Bajaj Finserv is using 442 autonomous voice agents.
According to their investor deck: These agents closed ₹1,980 crore (~$230M) in personal loans in 90 days (18% of call center revenue).
Peak volume is ~650k applications/day.
Collections saw a 55% lift with ~90% less human involvement.
The system operates across 22 Indian dialects with sub-second latency.
What surprised me isn’t that they’re using voice automation, but the depth of it.
This looks closer to full workflow ownership than IVR or call deflection.
According to their investor deck: These agents closed ₹1,980 crore (~$230M) in personal loans in 90 days (18% of call center revenue).
Peak volume is ~650k applications/day. Collections saw a 55% lift with ~90% less human involvement. The system operates across 22 Indian dialects with sub-second latency.
What surprised me isn’t that they’re using voice automation, but the depth of it.
This looks closer to full workflow ownership than IVR or call deflection.