Furhat Robotics.

The Situation

Furhat Robotics builds social robots designed for natural, human-like interaction. Their flagship platform combines speech, facial animation, and AI — the robot holds a conversation, reads the person in front of it, and responds expressively. The technology is used in education, healthcare, customer interaction, and research.

The platform had real depth, but it was carrying the weight of its own growth. Telemetry was fragmented, making it difficult to diagnose issues across deployments. Core systems had accumulated technical debt. And the tools available for customising robot behaviour — building characters, designing faces, defining how a persona speaks and responds — didn't match the ambition of what the platform could do.

What We Did

Indpro worked across several layers of the platform over an extended engagement.

Telemetry Architecture

The telemetry architecture was redesigned from the ground up. The original system made it hard to see what was happening across different deployments and harder still to act on it. The rebuilt version is centralised and modular — real-time logging, faster diagnostics, and a structure that can be extended without disrupting what's already running.

Face Editor

Indpro built a Face Editor: a dedicated tool for designing and customising robot faces, with dynamic expression support and real-time preview. It works for developers building with the Furhat SDK, and for non-technical users who need to customise without touching code. Both use cases are handled in the same interface.

Platform Performance

Core platform work ran in parallel — reducing interaction latency, refactoring legacy components, and improving the developer-facing UI. The goal was a platform that felt as capable as it actually was.

AI Creator

The most significant addition was the AI Creator — a tool for building and deploying AI-driven robot personas. Personalities are defined with traits, tone, and backstory, then connected to OpenAI for natural conversational capability. The full cycle from creation to deployment is handled in one place.

The Result

Furhat's platform became faster to monitor, faster to debug, and more accessible to the people building with it. Interactions became more responsive. The tools for customisation — both technical and non-technical — closed a gap between what developers and end users could do with the platform.

The engagement ran across system infrastructure, product tooling, and core platform performance. That breadth was deliberate: the individual pieces only work if the platform underneath them is solid.