TriFetch AI

TriFetch AI develops "Linda," an AI agent marketed to primary care and specialty medical practices as a replacement for front-office administrative labor. Linda covers three workstreams: answering inbound patient phone calls around the clock and in multiple languages (scheduling, rescheduling, confirmations, message-taking and routing clinical questions to staff, with capacity stated at up to 50 simultaneous calls); reading, sorting and entering faxes and referrals into the EMR the same day, flagging urgent referrals, faxing incomplete referrals back to the referring provider and calling patients to book; and prior authorizations, including eligibility checks, same-day submission, tracking and appealing denials.

Rather than integrating via APIs, TriFetch uses what it calls a "glass overlay": Linda is given her own system login and operates the clinic's existing browser-based screens the way a staff member would — opening the fax folder, opening charts, clicking buttons and filling fields. The company states this approach works with any browser-based system, naming Epic, eClinicalWorks, athenahealth, NextGen, Elation, Tebra, Office Ally and Oracle Health, as well as fax portals, payer portals and email. Linda can run roughly ten screens in parallel, and the company says nothing is finalized until a human approves it. Each deployment is configured to the individual clinic's scheduling rules, provider preferences and definitions of urgency, and the company lists specialties served including ophthalmology, cardiology, physical therapy, pediatrics, orthopaedics, allergy and asthma, ENT, dermatology, gastroenterology, pain management, primary care, nephrology, endocrinology, behavioral health, pain and spine, and pulmonology.

On data handling, TriFetch states it is HIPAA compliant and SOC 2 certified, signs BAAs with client practices and with every AI provider in its stack (including Google Cloud and Anthropic), does not use identifiable PHI to train AI, encrypts PHI in transit and at rest on US-based cloud infrastructure with least-privilege access, multi-factor authentication and audit logging, and backs each client engagement with $1M in cyber liability insurance.

Founding story

TriFetch was founded by two UCLA Computer Science PhDs who have published healthcare AI research in Nature and ICML. The company has been covered by Forbes under the headline "Meet The UCLA PhDs Ending The US Healthcare Admin Nightmare."

Business model

TriFetch sells a configured AI agent service to medical practices. Implementation is a structured, hands-on process: TriFetch shadows the clinic's team and drafts a structured SOP of its workflows, the clinic reviews and signs off, TriFetch builds and tests the agents on dummy patient data, takes Linda live with an accompanying handbook, and then jointly defines escalation points for human intervention. Engagements carry a 30-day money-back guarantee written into the contract; the company states no client has requested a refund.

Traction

TriFetch publishes a case study of a multi-provider gastroenterology practice that handed Linda its referral backlog: in the first two weeks, 767 backlogged referrals were cleared, 27 patients were verified, authorized and queued for procedures, and roughly 18,000 record updates were written to eClinicalWorks with no manual entry, which the company describes as a 5.4–6.6x ROI on the practice's own numbers. The company also states fewer than 2% of patients recognize Linda as an AI and that no client has invoked its 30-day refund guarantee.

Latest developments

The company reports a $1.9M pre-seed round led by Nexus Venture Partners, SOC 2 certification and HIPAA compliance, Forbes coverage of its founders, and the addition of NextGen Healthcare co-founder Tim Eggena as founding strategic advisor.