Loka’s AI agent is automating 100% of Epicor’s auto-service scheduling
Industries
SaaS & ISVs
Tech & TOOLS
AWS Lex, AWS Lambda, Amazon DynamoDB, Amazon RDS with pgvector, LangChain, Amazon CloudWatch, MLflow, AWS CDK, Figma
Teams & Services
AI & Agentic, Product Design & Development, AWS Engineering
milestones
Production-ready AI appointment scheduling
Industry
SaaS & ISVs
Tech & TOOLS
AWS Lex, AWS Lambda, Amazon DynamoDB, Amazon RDS with pgvector, LangChain, Amazon CloudWatch, MLflow, AWS CDK, Figma
Teams & Services
AI & Agentic, Product Design & Development, AWS Engineering
milestones
Production-ready AI appointment scheduling
Loka's AI agents on AWS handle 100% of Epicor's booking flow. No humans required.
The situation
Vehicle service scheduling should be conversational. Instead, many repair shops still depend on phone calls and manual scheduling, creating inefficiencies for staff and friction for customers.
Epicor partnered with Loka to explore how generative AI could transform the appointment experience. The objective was to build an intelligent conversational assistant capable of scheduling appointments, answering customer questions, and helping diagnose vehicle issues through natural language that would feel even more enjoyable and useful than appointment conversations we're all used to. The pilot would demonstrate how AI could automate routine customer interactions while providing the foundation for future product integration across Epicor's automotive software platform.
The challenge
With thousands of phone calls, online requests, and manual scheduling, customers needed a faster way to schedule service without waiting on hold or navigating complex booking forms. At the same time, they expected immediate answers to common questions about shop hours, pricing, available services and appointment availability.
The solution also needed to go beyond a traditional chatbot. It had to understand customer intent, classify vehicle problems accurately, remember previous conversation context, and retrieve up-to-date business information from live website content. All of this needed to happen through a conversational experience that felt natural and intuitive while minimizing the need for human intervention.
“Beyond a chatbot, the agentic solution understands customer intent, classifies vehicle problems accurately, remembers previous conversation context, and retrieves up-to-the-minute business information from the website.”
The solution
Loka designed and built an AI-powered, appointment-scheduling assistant using AWS-native services and modern large language model frameworks. The conversational agent guides customers through the complete booking workflow while gathering the information required to schedule service appointments.
The assistant classifies reported vehicle issues, recalls context throughout the conversation, answers questions using current website information, and helps customers identify the appropriate service before completing the scheduling process. To demonstrate how the experience could evolve into a production application, Loka also created interface wireframes illustrating future platform integration.
The production milestone provided Epicor a working example of how conversational AI can simplify customer engagement while reducing operational overhead for repair shops.
What we delivered
Conversational appointment scheduling: Customers can schedule service appointments entirely through natural language.
Intelligent issue classification: The AI agent captures customer descriptions of vehicle problems and categorizes them to help route appointments and improve service preparation.
Context-aware conversation: The assistant remembers interactions, creating a more natural and personal experience.
Live knowledge retrieval: By accessing current website content, the chatbot provides accurate answers about services, pricing, hours of operation and other FAQs without requiring static responses.
Future-ready product design: Alongside the working prototype, Loka developed product wireframes that demonstrate how conversational AI can be integrated into Epicor's broader customer experience.
Project results
The AI assistant enables conversational bookings, intelligent issue triage, and self-service answers through a single interface. The pilot establishes a production-ready foundation for AI capabilities across Epicor's automotive portfolio.
100K+: Epicor solutions run in 100,000+ automotive locations across North America, giving this capability a built-in path to industry scale.
5,000+: At just 5% adoption, Epicor turns one pilot into a product serving 5,000+ businesses, each one dissolving the time and cost of manual booking.
1 → All: One AWS-native architecture extends beyond scheduling to parts lookups, estimates and self-service across the portfolio.
Repeat ∞: A repeatable pattern for any appointment-driven business. Leaders in healthcare, home services and beyond can implement it with Loka.
