys. What happens when someone calls? How are transfers handled? What happens if no agent is available? How should callbacks work?
Technical dependencies matter as well. Salesforce permissions, automation, integrations, data quality and network readiness can all affect the implementation.
A structured Salesforce audit can help identify issues in the current setup before Voice is added on top of it.
Ownership after launch should also be clear. An Amazon Connect setup may require AWS knowledge, while partner telephony brings its own provider-specific responsibilities.
Finally, test the scenarios agents deal with every day, not only the ideal call flow. Routing failures, transfers, unavailable agents and fallback processes should all be tested before go-live.
What about cost?
Salesforce Voice costs depend on the telephony model, licences and actual usage.
Amazon Connect setups can include different types of usage. For example, Salesforce Voice platform minutes and telephony minutes are tracked separately, and telephony usage can generate additional charges.
For this reason, the licence price alone does not give a complete picture of the cost of running the contact centre.
Before budgeting, it is worth estimating expected call volumes, inbound and outbound usage, telephony requirements and any additional AI capabilities the organisation plans to use.
Final thoughts
Salesforce Voice can make voice a much more connected part of the customer service process.
Agents can work with calls, customer data, cases and routing in the same environment, while transcription, conversation intelligence and Agentforce add further options for automation and AI.
The main challenge is choosing the right setup. A Salesforce-native model, Amazon Connect and partner telephony can all make sense in different situations.
For teams already running customer service on Salesforce, Voice can be a logical next step when phone interactions need to work more closely with the rest of the service platform.
The technology is only one part of it. A successful implementation also depends on routing, data, integrations, processes and clear ownership after launch.
If you want to look at the wider Salesforce setup, you can also read our guides to Salesforce managed services and Salesforce implementation.
The original, more detailed version of this article is available on the Spyrosoft Salesforce blog.
Top comments (0)