Florida Prepaid’s Call Center Never Closes Thanks to Agentforce Voice

For many families, a Florida Prepaid account is one of the longest financial relationships they’ll ever have. Parents open an account when a child is young, then rely on Florida Prepaid for accurate information, secure transactions, dependable service, and financial confidence for years before those education funds are used. That long-term relationship makes every service interaction matter. A question about enrollment, eligibility, payments, scholarships, or plan usage is not just a call-center transaction. It’s a moment that can either strengthen or weaken trust.

Florida Prepaid saw an opportunity to reimagine the phone channel around that responsibility. The goal was not simply to answer more calls faster, but to make sure each customer reached the right resource at the right time — digital when speed and consistency mattered, human when judgment, empathy, or guidance mattered most. This year, Florida Prepaid turned to Agentforce Voice to scale up. Here’s what happened next.

Already built to move fast

Florida Prepaid did not start from scratch. By the time it moved to Agentforce Voice, the organization already had Agentforce running in chat, a strong Salesforce data foundation, Salesforce Voice, Amazon Connect, Service Cloud, Salesforce Knowledge, and connected service workflows. Agentforce Voice was not added as a standalone call-center tool, but integrated into a broader service model.

When a caller contacts Florida Prepaid, the call routes through Amazon Connect into Agentforce Voice. The agent uses approved Salesforce Knowledge Articles to identify intent and answer routine questions about enrollment, eligibility, and plan usage. The agent works only within the bounds of public knowledge — no account balances, no case history, no personal data. If a call involves a high-touch or sensitive issue, such as cancellation, the agent transfers the caller to a live representative with an AI-generated summary for context.

“We did not approach Agentforce Voice as a call-deflection tool. We approached it as a way to redesign the service model responsibly, using AI for routine questions, preserving human capacity for higher-value conversations and building the guardrails needed to scale with trust,” said Ashley Falls, chief information technology officer for Florida Prepaid.

Agentforce Voice also supports after-hours service. When representatives are unavailable, it can answer routine questions or create a Salesforce case for next-day follow-up, giving callers a path forward without reducing human accountability.

Scaling with confidence: From 10% to 75%

Florida Prepaid started with a narrow, measured rollout, allowing Agentforce Voice to handle 10% of calls. Within weeks, confidence grew, performance improved, and the team increased volume to 30%. After additional tuning and enhancements, Agentforce Voice moved to 50% of incoming calls and is now the first point of contact for 75% of business-hours calls.

That growth reflects more than call deflection. It shows a shift in how the contact center operates. Routine questions can be handled faster and more consistently by Agentforce Voice, while representatives focus on the conversations where human judgment, empathy, and personalized guidance matter most.

After hours, Agentforce Voice handles 100% of incoming calls by either resolving routine inquiries or creating a Salesforce case for next-day follow-up. Before Agentforce Voice, those callers had no clear path forward other than calling back during business hours.

For Florida Prepaid, the value is not simply fewer calls reaching representatives. It is better allocation of work across people, digital agents, workflows, data, and knowledge. On the chat channel, Agentforce has already been so effective at containing inquiries that approximately 35% of representatives were redeployed into proactive revenue generating engagement roles, while also creating more capacity for complex conversations, outreach, and education.

The shift is visible in the numbers. AI-handled calls averaged approximately 1.6 minutes, compared with 10.9 minutes for human-handled calls. For calls that still needed escalation, average handle time increased by about two minutes, a positive signal that representatives had more room to focus on issues that warranted deeper attention. 

“Florida Prepaid wanted to maintain a high-touch experience — which meant they were willing to increase average handle time for complex issues that truly matter,” said Kerri Bedrosian, senior project manager at NeuraFlash. “Agentforce Voice handles the common knowledge-based questions, freeing representatives to focus on deep customer understanding.”

What’s next

Florida Prepaid sees Agentforce Voice as a foundation to build toward more personalized, authenticated experiences where agents can help guide customers through additional steps and support more routine service needs. That expansion will depend on the same principles that shaped the initial deployment: start with clearly defined use cases, connect the right data and knowledge, set boundaries around what the agent can do, escalate when judgment or risk requires a person, and continuously monitor performance.

The goal is not to make Agentforce Voice do everything, but to ensure that more customers get the right level of support, at the right time, through a channel they already use, while maintaining the human accountability that protects trust.

Three lessons from Florida Prepaid’s deployment

Lesson 1: Start narrow

Florida Prepaid’s agent launched with a deliberately narrow scope — public knowledge articles only, no account data, no authentication. That wasn’t a compromise. It was a decision that made the deployment possible in a regulated financial environment. Starting narrow allowed the team to go live, prove value, and earn the organizational trust to expand.

Lesson 2: After hours is low-hanging fruit

100% of after-hours calls are now handled by Agentforce Voice. No representatives displaced, no complex handoff scenarios, no sensitive interactions to guard against. Callers who previously hit a dead end now get answers or a case logged for follow-up. After-hours coverage is the lowest-friction place to start: the agent has the channel to itself, the use cases are contained, and the value is immediate.

Lesson 3: Let representatives own the feedback loop

Florida Prepaid’s contact center team didn’t step back when the agent went live — they leaned in. Representatives are reviewing transcripts, flagging gaps, and pushing improvements back into configuration. NeuraFlash built tooling specifically to support this: a quality suite for queuing calls and generating custom deflection reports. The result is an agent that improves with each cycle, guided by the people who understand the customer relationships best. The technology scales, while the judgment that shapes it stays human.