Procurement creates the greatest value before delivery becomes urgent. Once the program is late, operationally disruptive, and integrator-dependent, most contractual provisions become reactive channels for documenting problems rather than preventing them.

Shelfware taught procurement to ask whether software was being used. Consumption pricing requires another question: Is the additional usage still worth what we are paying for it?

You start with a simple idea: automate a few tasks, save some time, and suddenly you’re drowning in complexity. But there’s a way out: use Microsoft 365 - it’s not just a suite of office apps; it’s an open buffet for workflow optimization, which is already there with you no matter what.

I hope the assessment encourages reflection—not on whether you're a better procurement professional than someone else, but on the multidimensional complexity of procurement decisions.

Most discussions about artificial intelligence have focused on comparing foundation models. More recently, the debate has shifted toward prompt engineering. Entire consulting offerings now promise better business outcomes through better prompts. The Rise of Organizational Intelligence The language model is becoming the most commoditized component of the solution. As foundation models become increasingly accessible and technically comparable, competitive advantage is shifting

Ultimately, the goal is not to produce the most impressive CV. The goal is to produce the strongest CV that remains completely supported by the candidate’s own experience.

If expertise could be created through prompting, organizations using the same models and similar prompts should produce broadly similar recommendations.

A corpus answers the question: “What do I know?” A persona answers the question: “How do I think and judge?” The combination is significantly more powerful than either component individually.

What makes sourcing analysis difficult in real life is rarely a lack of information. Usually, the difficult part is interpretation.

SaaS renewals should not be treated as an administrative activity. They are leverage events. In many companies, the renewal cycle becomes the only realistic opportunity to reassess platform strategy, challenge pricing structures, eliminate inactive subscriptions (shelfware,) review duplicated functionality, negotiate scale-down rights, benchmark commercial conditions, reconsider sourcing strategy, or reassess vendor dependency.

It always amuses me when people post "use my prompts", which sounds to me like "skip time on a diagnosis, take my medicine."

Ultimately, one of the most important procurement questions is not “Which supplier is cheapest?” but rather “What does the supplier know that we do not?”

The opportunity is usually not another platform purchase. The opportunity is operational integration and administrative simplification.

The Kraljic Matrix was never intended to become a yearly PowerPoint ritual. It was supposed to support procurement decision-making. The more repeatable, transparent, and automated the process becomes, the more useful the model actually appears.

Transforming procurement isn’t a walk in the park. It takes planning, patience, and a bit of grit.

Most supplier surveys don’t fail because of bad questions. They fail because of how they are performed.

Here we are expecting a breakthrough performance without: redesigning processes (e.g., digitisation vs digitalisation) relieving people's fears of job loss, developing new skill sets, assigning ownership, enforcing usage, measuring outcomes.

Whether the large SaaS vendors can keep control of the orchesration and integration layer remains one of the most interesting questions in the enterprise software industry.

Most organisational inefficiency comes from undefined states, unclear ownership, and uncontrolled transitions — not from a lack of intelligence. Those problems are solved with basic RPA and workflow design, not machine learning.

Nano-SaaS, therefore, emerges precisely where incumbents stop: at the level of micro-niche problems that only become viable when developed through vibe coding.