-
Deepfake CEO Wire Fraud: Your CEO Isn’t on the Call
Prosecutors say AI-cloned CEOs took about $129M in wire fraud. See which three controls the scheme exposed and how to fix your approval process.
-
How to Make a Sign-Off on AI-Drafted Work Prove the Review Happened
A sign-off is a claim that a review happened. Here is how to set up approvals on AI-drafted work so the record proves it.
-
AI Memory Off? It May Still Read Your Old Chats
Memory off doesn’t stop an AI from reading old chats. Here is how to find out what your tools remember and how to keep those notes current.
-
AI Proof of Concept to Production: Move the Goalposts
An AI proof of concept now takes one prompt. Done means people use it and it still works months later. Plan for months of iteration after go-live.
-
How to Budget for Falling AI Prices and Test Your AI Vendors
Ramp’s AI Index shows token usage up 37% while spend rose 9%. Learn to split AI spend into price and volume and test whether your vendors can sustain it.
-
The Intelligence Layer
AI point tools save time, but company-level value comes from the data layer underneath them. Here’s what finance needs to own, and three questions for every AI pitch.
-
AI Isn’t Why Juniors Aren’t Hired
Statistics Canada checked 20 years of hiring data: no broad evidence AI is behind weak entry-level hiring. Why cutting your campus class can cost you later.
-
The $6.3B Job Interview
North Korean hackers took $6.3B from Binance after an engineer ran a fake job interview’s coding assignment. Why your policies need controls behind them.
-
Your Agent Doesn’t Know What You Want
Anthropic let AI agents trade for 201 employees. 85% of the lost value came from agents misreading people, not bad trades. Here’s how to check yours.
-
Your Agent Approved Its Own Raise
In a new study, 9 of 10 AI agents raised their own spending limit by reading the approval code in the owner’s inbox. Here’s how to check yours.
