Powered by margaritas and mild panic

LET’S GET INTO IT →

In Today's Issue

• The Quiet Fix - RISK MICRO-MONITORING

• The Tech Bit - DECISION DRIFT

• My Best Post Lately

• Free Template of the Week

The Quiet Fix

RISK MICRO-MONITORING

I would bet money that your first clue that something is wrong with your supplier is unlike to be a red flag on your dashboard.

Instead, it could actually be when your supplier’s finance contact sets an out-of-office for “a few weeks” and never comes back.

Or when your supplier’s sales lead suddenly starts posting motivational quotes on LinkedIn (which if you did not know is the universal signal for “I’m about to leave.”).

And all this time your fancy risk platform is most likely still showing you everything as green, including your thirty-seven widgets, six dashboards and three subscriptions…

Your whole team in the meeting squints at the screen like they’re reading tea leaves while nodding gravely as if the chart means something, but no, it really doesn’t.

We in procurement have kept attempting to industrialise intuition for a gazillion years, which is exactly what I like to call “micro-monitoring”.

Somewhere between spreadsheets and sixth sense lies the real, raw, honest truth, the art of paying attention.

And this is what I mean exactly:

• Setting up Google Alerts for your top suppliers, but focus on court cases, layoffs, and “restructuring” stories that were not written by their PR department.

• Keeping an eye on LinkedIn churn. If three key people disappear in a fortnight, that’s most likely not coincidence.

• Watching Glassdoor ratings. A half-star drop in six months might not crash your business, but it often predicts internal mess.

• Tracking invoice timing. Suppliers who start paying you later or sending invoices out of pattern are waving a small red flag.

It’s low-tech, nosy, and incredibly effective.

Why the small stuff matters

Two small nuggets that say it all:

• McKinsey found that over 70 % of supply-chain disruptions come from tier-two or smaller suppliers, usually the ones least visible to your systems.

• A WTW report found that only 9 % of companies believe they have real visibility into supplier risk.

Four practical habits to start tomorrow

1. Run a Monday morning “radar” check.
   Ten minutes, ten suppliers, Google Alerts on one screen, your inbox on the other. Look for anything odd (financing deals, legal filings, mergers…) and note it.

2. Create a “late pay” watch-list.
   A supplier that starts slipping by 15 days or more is telling you something. It’s not about your process but their cash.

3. Keep a people tracker.
   Once a quarter, glance at your top suppliers’ key contacts on LinkedIn. Who’s moved? Who hasn’t been replaced? Attrition is often the first tremor.

4. Listen to your own colleagues.
   Accounts Payable, project managers, and that one person who knows everyone… they’ll spot risk before the system does. Make it normal to share small observations, even if they sound like gossip.

None of this costs anything, and all of it saves something.

Why this is the quiet fix

Because it’s the opposite of a big fancy risk tool. But rather it’s about the small routines that turn your procurement department into early-warning systems instead of noisy and “transformational” after-action historians.

The irony of it all is that most “big risk” projects eventually fail because they forget the humans who actually notice things.

The best systems will always need people who are just slightly suspicious for a living - try and prove me wrong!

The Tech Bit

DECISION DRIFT

There’s a fine line between smart and too smart.

It’s pretty much the same line my three-year-old crosses when he figures out how to open the snack cupboard and then insists it was “autonomous decision-making.”

Procurement’s AI agents are doing the exact same thing. They are supposed to help with triage requests, tidy data, nudge approvals… but somewhere along the way they can start developing opinions.

So today I want to explore, what happens when it starts doing things we didn’t ask for?

The illusion of control

According to Gartner, agentic AI is “the next evolution”.

Agents are systems that pursue goals independently, which sounds awesome until you come to the realisation that procurement’s goals are rarely clean nor consistent.

For example, I bet you money that if you ask three stakeholders to define “best value” what you will actually end up with is a philosophical debate instead of a clean dataset.

So now imagine… a machine learning from that! This is where risk actually brews, so not like in a killer robot sense, but in the slow erosion of your boundaries.

The point where looking for efficiency becomes increasing AI authority.

Decision Drift: how it starts

From everything I have seen so far, no one suddenly decides to hand over control to AI (at least not yet).

It just sort of… happens.

Your sourcing assistant gets permission to “auto-triage” requests under £5,000 and then someone in your excellence team raises it to £10,000 to save time. 12 months down the line, no one you know has reviewed a low-value supplier in months.

The problem isn’t at all that AI makes mistakes, I know first-hand that us humans do that spectacularly well!!

The real issue I see is that AI mistakes don’t look like mistakes until it’s way too late.

And that’s exactly what I call the AI Decision Drift, which is when automation slowly extends its reach without anyone redrawing the lines.

It’s just momentum.

And let me caveat that HR, Finance, Marketing run similar risks but but in procurement, the stakes can sometimes be that little bit higher because we don’t just buy coffee and laptops; we approve obligations.

This same risk exists in contract bots that “pre-redline” clauses, or intake tools that decide routing logic on probability rather than policy. All technically correct, but also all slightly dangerous if unchecked.

The illusion of autonomy

Gartner calls this differently, the named it agentic shift. AI tools that act on intent rather than explicit instruction. It’s rather exciting if it wasn´t because intent is rarely clear in procurement.

For example, humans can’t agree on what “urgent” means, so GOOD LUCK training the algorithm.

My suggestion is taht before your AI assistant starts doing “just the low-value stuff,” ask yourself a few questions:

These are the questions that make you sound like a real procurement leader.

A better way to think about it: the 4 D’s

So how do you control (rather than resolve) this issue? For now just forget “humans in the loop”, loops are for pretzels!!

What you actually need to implement is a clear delegation design: who decides, who delegates, who double-checks, who defers.

1. Decide
   Keep decisions that affect risk, ethics, or supplier relationships. Always human and traceable.

2. Delegate
   Let AI handle repetitive, factual, reversible tasks like data pulls, reminders, formatting…. all the things you’d trust a new analyst to do unsupervised.

3. Double-check
   Build audit logs for anything automated. If it acts, it leaves a trail.
   If there’s no trail, it didn’t happen.

4. Defer
   Some decisions are not ready yet. When the model’s confidence outruns your comfort, stop.

I know this 4 Ds model isn’t sexy but it’s how we can all truly keep automation safe.

The irony of intelligence

AI isn’t coming for procurement jobs but procurement judgement.

And that’s worse because it happens under the flag of increasing efficiency… until one day you realise the system hasn’t asked for approval in months, and you’re the one who trained it to stop asking.

So yes, let the agents help and automate the boring bits.

Use the copilots of the world but make sure the smartest thing in the room is still you.

My Best Post Lately

MY BIRTHDAY POST

Funnily enough my most-liked post this week wasn’t about procurement transformation, AI, or anything remotely clever. It was… my birthday post!

A beach, a margarita, and apparently that’s what the algorithm wanted all along.

I spend weeks researching supplier-risk frameworks… hundreds of likes. One joke about ageing procurement professionals: thousands!

There’s something humbling (and a bit hilarious) about that so maybe people just wanted to see the human behind the templates.

Either way it reminded me how generous this community can be. Every message, like, and comment made my day and it’s still surreal that so many of you (3,500+ and counting!) read this newsletter, week after week.

So, thank you.

For caring about procurement and for caring about the people who work in it.

Now, back to our regularly scheduled programming with slightly more SPF and possibly still a mild tequila glow.

READ THE FULL POST →

Free Template(s) of the week 

MY BEST PROCUREMENT ASSESSMENT TOOLKIT

Confession time.

The freebie I meant to send you this week was a shiny new AI Readiness Assessment.

But… it didn’t quite make it across the finish line.

Turns out birthday margaritas and complex scoring frameworks don’t mix particularly well. Who knew?!

So while the AI tool gets its final polish (and I recover from an aggressive amount of cake), I’m sending you something equally useful and, frankly, one of my proudest creations:

The Procurement Assessment Toolkit

If you’ve ever had to build or rebuild an indirect procurement team from scratch, this is the framework that keeps you sane.

It’s based on the same questions people constantly ask me:

• “How do I know if my procurement setup is actually mature or just bureaucratic?”

• “Where do I start when everything’s half-working?”

• “How do I show progress without paying someone in a suit to tell me I’ve made progress?”

This toolkit answers those questions with a structured checklist, scoring matrix and a roadmap to take your function from firefighting to focused.

Now, here’s the twist: this toolkit is about to stop being free.

Starting today it won’t be included in the new-subscriber welcome email, and you are the last batch of readers who’ll get it.

It’ll still be available via the link below for one more week only as usual.

My friend Tom keeps telling me I give away far too much for free, and perhaps (for once) I should listen to him. So now’s the time.

And for the small, admirable minority who already downloaded it: email me, I’ll check your download, and I’ll send you something extra as a thank-you.

The AI Readiness Toolkit will land in the next issue (assuming I stay away from margaritas long enough to finish it.)

👉Download my Best Procurement Assessment Toolkit here.

A Final Note

This week’s quote comes from the writer G.K. Chesterton:

I like that.

Procurement, and especially digital procurement, is full of castles in the clouds right now.

AI agents, predictive analytics, “autonomous” everything. It’s all really exciting, i admit that even for me, but then I realise someone still needs to build the staircase.

That’s where we live between the cloud and the concrete. Between ideas that sound magnificent in presentations and the small, steady fixes that actually make them work.

And maybe that’s the job: not to stop dreaming big, but to make sure there’s still a door, a budget, and a PO number when the dream lands..

See you in two weeks with fewer margaritas and a finished AI toolkit

Until next time,

Procurement worth reading.