- Client
- A US AI startup for government operations, automating public institutions' call centers
- Industry
- GovTech
- Country
- United States
- Project type
- Built from scratch · AI product
At a glance
Before
The client had an idea and a tight deadline, but no product.
After
In a little over three months, the idea became a voice agent answering real citizens, in a pilot delivered 10 days ahead of schedule.
The client
A US AI startup for government operations that automates the call centers of public institutions with voice and text AI agents.
The product and our role
An AI agent that answers calls and text messages, responds to questions, opens tickets and escalates to the human team when needed, with a real-time operations dashboard.
Codeminer42 built it from scratch, on a tight deadline, and delivered the pilot 10 days ahead of schedule. This case is about the delivery: we built it from scratch and put it live.
The challenge
- Public agencies receive a huge volume of repetitive calls, which take up the time and cost of human staff and cannot be covered 24/7.
- The client had the idea and a tight deadline, but not the product.
- Every agency served required the agent to know its processes, departments and internal rules in depth.
What we delivered
- A real-time call dashboard.
- Real-time transcription.
- Automatic summary and categorization of every call.
- SMS integration to send links and information during a call.
- A per-client agent configuration layer (prompt and knowledge base).
Results
- The pilot went into production 10 days ahead of schedule, running at a large public institution.
- About 200 calls handled by the agent in the first 10 days.
- 5 new prospective clients from the pilot.
