Skip to main content

The LayBuild blog

How AI support agents work, and where they break

We build an AI agent that answers support questions from your own documents, says so when they don't cover the question, and lets your team take over the chat. These posts explain how that works, where it falls short, and how to measure it honestly.

1 to 9 of 46 articles

How LayBuild works

Regression-testing a support agent with bun test

Two test layers for an AI support agent: fast contract tests with a stubbed model on every commit, and a golden question set against a real model.

10 min read

Running support

Preparing an AI support agent for a sale spike

What to set up before a sale: pinned sale terms, Q&A pairs for shipping and returns, plan limits, rate limits, handoff staffing and a load test.

8 min read

Security and deployment

Why PHI shouldn't go through LayBuild, and what can

LayBuild isn't HIPAA compliant and signs no BAA. What HIPAA asks of a vendor handling PHI, where we fall short, and safe patterns for healthcare support.

7 min read

Security and deployment

LayBuild's four roles and what each one can do

How access control works in LayBuild: Owner, Admin, Agent and Customer, what each role can reach, the audit log, TOTP two-factor, and what's missing.

6 min read