- the Procurement Blueprint
- Posts
- Cat-herding expert
Cat-herding expert
You are one too
HEY AGAINI started in procurement at a time when pressure equaled professionalism.Endurance was how you survived in the industry. I learned to absorb pressure quietly while coordinating everyone else’s: my stakeholders’, my managers’ and my suppliers’ chaos. That is the definition of cat herding. And procurement does a lot of it. Over time, that way of working leaves marks and burnout becomes normalised. That’s where bias can slip in too, whether it’s cultural or embedded in the AI tools many of us now rely on. This issue connects today’s two threads. The Quiet Fix is written by Hazel, our first ever guest-writer whose experience of burnout guided her new career. The Tech Bit looks at how bias shows up in everyday AI use in procurement and what to do about it. PS: Did you know Cat Herders Day is an unofficial observance on 15th December that honors people who manage chaotic, difficult, or seemingly impossible tasks?! This one is for us! | ![]() a ca |
In Today's Issue
Leadership Can’t Be Automated
AI can help you move faster, but real leadership still requires human judgment.
The free resource 5 Traits AI Can’t Replace explains the traits leaders must protect in an AI-driven world and why BELAY Executive Assistants are built to support them.
The Quiet Fix

BURNOUT & PROCUREMENT
by Hazel Cotton
Hazel is a Nervous System Specialist. She spent two decades working in Marketing Procurement - right from being a buyer at a Bluechip company through to being Global Practice Lead at leading consultancy. However, she spent most of her career chronically stressed, yet unaware of it at the time, which resulted in burnout and chronic illnesses. After healing herself through retraining the nervous system and brain, she pivoted her career. She now runs a wellness company helping high achievers prevent and recover from burnout and chronic illnesses.
Why Spotting the Signs Early Matters
“Burnout is becoming increasingly common in procurement. A significant proportion of procurement professionals report elevated stress and anxiety levels, driven by workload intensity, ongoing supply disruption, cost pressures, and risk exposure.”
This is a quote from the Chartered Institute of Procurement & Supply (CIPS) Q2 2025 Pulse Survey.
And it is why spotting the signs of burnout early are so important because it's no longer a fringe issue in procurement.
As sustained pressure becomes the norm rather than the exception (and let’s face it the economic environment at the moment is making things increasingly challenging), burnout is less likely to appear suddenly and more likely to develop gradually, making early awareness and recognition critical.
And let's be clear. Burnout doesn’t always look the same.
It doesn't mean that one day you are going to collapse on the office bathroom flow in a flood of tears (although you might).
It's not always dramatic.
In-fact it more commonly creeps up over time.
Burnout typically shows up in three key ways:
Emotional signs - include irritability, low motivation, anxiety, cynicism, or a sense of dread about work that never fully goes away.
Behavioural signs - working longer hours, difficulty switching off, withdrawing from colleagues, reduced decision confidence, or procrastination on tasks that once you could do in your sleep.
Physical signs - persistent fatigue, headaches, digestive issues, poor sleep, muscle tension, frequent illness, or brain fog.
Procurement roles carry a unique combination of pressures that make burnout more likely than some other professions.
Constant cost-saving targets, tight deadlines, supplier negotiations, stakeholder conflict, compliance demands, risk management, and the responsibility of getting decisions “right” all create sustained pressure.
Many procurement professionals also operate in reactive environments, firefighting issues while managing long-term strategy, often with limited resources or recognition.
Over time, this level of pressure doesn’t just affect your mindset.
It affects your nervous system.
Let me give you a whistle stop tour of how burnout manifests.
The autonomic nervous system governs how we respond to stress. It has two main branches: the sympathetic nervous system, responsible for fight-or-flight, and the parasympathetic nervous system, responsible for rest, recovery, and repair.
Short bursts of stress are normal and manageable. But if you remain in a heightened fight-or-flight state for weeks, months, or years, your body never fully recovers. This can lead to chronic exhaustion, impaired concentration, emotional reactivity, and eventually burnout or chronic illness.
This is why learning how to regulate your nervous system is not a “nice to have,” but a preventative skill.
Having simple techniques you can use in the moment helps shift the body out of fight-or-flight and back into balance.
Effective techniques include slow, extended exhale breathing, grounding exercises that engage the senses, gentle somatic movement or stretching to release physical tension. It is also valuable to take short breaks between meetings to reset rather than pushing straight through.
One way to assess whether your nervous system has become dysregulated and you are teetering towards burnout is to ask yourself (or your partner!) whether you have been displaying emotional, behavioural or physical signs for more than 2-3 months that’s are outside of your 'normal'. If so, and provided you have had tests to rule anything sinister out then I recommend to take preventative action before things escalate.
Burnout prevention isn’t about reducing your performance or excellence. It’s about supporting your system so you can perform, decide, and lead sustainably.
Try this technique I call TEBB when you are sitting at your desk.
Assess your Thoughts - are they positive, neutral or negative?
Assess your Emotions - are you happy, sad or anxious etc?
Assess your Breathing - are you breathing, fast or slow, shallow from your chest or deep from your diaphragm?
Assess your Body - are you tense, your shoulder and jaw clenched or are your muscles relaxed?
If you notice anything (and I expect that you will!) that could be informing your nervous system and brain that you are in a stressed state then reset your Thoughts, Emotions, Breathing and Body.
Then stand up and shake.
You have just reset your nervous system.
And if you think you might be in a state of chronic stress or overdrive then you will find my Free Stress Recovery Toolkit helpful - you can get the link here.
The Tech Bit

ERASING AI BIAS
This question came up during my AI Agents in Procurement webinar this week which I did with help of the lovely Zalion guys. By the way, 370 of you registered in the end, which I’m super grateful for.
During the Q&A, someone asked us “how can we remove bias from AI?”. And I was unable to give a quick answer so I said something along the lines of “how long is a piece of string”, and moved on.
I promise you this wasn’t a dodge. It was just the most accurate response I could give at the time.
Most procurement teams are already using tools like ChatGPT or Copilot for summaries, options, scoring ideas, supplier lists or initial recommendations. Some of you even have AI embedded inside sourcing platforms or agents that reach into ERP or contract repositories.
But Bias in AI is not a single problem, rather a whole pattern.
A pattern that comes from a model consistently leaning in certain directions because of what it has seen, how it has been asked to reason, and where it has been allowed to act.
The patterns you see in outputs come straight from what the model has learned and how you are using it.
I see three ways that plays out in our field.
Data bias
Most custom-AI that is being used in procurement today is guided by historic data, previous sourcing events, legacy supplier lists, old negotiations.
That data reflects how procurement behaved in the past. It reflects preferences, shortcuts, exclusions, and habits. The model learns those patterns and it also reproduces them at scale.
What comes out of those models, unsurprisingly, is outcomes that look conservative, incumbent-heavy, or narrow. But the model is doing exactly what it was trained to do.
What you should do instead
Treat data as a design choice.
Check who consistently appears in outputs and who never does
Supplement internal data with external benchmarks where possible
Remove attributes that act as crude signals for trust or quality
You cannot rewrite history. You can stop hard-coding it into future decisions.
Prompt and framing bias
This is very relevant to those of you who work constantly with ChatGPT or Copilot in day-to-day procurement work.
The way you ask the question matters.
If you ask for a supplier recommendation without specifying what good means, the model supplies something that fits the most common patterns in the data it knows.
It fills in the blanks using the defaults it has learned. That default behaviour can be similar to bias. The bias often sits in the question.
What you should do instead
Be explicit about what really matters.
State what matters and what does not
Ask for ranked options with reasoning and avoid single answers
Force the model to surface its assumptions
Clear prompts do not remove bias but at the very least they can expose it. Better prompts expose the underlying patterns you are unaware of or ignoring.
Decision and automation/process bias
I saw a very recent public-sector example where staff used ChatGPT to write bid language.
The AI phrasing ended up favouring certain suppliers. That event triggered a actual investigation of the Procurement team. The authority had to review the tender design, explain the criteria, and pause parts of the process. Time and trust were lost.
The bid criteria became skewed because the language came from AI with no guardrails.
So this is where bias stops being an abstract term and starts affecting real work outcomes.
I’ve seen AI used in procurement to:
rank suppliers after an RFP
prioritise sourcing requests in intake tools
flag “high-risk” contracts or suppliers
suggest negotiation levers or award scenarios
The problem is that once those outputs are inside a workflow, people treat them as truth.
A supplier ranked third gets less airtime.
A request marked “low priority” waits longer.
A contract flagged as risky gets escalated, even when the underlying issue is minor.
The model sets the tone and then the organisation follows.
This is how bias becomes operational.
What you should do instead
Decide explicitly where AI can influence and where it cannot.
Use AI to support evaluations rather final awards (remember “human-in-the-loop” from my webinar, if you attended)
Require human review for supplier exclusions or rejections
Make scoring logic visible and challengeable
If a supplier never makes it past an AI-generated shortlist, that is no longer decision support, it is is decision-making, which is what is dangerous.
And once procurement accepts it, the bias stops being just “accidental.”
Feedback Loop or “drift over-time” bias
AI outputs influence actions and actions change the data that gets stored.
The model learns from outcomes it helped create. Over time, patterns become stronger even if no one is watching.
There is a growing body of work showing how dataset skew can persist and amplify unless actively reviewed. The EU AI Act states how training data needs deliberate design and ongoing assessment to remove this type of bias.
If you only look at performance metrics like speed or accuracy, you miss the slow drift in outcomes and incentives.
What you should do
Plan for drift.
Separate training data from decision outcomes
Review models and prompts periodically
Look for changes in behaviour over time
So, can you remove bias from AI?
You cannot think of bias as a “bug” you fix, so no.
Think of it more like a property of systems built on human data and human judgement.
What you can do is shape the inputs, the questions and the decisions you allow it to influence. That gives you control over what kind of patterns it reinforces in your procurement ecosystem so you can stay intentional.
Intentional about data, framing, automation, and accountability.
AI does not make procurement fair or optimal by default. It makes the logic you give it louder.
AI learns patterns. It repeats patterns. It scales patterns. AI scales whatever logic you give it.
Procurement still decides what that logic is.
AI in HR? It’s happening now.
Deel's free 2026 trends report cuts through all the hype and lays out what HR teams can really expect in 2026. You’ll learn about the shifts happening now, the skill gaps you can't ignore, and resilience strategies that aren't just buzzwords. Plus you’ll get a practical toolkit that helps you implement it all without another costly and time-consuming transformation project.
My Best Post Lately

PROCUREMENT MEETINGS
I would say this is the post I’m happiest with this week and best fit’s today’s theme.
It names something I see constantly which is how we spend a lot of time fixing procurement with big things. All the new tools, new models and new frameworks.
Meanwhile, our calendars are doing more damage than any broken process.
When meeting cadence is off, work just circulates. Decisions drift and every risks gets quietly normalised. Soon your stakeholders bypass procurement rather than wait.
What my post was really about was tightening the rhythm instead.
Start having fewer meetings with sharper purpose that drive clear decisions. Otherwise: make it an email!
If your diary is full and progress feels slow, my post might be worth a read.
Free Template(s) of the week

ENTERPRISE-GRADE PROCUREMENT POLICY
This is an old one, actually my first ever freebie created for the newsletter, but one of my most downloaded ones.
Creating a procurement policy sounds simple until you actually try to do it. Which I have, many times!
Every policy must include a set of decisions about how your organisation buys, who gets to decide, and what happens when your people don’t do as they are told.
That’s where most teams I meet struggle.
They argue about thresholds. They struggle with global versus local rules. They write something for auditors that nobody can actually understand. Or they end up with a policy that looks OK but then slowly gets ignored and rarely applied.
This template exists to get you past that paralysis.
It gives you a full-length, usable structure with sensible defaults, clear roles, clear thresholds, and language written for humans. You can open it, adapt it to your reality, and actually roll it out without rewriting the whole thing from scratch.
Still free and still useful. And very much built from and FOR real corporate life.
Do you want access to other great templates from previous newsletters? Have a look at the full store below:
A Final Note
I feel this line from Hannah Arendt is scaringly appropriate to today’s topics:
“The sad truth is that most evil is done by people who never make up their minds to be good or evil.”
When I try to read into it, I do not see it as super dramatic, but more procedural.
Bias rarely comes from bad intent (although there are some real bad apples out there!), what I have experienced is that it usually comes from default behaviours.
From not stopping. From letting systems and habits run unchecked because everyone is busy herding the cats.
That’s true of AI. It’s true of organisations. It’s true of how we work.
Until next time,

Procurement worth reading.


