Work

Compliance-gated AI

Telehealth Client

I built an AI assistant for a telehealth company's website, designed to answer only from approved content, with a separate rules-based safety check on what it says.

Private telehealth service assistant with appointment instructions, approved sources, scope limits and a staff contact option.
0network calls in the button-only build
Client
Telehealth, pre-launch
Role
Built
Timeline
2026 · Delivered
Stack
Next.js, Preact widget, Claude (client's own account), HubSpot, Resend, Jotform HIPAA, Vercel

Way faster than any chatbot I have used previously for any company.

Client founderTelehealth, pre-launch

A telehealth business preparing to launch needed an AI assistant with strict limits on the questions it could answer. The work included checks for those limits and material for the client's legal review.

  1. Built an AI assistant that answers from supplied information, with a separate rules-based check for restricted questions and responses.
  2. Connected inquiries to the client's customer-management tools and drafted six intake and consent forms.
  3. Prepared a button-only alternative that made no network requests, and documented how to shut the system down.

Checks behind the work

622automated test casesgrown from 80
3 of 3challenge questions blocked by the added checkThese same questions passed the earlier prompt-only version. This small test does not establish overall safety.

Say hello.

Tell me what you'd like to build or improve.

Quick navigation

Find a page.

Use the arrow keys to choose a result, Enter to open it, or Escape to close.