Let me tell you about an ordinary Tuesday.
Fourteen MRO supplier quotes in my inbox. Different formats, different currencies, different levels of… let's say effort. One of them was a scanned PDF from a vendor who apparently hasn't heard about email attachments since 2009.
I needed a normalised comparison table and a commercial recommendation on the desk before a 2pm committee meeting.
I had a 10 o'clock. Another at 11:30.
I was doing the mental maths — which corners could I cut? Which vendor was I going to shortchange in the analysis because I simply didn't have time to do it properly?
That was the moment I decided to try something different.
I spent 15 minutes building a prompt in Microsoft 365 Copilot. Fed it all fourteen quotes — scanned PDF included. And got back a normalised comparison table, flagged anomalies across commercial terms, and a plain-English summary of the trade-offs.
Total time: 23 minutes.
I sat with the output for a moment. Not because it was perfect — I edited it. But because a task that had eaten two and a half to three hours of my week, every single week, for twenty-eight years had just become a 23-minute job.
"The AI didn't replace my judgment. It reclaimed 23 minutes so I could actually use it."
Here's the thing I want you to sit with, because this is what separates useful thinking about AI from the noise that's everywhere right now:
The 23 minutes wasn't the point.
The point was what I did with the two and a half hours I got back.
For the first time in longer than I can remember, I had time to actually think about that supplier relationship. To review their performance trajectory properly. To build a real commercial position before walking into that committee room — instead of arriving with a hastily assembled table and hoping nobody asked too many questions.
AI didn't make me faster.
It made me strategic.
That distinction is what this newsletter is built on.
THE FRAMEWORK: Human Time vs. Human Judgment
After that Tuesday, I started paying attention to something I'd never quantified before — the difference between tasks that consume my time and tasks that actually require my judgment.
They're not the same thing. And most of us spend far too much of the former pretending it's the latter.
Human Time = tasks that are repetitive, structured, and rule-based. Data normalisation. Format conversion. Summary generation. Comparison tables. These tasks eat hours. AI handles them well.
Human Judgment = decisions that require context, relationships, risk assessment, and experience. Commercial positioning. Supplier relationship management. Risk trade-offs. Escalation decisions. These tasks require you. AI cannot do them — and anyone who tells you otherwise hasn't sat across a supplier for 28 years.
The problem isn't that procurement professionals are bad at their jobs.
The problem is that we spend 60–70% of our time on Human Time tasks — and arrive at the moments that require Human Judgment already exhausted, already behind, already cutting corners.
That's the gap AI closes. Not by replacing what you know. By clearing the runway so you can actually use it.
THE TOOL: Microsoft 365 Copilot for RFQ normalisation
What it does: Takes supplier quotes in any format — spreadsheets, PDFs, Word docs, scanned images — and produces a normalised comparison table with commercial flags.
Why M365 Copilot specifically: If your quotes land in Outlook and live in Excel, Copilot has the home-court advantage. It pulls directly from your inbox and SharePoint without manual upload. Most procurement teams already have the licence. Most aren't using it for this.
The prompt that works:
"I have [X] MRO supplier quotes attached in different formats. Normalise them into a single comparison table with these columns: supplier name, unit price, lead time, payment terms, MOQ, and any flagged anomalies. After the table, give me a plain-English summary of the key commercial trade-offs between the top three options."
Honest caveats: → Review every output before it goes anywhere. The normalisation is usually accurate; the commercial interpretation needs your judgment applied on top. → Scanned PDFs vary in quality. If a vendor's scan is poor, the extraction will reflect that. → This doesn't replace your knowledge of the supplier relationship. It gives you time to actually use it.
SSC PREDICTION TRACKER #001
Every issue, I put something on the record. Dated. Falsifiable. Tracked publicly.
Prediction: By the end of 2027, procurement professionals who haven't built at least one production AI workflow will be passed over for promotion in favour of less experienced colleagues who have.
The conventional wisdom is that experience and relationships protect senior procurement careers. I think that's about to flip. The 26-year-old buyer who automated their RFQ workflow and freed up 15 hours a week for strategic work will out-deliver the 20-year veteran who didn't — and leadership will notice.
This isn't about AI replacing procurement. It's about AI-fluent procurement people replacing the rest of us.
Filed: May 2026 · Resolution: December 2027 · Confidence: High
BEFORE YOU GO
If you found this useful, the single best thing you can do is forward it to one colleague in procurement or supply chain who'd appreciate a straight-talking take on AI that isn't vendor spin.
And if someone forwarded this to you — subscribe here: thesmartersupplychain.com
Refer one person who subscribes → I'll send you the Procurement AI Prompt Pack: 15 field-tested prompts for the tasks that eat your week.
NEXT ISSUE
I tracked every task I did at work for a full week. Then I rated each one:
🔴 RED — AI can do this now 🟡 AMBER — AI assists, you decide 🟢 GREEN — this needs you
Next issue, I'll show you the full breakdown — and the framework you can use to audit your own week.
The Smarter Supply Chain is written by a procurement professional with 28 years in the field. No vendor relationships. No sponsored claims. Just what I've actually tested.