Finch

Finch Legal handles pre-litigation tasks for personal injury cases, performing repetitive phone calls to insurance companies and medical providers. After adopting ElevenAgents, the call success rate rose from 59% to 93%, and weekly call volume increased from approximately 500 to over 3,800. Monthly call volume exceeded 15,000.

Previous solutions offered limited control over internal workings, voice quality, latency, and reasoning model adjustments. With ElevenAgents, Finch could directly configure various LLMs, voices, and conversational styles, balancing trade-offs between response latency, voice quality, and reasoning depth, while continuously improving agents using analytics tools.

In calls with insurance representatives, overly robotic voices can lead to short responses or call transfers. Finch applied ElevenLabs' Expressive Mode to adjust the agent's tone and speed according to the conversation flow, encouraging representatives to continue the call.

ElevenLabs' forward-deployed engineers built the first agent directly on Finch's account. Finch went live in days rather than months, and technical issues encountered during the pilot were resolved within 72 hours.

Finch's key workflows are as follows:

• Insurance Calls: Navigate IVRs using DTMF, wait 10–30 minutes or up to 2 hours to connect with a representative, submit claims, and structure the results.
• Medical Provider Follow-ups: Identify cases where only records were sent and bills were missing, then re-initiate bill requests.
• Park on Hold: When employees transfer calls on hold to the agent, the agent handles music and IVRs, returning the call to the agent as soon as a representative connects.

All three workflows use the same infrastructure. Agents navigate phone trees using DTMF and rotate through multiple phone numbers on SIP trunks to reduce the likelihood of specific numbers being blocked by telecom carriers.