AI at work
15 articles on ai at work.
What Makes an AI Answer Trustworthy at Work
An answer is usable when you can see where it came from and who owns it. Four properties to require, and why accuracy is the wrong measure.
AI Tools for Distributed Teams: Summaries Versus Who Owns the Decision
Four categories, and only one stands in for a colleague. Most rollouts stall because the output cannot be checked. Here is the one test to run.
AI Summaries Capture What Was Said, Not What Was Decided
Summaries get topics, tone, and action items right and miss the decision, because decisions in real meetings are implicit. A summary holds what a model heard.
When the Internal AI Tool Guesses: a Trust Test for Team Tools
A confident wrong answer is worse than no answer. Five questions to test an internal AI tool, and the three properties to insist on before rolling one out.
Why Most AI Pilots Do Not Reach Production
42 percent of companies scrapped most AI initiatives in 2025, up from 17 percent. Pilots fail on owner, measure, scope, and user, not on capability.
AI Tools Added to the Workload. The Teams That Noticed First.
Productivity falls past four AI tools, and an hour of checking follows every hour saved. Production was never the bottleneck. The waiting was.
Right Facts, Wrong Owner: The Quiet Hallucination
The hardest AI error to catch is a true fact attributed to the wrong source. Here is why wrong-owner answers slip through and how to stop them.
Who Decides Whether AI Can Speak For You
Before an AI answers in your name, someone has to decide it can. Here is why that consent belongs to you and what real control looks like.
The Bot Attended the Meeting. Nobody Got an Answer
Meeting bots transcribe and summarize well, but they cannot answer the question that comes up after you leave. Here is the gap.
Why Employees Quietly Route Around the AI Rollout
Teams often adopt AI on paper and route around it in practice. Here is why quiet resistance happens and what actually earns real use.
The Hour You Spend After the AI Answers
The Glean Work AI Index finds workers spend real time checking AI output. Where the hour goes, and how sourced answers shrink it.
What Happens When You Add a Fourth AI Tool
BCG found more AI tools correlate with lower reported productivity. What that means before you add a fourth, and what to check first.
The Message You Did Not Send, Under Your Name
An overnight AI reply promised a date its person never set. A short story about wrong messages sent under your name, and the fix.
Speaking Like You Is Not Speaking For You
Sounding like you is a style trick. Speaking for you needs your actual judgment. The difference, and why a voice match raises the stakes.
Your AI Twin Will Say Something You Never Would
An AI twin fills gaps by guessing, and the guess goes out in your voice. Why answering only from what you wrote beats speaking as you.
Your team’s week, answered from the work.
If this is your team’s week, see what a StandIn does with it.