The questions buyers actually ask.
Direct answers to the questions that come up before someone hires us — what drives the cost of a build, why AI projects stall, and how to tell production-ready from demo-ready.
What to ask an AI vendor
Ask for artefacts, not opinions. Anyone can say yes.
Read →
How to add RAG to an existing SaaS product
Tenant isolation first. Everything else is a smaller problem.
Read →
How to run evals on an AI agent before launch
Write the test set before the demo, not after the incident.
Read →
How do you know if an AI agent is production ready?
A demo proves it can work. Production asks what happens when it doesn't.
Read →
Is premium audit AI regulated
No rule names premium audit. One reaches it anyway.
Read →
Can you scale a Bubble app
Further than critics claim. Not as far as you eventually need.
Read →
Does OSHA PSM apply to fuel storage
Usually not. Atmospheric tanks are written out of the rule.
Read →
Do I need RAG or a bigger context window
Long context is real now. It still is not a document store.
Read →
Does sending client data to an LLM waive privilege?
A court has now answered this. The tool's terms decided it.
Read →
Fair housing rules for AI tenant screening
Buying the model does not move the liability.
Read →
How to connect an AI agent to your CRM
Start read-only. Earn write access with an audit trail.
Read →
How to migrate from Zapier to n8n
Move the workflows that hurt. Leave the rest alone.
Read →
Is algorithmic rent pricing legal
Legal in most states. The inputs are what got sued.
Read →
How do you build HIPAA-compliant AI
Decide what the model never sees. Then build.
Read →
Is your AI build EU AI Act compliant
The high-risk deadline moved. The transparency one did not.
Read →
Should an AI receptionist handle emergency calls
No. Route it to a person, and design that route first.
Read →
TCPA rules for automated service texts
Add one promo line to a service text and the rules change.
Read →
What a BAA does not cover
A signed BAA is a promise, not a control.
Read →
What does an AI agent project actually cost
The model is the cheap part. Your data is the bill.
Read →
Why clinical AI models fail in production
The model kept working. The hospital changed around it.
Read →
Why do AI projects fail
Most AI projects die of the wrong problem, not the wrong model.
Read →
What happens after the AI pilot
Most pilots do not fail. They just never become anything.
Read →
What is an idempotency key and why payments need one
The retry is not the bug. Treating it as new is.
Read →
What Part 11 requires of a system
Part 11 only bites where another FDA rule already applies.
Read →
What PCI scope actually means
Your PCI scope is set by how you embed the form.
Read →
When does my software become a regulated medical device
Time pressure is what pulls software into device territory.
Read →
Why you never use floats for money
Floats cannot hold 0.10. Store money as integers instead.
Read →
Why your RAG returns wrong answers
The model is fine. It never saw the right passage.
Read →
