Salesforce Voice: enabling AI capabilities in your contact center

Salesforce has announced the planned retirement of Open CTI, officially scheduled for February 28, 2028, and customers are being pointed at Salesforce Voice for their telephony. The migration also decides how much AI your contact center can run.

From Open CTI to Salesforce Voice: the AI imperative

Open CTI arrived in 2012 as a bridge between Salesforce and external telephony systems, giving you softphone features such as call controls, logging and click-to-dial. It was built for a different era, and its integration of interaction records is limited. What customers expect now is immediacy, personalization and unified interaction data. Salesforce Voice answers that with a single integrated platform where customer data fuels real-time actions and insights instead of arriving as a post-interaction update.

That real-time exchange is what agentic and generative AI functionality such as Salesforce Agentforce needs in order to run inside contact center operations.

Migration status and benefits

Research indicates 55% of Salesforce Open CTI and Voice customers are actively exploring or already underway with their migration. There is time before the retirement deadline, and moving early buys you AI-powered contact center functionality that can shorten resolution times, cut after-call work and make the customer experience more consistent.

Key AI use cases enabled by Salesforce Voice

With AI on top of it, Salesforce Voice supports several changes to how a contact center runs.

1. Real-time compliance

AI monitors agent conversations against company guidelines and compliance documentation, then prompts and guides the agent in the moment. Interactions stay accurate and consistent, and the risk of a compliance breach drops. Salesforce Voice users report 18% higher usage of "call quality assurance" than Open CTI users.

2. Automated knowledge retrieval

AI reads the conversation as it happens and surfaces what the agent needs: an answer to the customer's question, a help article, a policy document. That cuts search time mid-call, improves the customer experience and gets to resolution faster. 76% of respondents using agentic AI report using or implementing "automated knowledge retrieval."

3. Handling routine tasks

AI can take routine requests on its own, things like address changes, refunds and appointment bookings, either asynchronously or in the background while a human agent is on a call. Human agents get freed up for complex, high-value interactions, cost per interaction falls and capacity grows.

Salesforce Voice customers show 28% higher adoption of "end-to-end handling of transactional workflows."

4. Automating post-call tasks

AI can compile call notes, populate CRM fields and assign follow-up actions. That cuts the hidden cost of post-call work along with the human error that comes with it. Salesforce Voice respondents are 81% more likely to have implemented "automated post-call work" and 31% more likely to rank "real-time 'co-pilot' guidance for human agents" as a top outcome.

5. Proactive customer service

AI can watch metrics such as failed transactions and delayed deliveries and spot a problem before the customer reports it. It can then resolve the straightforward cases, assign the complex ones to humans, or notify customers up front. Call volumes come down and the customer experience improves. Salesforce Voice customers show a 28% higher adoption of "proactive customer service."

6. Agentic fraud prevention

AI can analyze live call data and telephony signals to detect fraudulent behavior, which takes the weight off cumbersome security questions. Calls run more efficiently and the risk drops. Enterprises combining Salesforce Voice and Agentforce are 2.6x more likely to adopt agentic fraud detection.