Agentforce 360 Deep Dive: Let’s hear it for Agentforce Voice
When your brand talks, it doesn’t matter if it’s a human or AI. Customers expectations are high when on the phone, and agents need to be faster and conversational. And above all else, avoid transfers! Let’s see how Agentforce Voice does it.
Welcome to the next article in our series on Agentforce 360. Today we’re tackling Voice, where they take AI to the phones. To recap, we’ve addressed the Agentic Divide (the problem), the Agentic Enterprise (the solution), and the platform Salesforce has introduced to get there. We talked about Builder, Script, and my favorite so far, Intelligent Context. Agentforce Voice is the natural next component to deconstruct.
Capabilities like Builder, Script and Intelligent Context design the agents to be fast, accurate and skilled. Agentforce Voice enables those agents to deliver responses via natural communication and stay true to your brand.
Salesforce says customers still prefer to talk to support or customer service on the phone. I don’t personally agree, but perhaps with certain industries this is still very much true. Even so, I do agree that a voice component needs to be inherently built into every customer support AI initiative. Whether it’s 2 people or 2,000 calling your 1-800 number, consistency of experience needs to be there.
On top of that, the expectation of an AI agent on the phone versus a text box changes a little. There is a raised bar for speed (no dead air), instant solutions (no phone trees or transfers), or it lacks conversational flow.
If you’re convinced you need to incorporate voice into your AI agent program, read on to see what it does.
What does Agentforce Voice do?
At its core, Agentforce Voice allows AI agents to participate in real-time phone conversations the same way they already do in chat, but now the interaction happens through spoken dialogue instead of text.
When a customer calls in, the voice agent can answer questions, guide workflows, capture information, and trigger actions inside Salesforce, all while maintaining conversational flow. It’s reasoning in real time, using the same context, logic, and guardrails you’ve configured across Builder, Script, and Intelligent Context.
The agent is doing more than reading from a script or routing calls. It can recognize intent, pull relevant data, follow defined processes, and escalate when needed, all without forcing the caller through traditional phone trees.
Agent Voice, now in GA
Consulting Partner
Managed Services Provider
Voice also introduces an additional layer of experience design. You get the right answer, sure, but it also takes into consideration pace and tone.
Your agent is grounded in the same data, workflows, and brand standards as every other interaction point. If you’re going to implement AI for voice, this is the way to go.
Voice may not be every organization’s first AI use case, but when customers pick up the phone, the experience still needs to reflect the same intelligence as every other channel. Let’s look at how this plays out in a real MedTech scenario.
Agentforce 360 Use Case
Agentforce Voice for a Medical Device Manufacturer: Provider Support and Device Escalation
So far, many of our use cases and examples have focused on convenience. Is AI making it easier to engage with you? But sometimes speedy guidance directly impacts patient care.
Let’s say a hospital calls in to check the status of an implant or device order tied to an upcoming procedure. Time matters, and they need clarity quickly.
With Agentforce Voice, the agent can recognize the urgency of the request, identify the hospital and order context, and retrieve fulfillment details in real time. It can confirm shipment status, expected delivery windows, and whether any delays or substitutions are in play.
If the order is at risk of missing its procedural window, the agent can escalate immediately — routing the call to the right logistics or account team without forcing the hospital staff through phone trees or manual transfers.
In scenarios like this, Voice becomes the fastest way to deliver accurate updates, helping providers plan with confidence and avoid last-minute disruption.
What Comes Next
Agentforce Voice brings agents into one of the most real, high-expectation channels enterprises operate. Conversations happen live, context has to surface instantly, and responses need to balance speed with accuracy.
It’s where everything we’ve covered so far comes together. Builder creates the agent. Script governs behavior. Intelligent Context ensures responses are grounded in the right data. Voice is where all of that gets pressure-tested in real time.
But as agents expand across channels, workflows, and use cases, another question starts to surface — how do teams actually build, customize, and evolve these experiences fast enough to keep up?
Next, we’ll explore how Salesforce is approaching that challenge through Agentforce Vibes and the concept of vibe coding — where natural language becomes the starting point for building agents, workflows, and platform customizations.
