Home / Case Studies /
Voice Agent
Case Study — Voice AI
A Voice agent in hindi that answers customer support calls with no
human intervention
A production voice agent automating inbound customer support for Battery
Smart, India's largest EV battery swapping network - it speaks Hindi,
pulls live data from database, battery, and station APIs and resolves
in scope calls end-to-end with zero human handoff.
0
Human handoffs for in-scope queries
1000+
Calls handled per month
70%
Of inbound volume fully automated
Client Context
Phone-first support at fleet scale
Battery Smart operates India's largest EV battery-swapping network.
Its users are drivers on the road all day, Hindi-speaking and
phone first. When a driver needs to know their battery status, find
the nearest station with available batteries or check a payment, they
call.
That made the call centre the operational bottleneck: every new cohort
of drivers added call volume and the only way to absorb it was more
headcount.
The Problem
Support costs scaled linearly with growth
-
Most inbound calls were repetitive, structured lookups, battery
status, station availability, swap history, account and payment
queries; consuming trained agents on work a system could do.
-
Peak hour queues left drivers waiting at swap stations for answers
they needed immediately.
-
The caller base is Hindi first. Conventional IVR menus and
English centric bots performed poorly and drove calls straight back
to human agents.
-
24/7 coverage meant night shifts and redundancy, expensive to
staff, hard to retain.
The Solution
A voice agent wired into live production data
We built a Hindi voice agent that answers the call, understands the
request, fetches the answer from live systems, and speaks it back in a
full conversation loop with no human in it.
Inbound Call
Telephony layer
Twilio
→
Speech to Text
Sarvam AI STT, Hindi-tuned
→
LLM Agent Layer
Intent, dialogue state, tool calling
→
Live APIs
Driver · Battery · Station data
→
Text to Speech
Sarvam AI TTS, natural Hindi voice
-
Hindi native speech pipeline. Sarvam AI STT/TTS
tuned for Hindi and Hinglish code switching, robust to noisy
roadside audio.
-
Real answers, not scripts. The LLM agent layer
calls production APIs at answer time with battery state, station
availability and account data are live, never canned.
-
Hard scope guardrails. The agent resolves in-scope
queries end-to-end and gracefully hands anything else to the
escalation path; it doesn't guess.
-
Conversational latency. A streaming STT/TTS
pipeline keeps response times natural: ~200ms latency on average.
Results
Support capacity decoupled from headcount
In scope driver queries no longer touch a human. The call centre
workforce is reserved for genuinely complex cases; support capacity
stopped scaling with customer growth.
Zero
Human handoff for in scope queries resolved fully by the
agent
1000+
Inbound calls handled per month
70%
Share of total inbound volume automated
30%
Reduction in call-centre workforce dependency
24/7
Coverage — nights and holidays included, no shift staffing
+70%
First-call resolution rate for in-scope queries
Tech Stack
What it's built with
Sarvam AI STT
Sarvam AI TTS
LLM Agent Layer — Tool Calling
Live REST API Integrations
Twilio
Meta Llama 3.3 70B