Toma operates a conversational AI platform purpose-built for automotive dealerships, with a modern TypeScript + Node.js + React stack deployed on Kubernetes and AWS. The tech choices—real-time voice processing, PostgreSQL + Prisma for structured data, Prometheus + Grafana for observability—reflect a company solving for low-latency, high-reliability agent interactions in a safety-critical domain. Recent adoption of Plaid signals expansion into dealership financial workflows beyond voice automation.
Toma builds AI voice agents that automate inbound and outbound communications for automotive dealerships while maintaining guardrails against hallucination and customer experience degradation. The platform handles call answering, lead qualification, appointment scheduling, and integration with dealership management systems (DMS). Since launch in 2024, the company reports processing over 1 million calls for dealerships across the United States and recovering millions in revenue through automation and staff time savings. The engineering-focused hiring profile (5 engineers, senior-skewed) and active project list (real-time voice improvements, ML/LLM deployment, observability, DMS integrations) indicate a young company iterating on agent reliability and integration depth.
TypeScript, Node.js, React, Next.js, PostgreSQL, Prisma, AWS (ECS, EKS, RDS), Kubernetes, Docker, Prometheus, Grafana, and Elasticsearch. Recently adopting Plaid for financial integrations.
Real-time voice AI, dealership integrations (DMS, scheduler, Plaid), ML/LLM deployment pipelines, observability infrastructure, customer onboarding tooling, and latency optimization for agent responses.
Other companies in the same industry, closest in size