Ferrellgas

How Condado Helped Ferrellgas Reduce Abandonment by 60% with Yellow AI & NICE CX

Facing 90% abandonment rates, frustrated customers, and long hold times, Ferrellgas required an AI-driven solution to improve efficiency and enhance customer experience.

Condado designed and implemented an advanced AI orchestration layer, integrating Yellow AI with NICE CX, a previously unsupported integration. By enabling AI-powered automation, Ferrellgas saw:

• Abandonment rates drop from 90% to 30%
• A 200% increase in phone payments—without human intervention
• A nearly 50% reduction in agent headcount

About the Client

• Client Name: Ferrellgas
• Industry: Energy (Propane Distribution)
• Company Overview: Ferrellgas is one of the largest propane distributors in the U.S., providing fuel solutions for residential, commercial, and industrial clients. 

The Problem: A Call Center at Breaking Point

By 2023, Ferrellgas was facing a customer service crisis:

• Abandonment rates were at a staggering 90%. Customers waited on hold for 10-15 minutes, only to give up before reaching an agent. Long wait times resulted in frustrated customers, leading to poor agent experiences when calls finally connected.
• The existing IVR system was outdated and convoluted, forcing customers through multiple confusing menu layers before reaching an agent.

Ferrellgas recognized that AI-powered automation was the only way forward. They had already chosen Yellow AI as their AI platform, but there was no native integration between Yellow AI and NICE CX (Ferrellgas’s existing CCaaS platform).

To unlock AI’s full potential, Ferrellgas needed:

1. Seamless NICE CX + Yellow AI integration for end-to-end call handling and reporting.
2. An AI-powered self-service system (Autopilot) to handle inquiries without human intervention.
3. AI-assisted support for agents (Copilot) to improve call resolution efficiency.

Recognizing Condado’s expertise in engineering industry-first AI integrations, Ferrellgas selected us as their Yellow AI and NICE CX integration partner. 

Immediate Considerations: Challenges Condado Had to Solve

While Ferrellgas had a clear vision, several technical and operational roadblocks stood in the way of execution:

1. No Native Integration Between Yellow AI and NICE CX

At the time, Yellow AI was not a NICE-certified partner, meaning:

• There were no pre-built connectors or APIs to link the two platforms.
• Real-time reporting and data exchange between Yellow AI and NICE did not exist.
• Custom-built AI capabilities were required to serialize and memorialize AI-driven interactions within NICE CX.

2. Ensuring AI-Powered Automation Without Losing Visibility

While Yellow AI could automate call flows, there was no way to track AI-driven interactions within NICE’s reporting system. Ferrellgas needed complete visibility into:

• How AI was training itself
• How AI-handled inquiries were classified and resolved
• What customers were asking AI versus what required agent escalation

Since Yellow AI lacked a reporting API, Condado had to engineer a solution to sync data between Yellow AI and NICE without breaking AI automation.

3. Streamlining Customer Experience

• AI interactions had to feel natural, ensuring customers did not get stuck in a loop or feel "trapped" in an automated system.
• AI had to seamlessly escalate calls to live agents when necessary, avoiding drop-offs.

Our Approach: Engineering a Seamless AI-Powered System

Condado built an AI-powered integration layer that allowed seamless communication between Yellow AI and NICE CX. 

Key Capabilities of the Solution

1. AI-Powered Self-Service (Autopilot)

• Automated call classification, routing, and self-service options to reduce customer reliance on agents.
• Allowed customers to handle routine inquiries and transactions entirely through AI.

2. AI-Augmented Agent Support (Copilot)

• Enabled AI-driven insights for agents, ensuring context-aware call escalations.
• Provided real-time interaction history from Yellow AI to NICE agents.

3. Seamless AI + NICE CX Integration & Reporting

• Engineered custom APIs to serialize and memorialize AI data within NICE CX, ensuring Ferrellgas had complete reporting visibility.
• Worked directly with Yellow AI to develop custom endpoints that allowed NICE CX to pull or receive AI-driven call data.

4. Enabling AI-Powered Payments

• Integrated payment flows within Yellow AI, allowing customers to complete transactions without human intervention.
• This feature drove a 200% increase in AI-handled payments.

Deployment & Integration Strategy

Since Yellow AI had no built-in NICE CX integration, Condado had to engineer every component from scratch:

1. Built custom listeners within NICE CX to receive and store AI-generated data.
2. Worked with Yellow AI to create custom API endpoints for bidirectional data flow.
3. Configured NICE CX’s reporting tools to recognize AI-driven interactions.
4. Rolled out the solution in phases, monitoring abandonment rates, customer interactions, and AI accuracy.

Key Impact & Business Outcomes

Ferrellgas experienced massive improvements post-deployment:

• 90% abandonment rate dropped to 30%
• 200% increase in AI-handled payments—with zero human intervention
• Agent headcount reduced by nearly 50%, improving operational efficiency
• Seamless AI-human handoff, eliminating frustrating IVR loops
• Complete reporting visibility, allowing Ferrellgas to track AI-driven interactions

Foundation for Future AI Integrations

This AI-powered transformation positioned Ferrellgas to scale its automation strategy further. Building on this success, Ferrellgas is now:

• Expanding AI-powered automation into chat and SMS (in active development).
• Exploring deeper AI-personalization strategies based on real-time insights.
• Further optimizing AI-human collaboration for even better customer experience.

Final Thoughts 

This AI integration showcases Condado’s ability to tackle complex, industry-first engineering challenges for enterprise clients. Our work with Ferrellgas demonstrates why industry leaders turn to Condado for solutions that go beyond off-the-shelf technology, addressing real operational challenges with precision and scalability.