AI Checks Now Understands What Changed. Not Just What's Wrong.
I've been in freight since 1999. The hardest part of the job has never been finding data. It's always been knowing which data matters right now.
Here's a scenario I've seen play out a hundred times. A shipment shows an ETA of next Tuesday. Required delivery is next Wednesday. One day of buffer. Looks fine.
But yesterday that same shipment had 15 days of buffer.
Overnight, you went from comfortable to needing to call the carrier. The current state didn't change. The situation did.
That's not a data quality problem. That's a memory problem.
We just shipped history awareness to AI Checks.
AI Checks now compares each shipment against its prior version. When a check runs, the AI knows what changed. Not just what's true right now.
That's the difference between 200 alerts and 8.
Without memory, every shipment with a tight deadline looks like a problem. With memory, only shipments that recently developed a tight deadline surface. The ones that have always been tight stay quiet. You already know about those.
Same logic applies everywhere: delays getting worse, exceptions that cleared and then reappeared, routing that's inconsistent with the last submission. Change-triggered flagging. Not state-triggered.
What this looks like in practice
A large freight forwarder we work with has a rule: capture a reason code whenever shipment data changes. Simple requirement. Good operational discipline.
In practice, their TMS buries that field two screens away from the shipment record. It gets missed constantly.
With history awareness, AI Checks catches it. When a value changes without a reason code, the shipment surfaces immediately. Before the customer sees it.
That's the kind of alert that actually matters. Not a report of everything that's wrong. A specific shipment that needs a specific action today.
Supply chain is a process, not a photograph.
ETAs shift. Exceptions resolve and come back. A shipment that looked fine two days ago is broken today.
AI built on snapshots will always be noisy. It flags things already handled. It misses things that just went sideways. The more shipments you run through it, the less signal you get.
Memory fixes that. It's what turns an alert tool into something that actually understands where a shipment has been.
AI Checks doesn't replace your TMS. It sits next to it.
Your TMS holds the data. That's what it's built for. But it wasn't built to tell you which of your 500 open shipments needs your attention at 8am on a Tuesday. That's a different problem.
AI Checks adds the intelligence layer. It reads what your TMS knows, runs your rules against it, and hands you a short list of shipments that need action today. When something changes, it resurfaces. When it's handled, it stays out of your way.
Most teams are filling this gap with spreadsheets, morning reports, and someone manually comparing yesterday's data to today's. That works until it doesn't. It doesn't scale. It misses things. And it burns out good operators.
This is what that gap looks like when you actually solve it.
This is live. Not a roadmap item. Not a beta.
Try it at checks.chain.io. Free tier is 500 checks per month. No credit card.
If you're a freight forwarder or 3PL dealing with this every day, reach out. This is exactly what we built it for.
