Customer Story
How an AI-powered call center built by two developers serves hundreds of SMBs on Akka.
MrCall built an AI-powered call center that handles thousands of simultaneous VoIP requests — serving hundreds of small and medium businesses across Italy — powered by Akka and maintained by just two developers.
Akka ROI Scorecard
developers built and operate the entire AI call center — from voice AI to infrastructure.
infrastructure — resources allocated per call, eliminating idle overhead and scaling linearly.
of simultaneous VoIP calls handled with consistent low latency.
Business outcome: hundreds of SMB customers served daily with AI-powered call management.
Small and medium businesses need intelligent call handling — screening, routing, scheduling, answering common questions — but can't afford dedicated call center staff. MrCall set out to automate this entirely with AI-powered voice agents.
The technical challenge: each inbound call requires real-time speech recognition, natural language understanding, and voice synthesis — all running concurrently across thousands of simultaneous connections. A single dropped call or delayed response breaks the customer experience. MrCall needed a platform that could handle massive concurrency with predictable latency, recover instantly from failures, and be operated by a very small team.
Akka provided the concurrency model MrCall needed to handle thousands of simultaneous VoIP sessions. Each call becomes an independent process with its own lifecycle, state, and failure boundary — meaning one problematic connection never cascades to affect others.
“With Akka, we were able to build the entire platform as a distributed system from day one,” says Angelo Leto, founder. “We didn’t have to retrofit concurrency or resilience — it was there out of the box.”
Akka also kept the codebase small enough for two developers to own and operate. Instead of managing thread pools, connection queues, and failure recovery manually, Akka handles these concerns at the infrastructure level.
Speed to production. Two developers built and shipped a production AI call center. No DevOps team, no dedicated infrastructure engineers — just two people delivering a complete voice AI platform.
Scale. Thousands of concurrent VoIP sessions with consistent latency. Each call is independently managed and isolated from failures in other calls.
Cost to operate. Minimal infrastructure footprint. Resources are allocated only when calls are active and released immediately on completion.
Business impact. Hundreds of SMB customers across Italy rely on MrCall daily for AI-powered call screening, routing, and scheduling.
MrCall’s voice agents already behave like autonomous AI agents — they answer calls, understand intent, and take action without human involvement. On the Akka Agentic AI Platform, this maps directly to governed agentic workflows: each call becomes a stateful agent session with durable memory, tool access (calendar booking, CRM lookup, order status), and auditable decision trails.
As MrCall expands to handle more complex multi-turn conversations — scheduling follow-ups, processing payments, escalating to humans — Akka’s event-sourced state and supervision model ensures every agent interaction is traceable, recoverable, and compliant. Enterprises build on the Akka Agentic AI Platform directly, or have a system delivered and operated through Akka Specify.
