How Ace Waves automated 72% of customer service for Reloe in 7 days

TL;DR

• Challenge: 20+ supplement brands, fast growth, constantly evolving procedures, high variance in decisions, and increasing ticket backlog

• Goal: Scale customer service without adding headcount, improve speed and accuracy, prevent escalations, and retain revenue from cancellations and churn

• Solution: AI agents deployed inside existing systems, resolving issues end-to-end

• Delivery model: built on Ace Waves’ multi-agent platform and implemented by an embedded forward-deployed team

• Direct testing: Ace Waves achieved 5 times higher automation than Zendesk AI

• Timeline: 72% automation in the first week; 84% within two weeks

• Costs: 40% reduction in customer service costs

• Retention: 15% uplift by retaining customers who intended to cancel

The situation Reloe was in

Reloe runs a global subscription-based health supplement business across the US, Europe, and Asia. They manage more than 20 brands and several health apps, serving over 800k active customers and generating €60M+ in annual revenue, adding at least 3-5 new brands each quarter. With this pace of growth, operational complexity grew faster than the systems behind it.

Customer conversations were rarely single-issue. One ticket could blend subscription and billing questions with shipment problems and refund requests, sometimes with health-adjacent context that requires careful handling.

The visible pain was a ticket backlog that had grown alongside rapid scale. Underneath it was a consistency problem that kept recreating volume. Procedures, tooling, and governance had not kept pace with growth. Customer Service Debt had accumulated across workflows.

Here’s what that looked like in practice:

• Execution variance increased as volume and brand complexity grew, especially in sensitive flows like refunds and cancellations

• Procedures that needed to be frequently updated as the business evolved

• Manual glue work in spreadsheets and different tools to fill integration gaps

• Policy drift across brands, especially in refunds, cancellations, and retention steps

• Slow reply times: orders move quickly through fulfillment - decisions do not have the luxury of delay.

CEO had a clear objective: reduce customer service costs while protecting revenue from unnecessary cancellations, returns, and subscription churn.

Reloe’s Move from Zendesk AI to Ace Waves

Reloe wasn’t starting from zero. Before working with Ace Waves, Reloe relied on Zendesk AI automation and Copilot. In practice, it improved routing and response drafting, but required constant manual configuration and ongoing updates from the team. The part that remained heavily human-dependent was execution.

For Reloe, where most customer conversations require transactional actions rather than simple replies, Zendesk AI automation reached only 15%.

Subscriptions were still updated manually. Refunds were still reviewed and issued by human agents. Retention decisions depended on interpretation. Cross-system actions required switching between tools. The operating model remained human-led, with AI acting as an assistant.

In direct testing, Zendesk AI could not match Ace Waves results. AI automation coverage increased 5 times, resolution time dropped from multiple days to minutes, and for the first time, AI-driven workflows began actively retaining customers.

“

As we scaled, keeping decisions aligned across brands became more demanding. We needed more than automation. We needed consistency we could trust. Ace Waves enabled us to keep scaling the business without losing control or quality.

Lina Plieniūtė,
Head of Customer Support Ops, Reloe

From deployment to operational scale

Within the first week of deployment, AI agents automated 72% of customer service. Over the following weeks, automation expanded to 84% as additional workflows were rolled out.

The coverage was not limited to one workflow or one brand. Core customer service processes across brands were rebuilt and executed through structured Agent Procedures, while maintaining each brand’s voice and commercial priorities.

What shipped (scope of execution)

AI agents execute the same work inside the same tools human agents previously used, including:

• subscription cancellations and modifications

• retention flows

• full and partial refunds

• shipment issue handling

• order detail updates

• automated triage, tagging, and structured routing

• general customer inquiries and product-related questions

Sensitive cases or ambiguous scenarios are escalated to human agents with full context under defined guardrails.

Ace Waves AI agents operate inside Reloe’s existing Zendesk environment and connected systems, providing 24/7 availability and supporting over 90 languages across Reloe’s global customer base.

How the forward-deployed model made it work

Reloe didn’t need to assemble an internal AI project team or migrate systems. Ace Waves didn’t just add an AI layer and hope it would “figure it out.” We rebuilt the foundations so the work could finish reliably across brands.

Our forward-deployed engineers worked directly with Reloe’s CX and operations teams, taking full ownership business outcomes and translating evolving policies into executable Agent Procedures, as well as shipping AI agents into production.

The work unfolded in three phases:

Phase 1: operational foundations & procedure redesign

• mapped and redesigned end-to-end workflows across brands

• built and formalized support procedures

• centralized fragmented tools and manual processes into unified system integrations

Phase 2: guardrails & revenue logic

• created custom guardrails and human escalation paths

• added geo-specific business rules and boundaries

• designed retention and refund logic aligned with commercial goals

Phase 3: integration & production deployment

• built reporting dashboards to monitor performance and consistency

• integrated AI agents into Reloe’s existing systems and data sources

Refund issuing and cancellation handling were introduced gradually, allowing the team to validate performance before expanding coverage.

When procedures change, they are updated centrally. There is no need to retrain multiple human agents or rely on everyone remembering new rules. The change applies immediately across production. That centralized control made scaling predictable.

Business impact

Coverage and speed:

• Automation rate: 72% in the first week; 84% within two weeks

• Response times: reduced from 3-4 days to 1-2 minutes

Costs and retention:

• Customer service costs: reduced by 40%

• Retention rate: 15% uplift from retaining customers who intended to cancel

Operational quality:

• Fewer errors and less variance through procedure-based execution

• Enabled backlog clearance as capacity returned to the human team

• Customers frequently expressed appreciation for Reloe’s customer-facing AI agent, Ace

What changed inside Reloe

Before Ace Waves, customer service performance depended heavily on people - their memory, judgment, and experience. After deployment, execution is governed by Agent Procedures.

Instead of coaching individuals to reduce variance, Reloe now adjusts the system itself. Instead of scaling through hiring, they expand workflow coverage. Instead of reacting to backlog and inconsistency, they operate with defined logic and measurable outcomes.

Customer service moved from people-dependent and interpretation-heavy to engineered and controlled.

Future outlook

Reloe and Ace Waves continue to build and improve Agent Procedures, with the forward-deployed team working alongside as a long-term partner. As new policies, products, and markets are introduced, procedures evolve with them.

The objective is not only high automation, but a customer service system that evolves with the business.