AI governance is an operations problem, and most companies are handing it to people who write policy.
Pick one workflow that already uses AI. Write down who owns it, what it can touch, and where you would look to see what it did yesterday. Which of those would take you longest?
More like this in the newsletter: nextlinklabs.com/newsletter
Somebody asked me last week if AI has made running NextLink easier. It has not, and I don't think it's supposed to.
Ten years into running this business, here is the uncomfortable part. The technology moves faster than my own judgment can update, and I have to keep checking which of my instincts are still correct. What has this year changed about how you think your own job works?
I share lessons on building better software, stronger teams, and smarter infrastructure. Subscribe to the newsletter: nextlinklabs.com/newsletter
There is a real argument happening among developers right now about whether AI agents make buying software pointless, and both sides are partly right.
Which one of your current subscriptions is really standing in for a workflow specific enough to you that it's worth building instead?
I share lessons on building better software, stronger teams, and smarter infrastructure. Subscribe to the newsletter: nextlinklabs.com/newsletter
MIT looked at 300 real AI deployments and found that 95 percent of them produced no measurable financial return.
The 5 percent that worked were built or adapted with people who understood both the technology and the specific workflow. If your pilot is stalled, did anyone redesign the workflow around it, or did it just get pointed at the old one?
I share lessons on building better software, stronger teams, and smarter infrastructure. Subscribe to the newsletter: nextlinklabs.com/newsletter




