Squeeze didn't buy an off-the-shelf AI tool—they built one. The team behind PEEL Voice breaks down how a QA problem turned into a full-stack AI analytics platform that's reshaping sales coaching, compliance monitoring, and lead conversion.
Listen wherever you get your podcasts
Key takeaways
- PEEL Voice grew out of a QA initiative and a billing dispute—not a top-down AI strategy—proving iterative, problem-first development works.
- The platform is custom-built in-house, allowing Squeeze to swap AI models (including OpenAI's latest) without rebuilding core infrastructure.
- Five pillars drive current value: compliance monitoring, quantifying qualitative data, sales enablement, agent performance coaching, and operational efficiency.
- Customer sentiment is tracked before and after call transfers and correlated with close rates to identify which handoff behaviors actually move the needle.
- PEEL Voice can identify high-performing partner agents by name (e.g., MLS number) so their winning behaviors can be replicated across a sales team.
- The roadmap includes real-time dashboards, CRM-triggered marketing actions based on call outcomes, and third-party call-flow integrations.
- AI named the platform: ChatGPT generated the PEEL acronym—Performance Enhancing Engagement Learning—when prompted with a citrus-themed naming brief.
From QA Workaround to Full-Stack AI Sales Platform
PEEL Voice started as a practical fix: Squeeze needed to review more calls faster for quality assurance. A billing dispute with a client—who questioned which calls were payable—forced the team to mine call transcripts systematically. That moment revealed a much bigger opportunity: if AI could surface compliance flags, why couldn’t it surface everything meaningful happening on a call?
Built entirely in-house over roughly six months, PEEL Voice is a custom AI-powered voice analytics platform. Rather than licensing a generic solution, Squeeze engineered it to swap AI models as better ones emerge—including OpenAI’s latest reasoning models—without rebuilding the underlying architecture.
Five Core Value Pillars
- Compliance: Automated review of every call to confirm required disclosures and opt-outs are delivered.
- Quantifying qualitative data: Converting call transcripts into structured data points that inform marketing funnels and sales team management.
- Sales enablement: Recommendations to partner sales teams on product pitching, objection handling, and customer sentiment—both before and after a transfer.
- Performance enhancement: Fast coaching feedback for agents, identifying where new hires deviate from proven call tracks and where top performers excel.
- Improved efficiency: Replacing hours of manual call review with AI-generated insights that leadership can act on immediately.
Insights the Dispositions Miss
Standard call dispositions—”not interested,” “follow-up,” “closed”—collapse nuance. PEEL Voice drills into why a lead went cold: was a mortgage prospect put off by interest rates, bad timing, or a mishandled objection? On the partner side, the platform can even track named agents by MLS number, identifying which loan officers convert at the highest rates so those behaviors can be systematized in training.
The team also measures customer sentiment shift—comparing sentiment at the moment of transfer with sentiment afterward—and correlates that shift with close and application rates.
The Road Ahead
The near-term roadmap centers on real-time reporting: moving insights from periodic offline reports to live dashboards at clients’ fingertips. Longer term, the platform is architected to trigger downstream CRM actions—moving a lead into a new marketing funnel, firing an email sequence, or scheduling a callback—based on what actually happened on the call. Third-party integrations, allowing external call centers to plug their own call flows into PEEL Voice’s analysis layer, are already in early conversations.
The name itself came from AI: prompted with a brief about Squeeze and a request for a citrus-themed acronym, ChatGPT returned PEEL—Performance Enhancing Engagement Learning. The team acquired the domain the same week this episode recorded.
The Human Edge
The hosts are clear that AI alone isn’t the differentiator—it’s the team of prompt engineers with deep call-center backgrounds asking the right questions of each campaign. As Brett put it, even an off-the-shelf tool still requires someone who understands the sales process to configure it meaningfully. That institutional knowledge, combined with a scalable custom build, is what the team believes will produce better insights than any technology-first competitor.
AI is not going to take your job—someone using AI is going to take your job.
— Jacob Thorb
We are technology first, but we have the backing of the call center—years of history and experience that informs the technology. That's why we're going to win.
— Brett Evanson
It's not just are they saying things they should or shouldn't from a QA perspective—what are the good agents doing that are making them convert at a higher percent?
— Brett Evanson
Performance enhancing engagement learning… the meeting room went silent.
— Carson Poppinger
Episode chapters
- 00:10 — Welcome & Episode Introduction
- 01:01 — Origin Story: How PEEL Voice Was Born from QA
- 03:15 — From Squeeze-Centric to Partner Sales Enablement
- 07:55 — Five Core Value Pillars of PEEL Voice
- 10:28 — Custom Build vs. Off-the-Shelf: The Architecture Decision
- 12:38 — Mining the 'Not Interested' Bucket for Hidden Insights
- 15:58 — Vision for PEEL Voice in 2–3 Years
- 21:36 — The Name: How ChatGPT Created the PEEL Acronym
- 26:08 — AI as a Personal Productivity Tool
- 30:31 — Closing Thoughts & Final Reflections
Frequently asked questions
What does PEEL Voice stand for?
PEEL is an acronym for Performance Enhancing Engagement Learning. The name was generated by ChatGPT after the Squeeze team prompted it for a citrus-themed acronym that reflected the platform's purpose.
How is PEEL Voice different from off-the-shelf call analytics tools?
PEEL Voice is custom-built by Squeeze, which means it can be tailored campaign by campaign, upgraded to new AI models without a full rebuild, and configured with prompts written by analysts who have deep call-center expertise—something generic platforms don't offer.
What kinds of sales insights does PEEL Voice surface?
It flags compliance issues, measures customer sentiment before and after transfers, identifies why leads go cold (e.g., interest rates vs. bad timing), tracks whether agents schedule a concrete next step, and pinpoints which agent behaviors correlate with higher conversion rates.
Can PEEL Voice be used by companies outside of Squeeze's contact center?
Early conversations are underway with external companies who want to keep their existing call flows and simply integrate PEEL Voice's analysis layer on top—though that capability is still maturing.
How does PEEL Voice help with sales agent training?
By analyzing transcripts at scale, it quickly identifies when agents deviate from trained call tracks, which objection-handling techniques underperform, and what top agents do differently—giving sales managers data-backed coaching material instead of anecdotal feedback.
What is the future roadmap for PEEL Voice?
The team plans to deliver real-time client dashboards, CRM integrations that trigger automated follow-up actions based on call outcomes, and third-party call-flow integrations—all built on the same flexible, model-agnostic architecture.