Overview

AAA — The Auto Club Group is the second-largest club in the AAA federation, serving nearly 13 million members across 14 states. Every year, ACG responds to 6.5 million roadside service rescues — tow calls, battery failures, lockouts, accidents — reaching members at their most stressed and most vulnerable. The contact center operation that supports this runs 24/7, coordinating dispatchers, third-party technicians, and agents across an enormous geographic footprint.

Supernova Award Category

Next Generation Customer Experience

The Problem

Roadside service operations run on time and information so providers struggle to deliver both. Communication between members, dispatchers, and field technicians was fragmented and one-directional: members received status alerts but had no way to reply, ask questions, or provide real-time updates without calling back in and waiting on hold. Contact center agents spent the majority of their time on routine status inquiries rather than complex member needs — a workload that proved nearly impossible to staff for during peak demand in 2021 and 2022. The result was slower service, frustrated members, and agents burning out on repetitive, low-value work.

The Solution

ACG partnered with Riptide to deploy Ripley™, an AI customer advocate built on a patented multi-party orchestration framework, integrated directly with ACG's existing Salesforce Service Cloud. The key insight was reframing the problem: the goal wasn't to deflect calls, it was to give every member an AI advocate who could act, coordinate, and communicate on their behalf across the full service journey. Ripley was deployed on frontline roadside service conversations at scale, with its “no customer escalation” approach that empowers the AI advocate to consult with human agents behind the scenes on the customers’ behalf whenever it’s uncertain. No guessing. Ramping quickly, ACG tuned Ripley's advocacy continuously, reaching full-scale production across 13 states within 90 days.

The results

Before Ripley, members in distress on the side of the road had one option: wait on hold, re-explain their situation, and hope the agent they reached had the full context. Agents were so overloaded with status inquiries that average speed to answer was measured in minutes. After deployment, ACG achieved average speed to answer of four to six seconds on phone channels, not because they added headcount, but because Ripley now handles 30,000 roadside service conversations per day. Of those, Ripley resolves 24% entirely on its own, advocates and coordinates by collaborating with human agents for another 71%, and includes direct agent involvement only 5% of the time. Members who previously couldn't reply to an ETA alert can now hold a real conversation: asking questions, changing service details, or confirming locations with all parties (member, dispatcher, technician, agent) kept in sync through a single shared thread. One member, after a frustrating multi-channel experience, described the Ripley interaction as the only one where she felt someone truly cared.

Metrics

Average speed to answer (phone): (before) Minutes (after 90 days) 4–6 seconds Dispatch handling effort: (before) Baseline (after 90 days) 50% reduction Member satisfaction (CSAT): (before) Baseline (after 90 days) 4%+ increase Agent concurrency: (before) 1:1 (member to agent) (after 90 days) 3–4x/agent w/Ripley Agent attrition: (before) Baseline (after 90 days) ~50% reduction over 3 years Daily AI-handled conversations: (before) 0 (after 90 days) 30,000 Advocacy rate (AI-driven CX): (before) 0 (after 90 days) 95%

The Technology

Riptide's Ripley is the world’s first AI customer advocate powered by a patented LLM-intermediated multiparty conversation framework (U.S. Patents 12,199,936 B2 and 12,101,362 B2), integrated with Salesforce Service Cloud.

Disruptive Factor

The conventional wisdom in enterprise AI is to keep AI away from high-stress, high-stakes interactions. ACG and Riptide did the opposite: they deployed Ripley directly into the most emotionally charged moment in the service relationship — a member stranded on the road, possibly in an accident, possibly scared. It worked because the design was fundamentally different from chatbot deflection. Ripley doesn't pretend to know things it doesn't; it asks for help, bringing a human agent into a parallel conversation behind the scenes to answer the specific question it can't resolve, then carries on without forcing the member to repeat themselves. This "advocate, don't escalate" model changes what AI in customer service can mean: not only a cost-reduction tactic or a boost in productivity, but a genuine upgrade in the quality of care. Any organization running complex, multi-party service operations — utilities, insurance, healthcare, logistics — can apply this model. The architecture is already proven at scale.

Shining Moment

Agent attrition at ACG dropped by roughly half over the three years since Ripley was deployed. That wasn't a goal anyone set out to hit. It happened because agents stopped spending their days on data entry and status lookups (the work that wore people down) and started spending them on the interactions that actually require human judgment and human empathy. Joe Russo, ACG's AVP for Automotive Services, noted the change directly: agents feel more rewarded because they're doing more meaningful work. That outcome — a technology deployment that made the jobs of the people it was supposed to replace more satisfying — is rare. It's worth recognizing.