Poszo
Exploring whether conversational AI can become part of a restaurant team.
Ordering food sounds simple.
In reality it requires understanding people, handling ambiguity, asking the right questions and keeping service moving under pressure.
Poszo explores whether modern AI can perform that role naturally and reliably.
The Question
Voice AI has improved dramatically.
But can it move beyond answering questions and become an operational team member trusted to interact with customers?
Poszo exists to explore that question.
Why I’m Building It
Quick-service restaurants operate in fast, noisy environments where every second matters.
Rather than building another chatbot, Poszo explores whether conversational AI can understand customers, manage conversations naturally and know when a human should take over.
The challenge isn’t speech recognition.
It’s building trust through conversation.
Current Areas of Exploration
Natural Conversation
Creating interactions that feel human rather than scripted.
Real-time Decision Making
Understanding incomplete orders, ambiguity and changing customer intent.
Human Escalation
Knowing when to transfer a conversation smoothly to a staff member.
Restaurant Integration
Exploring how conversational AI fits into real operational workflows rather than existing as an isolated demo.
Latest Lesson
Fast inference alone isn’t enough.
The real challenge is maintaining a natural conversation while making accurate operational decisions under real-world conditions.
The hardest engineering problems turned out to be conversational rather than technical.
Roadmap
- 01Research
- 02Proof of Concept
- 03Pilot Demonstration
- 04Operational Trials