AI & data
AlmostHuman.ai — Voice to analytics AI POC
Phykon (AlmostHuman.ai)Visit site ↗
A proof of concept that turns spoken questions into trade and operations analytics—voice as the front end to structured insight.
AlmostHuman.ai needed a fast way for operators to interrogate trade and order data without building another static dashboard for every question.
ADVOP delivered a voice-to-analytics AI POC: speech is transcribed and interpreted, grounded in the product’s data model, and answered with charts, summaries, or tabular follow-ups—so stakeholders could explore metrics conversationally during reviews.
The POC validated interaction patterns and latency targets before committing to a full production rollout, and informed how voice fits alongside the existing agent dashboards and automated order workflows.
Scope & outcomes
- Voice capture and intent routing for analytics queries
- Grounding responses in trade and operations data
- Summaries and visual analytics surfaced from natural questions
- POC scope: validate UX, accuracy, and integration hooks for a later production build