Sourcing Memo: Vertical AI Agents in Regulated Industries (Pre-Series A)
Note: stage data moves fast in this category — I've flagged stage as of my last reliable information; several may have raised since. Verify rounds before outreach.
- Tennr — Automates referral/fax document processing workflows for medical practices. Fit: started as a seed-stage YC company (W23) attacking an unglamorous, deeply regulated healthcare paperwork wedge; strong early revenue signal before institutional rounds. (May have graduated past Series A — worth confirming, but the profile is the archetype.)
- SmarterDx — Excluded on stage; replacing with: Anterior (formerly Co:Helm) — AI clinical reasoning agents for prior authorization on the payer side. Fit: founded by an NHS physician, YC-backed, seed stage; prior auth is a regulatory choke point with clear buyer pain.
- Solve Intelligence — AI agent for patent drafting and prosecution inside attorneys' existing workflows. Fit: YC S23, seed-stage, IP law is a high-billable-rate vertical with structured documents — ideal agent terrain.
- Greenlite — AI agents for AML/KYC compliance operations at fintechs and banks. Fit: seed-stage (Greylock-backed), selling into compliance teams drowning in alert-review headcount; strong "agent replaces BPO seat" economics.
- Marveri — AI for legal due diligence document review in M&A. Fit: MIT-founded, seed stage; narrow wedge (diligence checklists) rather than generic legal copilot.
- Cascading AI (Casca) — AI loan origination agent for banks and credit unions, starting with SBA lending. Fit: YC S23, seed; SBA loans are painfully manual and banks pay for throughput.
- Hona — Client-communication automation for law firms (status updates, intake). Fit: seed/early stage, plaintiff-side firms are underserved by legaltech; retention-driven wedge into broader case automation.
- Sully.ai — AI "medical employee" agents (scribe, coder, receptionist) for small physician practices. Fit: early-stage, YC-backed; targets the long tail of independent practices ignored by enterprise scribe vendors.
- Hamming AI — Automated testing/auditing of voice AI agents, including compliance-sensitive deployments (healthcare, finance). Fit: YC-backed seed; picks-and-shovels for regulated agent deployment — every buyer in this thesis needs it.
- Andi (fka others) — low confidence, replacing with: Corgi (insurance) — also low confidence. Safer pick: Basepilot — AI agents automating back-office browser workflows, with traction in insurance and healthcare ops. Fit: YC W24, pre-seed/seed; horizontal tech but landing vertically in regulated back offices.
---
White-space gaps noticed
- Post-deployment compliance & audit infrastructure for agents themselves. Lots of companies deploying agents into healthcare/finance; almost nobody building the audit-trail, model-governance, and regulator-facing evidence layer those buyers will be forced to procure (Hamming is adjacent but focused on voice QA). This becomes mandatory spend as EU AI Act / state healthcare AI rules bite.
- Agents for government-adjacent regulated verticals — Medicaid/Medicare enrollment brokers, court filing/clerk workflows, municipal permitting, and insurance state-filing compliance. Sales cycles scare off founders, so quality of competition is low, but document volume and rule-density are exactly where agents outperform. Almost no credible pre-Series A players here.