🎧 NEW TODAY — THE SMARTER SUPPLY CHAIN PODCAST IS LIVE
As of today, everything you read here also has a voice. The podcast launches with its first episodes ready to binge - same practitioner lens, same no-hype rule, now in your ears on the commute.
Listen on Apple → 【 APPLE PODCASTS SHOW 】
Listen on Spotify → 【SPOTIFY】
FROM THE DESK
It's a Tuesday morning. Fourteen supplier quotes to compare. A contract clause to check before a 2pm renewal call. A performance review to prep for Thursday. A year ago, that was the whole day gone.
This week, three of those four things barely need me for most of the hours they take. The trick was never the AI. It was working out which three.
Every task in a procurement week splits into two piles: work that needs your judgment, and work that just needs your time. For 28 years the tools meant both piles needed me — the admin and the strategy each demanded a person in the chair. For the first time, one of those piles doesn't.
So I did the unglamorous thing. I took the twelve tasks that actually fill a typical week and rated each one by exactly how much AI can take off my plate. Red, amber, green. Not a vendor demo. Not a conference slide. My own week, my own work, with the time it cost before and after sitting right next to it.
THE SIGNAL — WHAT CROSSED THE DESK THIS WEEK
🟡 Gartner puts a number on it: 28% of procurement time is automatable
Gartner Supply Chain Practice
Gartner now says nearly 28% of procurement time is ripe for automation, and that one in five procurement roles will be reshaped by 2030. Read the fine print and they bury their own lede: teams that automate the work without redesigning the role "will struggle to capture AI's full value." That sentence is the whole game. The 28% figure isn't the story — what you do with the reclaimed time is.
🔴 80% of P&G's procurement transactions are now "touchless"
Gartner survey, June 2026
Reported alongside a finding that 78% of Fortune 500 firms have automated at least one procurement function. "Touchless" is the RED category running at industrial scale — and it's a trap dressed as a win. A touchless purchase order saves the time. It only banks the value if the freed hour moves somewhere that needs judgment. Most don't move it. They just process more POs, faster. That's the treadmill, not the exit.
🟢 The projection nobody's framing correctly: agents 70%, CPOs near-zero
Suplari, 2025–2035 role projections
One analysis puts purchasing agents at a ~70% chance of role reduction and the CPO role at almost nil. That's not two forecasts. It's one — RED versus GREEN drawn as a career chart. The role doesn't disappear. The share of it that was always admin with a senior title does. Which is exactly why the map below matters more than the forecast.
THE MAP — 12 TASKS, 3 COLOURS
Human Time vs Human Judgment
Here's the rule in one line: RED means AI does it better and faster — your job is to review and approve. AMBER means AI does the data and the draft, you do the interpretation and the decision. GREEN means it's yours, and it stays yours.
🔴 RED — FULLY REPLACEABLE (review & approve)
Supplier quote normalisation — 2–3 hrs → 20 min
First-draft RFQ / RFP documents — 1.5–2 hrs → 25 min
Routine supplier correspondence — 45–60 min → 10–15 min
Contract clause extraction — 1.5–2 hrs → 15 min
Four tasks. Six to ten hours a week, depending on your workload. That's the time AI gives back — not by cutting corners, but by removing prep work that never needed expert judgment in the first place.
🟡 AMBER — AI-ASSISTED (you keep the decision)
Supplier research & pre-meeting intelligence — 1–1.5 hrs → 20 min + analysis
Spend analysis & category reporting — 2–3 hrs → 40–60 min
Market & commodity price monitoring — 45–60 min → 15 min
Supplier performance review prep — 3–4 hrs → 1–1.5 hrs
Tender evaluation scoring & shortlisting — 2–3 hrs → 50–70 min
The amber principle: AI handles the data, the draft, and the first pass. You handle the interpretation, the context, and the decision. The skill to build is knowing exactly where the handover sits.
🟢 GREEN — HUMAN REQUIRED (permanently yours)
Supplier negotiation — AI preps only
Strategic supplier risk & the decisions — AI preps only
Supplier relationship development — can't be replaced
Only three tasks land squarely here. That low number isn't a sign procurement is simple — it's a sign of how badly the ratio of strategic to admin work has been skewed by the absence of better tools. The green work is what the reclaimed red-task time is for.
THE HONEST WARNING
When I first ran these workflows, I clawed back six to eight hours a week. Then I quietly filled almost all of it with more admin. More emails. More reactive work. A modest efficiency gain and no real change in what I was creating. That's the default outcome, not the exception.
Before you deploy a single one of these, answer one question honestly: what will I do differently with the time? If the answer is "handle more of the same," you're just running faster on the same treadmill. If it names a specific green task — a relationship on autopilot, a category strategy built on last year's assumptions, a negotiation you've only ever prepped on the fly — then you're using AI the way it's actually worth using.
TOOL SPOTLIGHT — THIS WEEK'S RED TASK TO STEAL
Contract Clause Extraction 🔴
This is the rating that surprises people most — and the one that should've been automated a decade ago. Upload a supplier contract to a document-capable tool like Claude or Microsoft CoPilot, ask the five questions below, and you get cited answers in about 15 minutes instead of the two hours you used to spend just finding the relevant clauses before review could even start.
The honest caveat: this replaces the hunting, not the judging. What those clauses mean for your commercial position is still yours. It's not legal review — it's the prep that makes legal review faster. Got a renewal on the calendar this fortnight? Run it on that.
Please review the attached contract and answer these questions, citing the relevant clause for each: 1) Termination — grounds and notice for each party? 2) Liability — what's the cap, any unlimited carve-outs? 3) Exclusivity — any provisions, and their scope/duration? 4) IP & tooling — who owns it, and what happens on exit? 5) Price review — what triggers a change, and who can invoke it? Where a provision is absent, say so explicitly. This is for commercial review, not legal advice.
PREDICTION TRACKER — ON THE RECORD
🟢 Prediction #001 · On Track
By end of 2027, AI-fluent procurement professionals get promoted ahead of more experienced but non-fluent veterans. Holding up: 94% of procurement execs now use generative AI at least weekly. Individual fluency has quietly become table stakes. The promotion gap is what comes next.
🟡 Prediction #002 · Watching
The 4% of teams that have scaled an AI workflow to production will triple by end of 2027. Gartner's 28%-automatable figure and the rush of "touchless" enterprise reporting are early tailwinds. The dividing line stays the same: not budget, but specificity. The teams that win pick one high-friction workflow and prove it.
🔵 Prediction #003 · New This Issue
By the end of 2027, the first public wave of "AI redesigned my role and I do more strategic work now" will be shadowed by a quieter wave of redundancies. The dividing line won't be who adopted AI. It'll be who was doing green work — and who was doing administration with a senior title. AI isn't going to replace procurement professionals. It's going to expose which ones were which.
Start with the free playbook
If this map was useful, the full version is The Procurement Professional's AI Playbook — 39 pages rating every task, with the prompts behind each one. Free.
Get the free AI Playbook → thesmartersupplychain.com
New here? Subscribe to The Smarter Supply Chain Chronicle and get the next issue in your inbox on Tuesday.
Know someone still doing red work by hand? Share your link — two subscribers unlocks the full 24-prompt Procurement Prompt Library.
The Smarter Supply Chain
Practical AI for procurement and supply chain professionals. Written by a practitioner with 28 years in the field. No theory. No hype. No borrowed content.