.jpg)
AI-enabled IT Sourcing Strategy Tool for Procurement Professionals
- ukrsedo
- Jun 5
- 4 min read
Updated: 49 minutes ago
Procurement professionals rarely disagree because they have different information. They disagree because they interpret the same information differently. Years of experience shape how they recognize commercial risks, evaluate supplier behavior, balance governance against agility, or judge when a sourcing strategy is becoming too risky. This raises an interesting question: can AI be taught not only procurement knowledge but also procurement judgement?
This AI-enabled IT sourcing strategy tool helps procurement professionals develop sourcing recommendations by combining structured analysis with procurement judgement.
The Interpretation Game
While experimenting with AI-supported sourcing analysis recently, I noticed something interesting. The outputs became substantially more useful once I stopped trying to make the environment behave like a generic procurement copilot and started injecting actual procurement reasoning patterns into it instead.
What makes sourcing analysis difficult in real life is rarely a lack of information. Usually, the difficult part is interpretation.
Two procurement professionals can review exactly the same IT sourcing scenario and focus on completely different risks, leverage points, governance concerns, implementation problems, supplier behaviors, or commercial exposures. One immediately starts worrying about lock-in risk and transition complexity, another about negotiation positioning or operational continuity.
The difference is not data. It is an accumulated commercial judgement developed through years of dealing with suppliers, projects, stakeholders, failed implementations, contract structures, organizational politics, and category-specific frustrations.
Most procurement AI tools eliminate that human layer in favour of generic consistency.
What Is an IT Sourcing Strategy?
An IT sourcing strategy determines how an organization should acquire, deliver, and manage technology products and services to achieve its business objectives in line with the relevant category strategy.
Rather than focusing solely on selecting suppliers or negotiating contracts, it considers the broader commercial, technical, financial, and operational implications of sourcing decisions.
Developing an effective sourcing strategy requires balancing multiple factors, including business priorities, internal capabilities, market maturity, supplier competition, implementation risks, regulatory requirements, and the total cost of ownership. Procurement professionals must also decide whether a requirement is best fulfilled through in-house development, commercial software, cloud services, outsourcing, managed services, or a hybrid sourcing model.
A well-designed sourcing strategy provides a clear rationale for procurement decisions, aligns stakeholders around common category objectives, and reduces the risk of inconsistent or reactive purchasing. It also creates a transparent framework for evaluating sourcing options and selecting the commercial approach that delivers the greatest long-term value rather than simply the lowest acquisition cost.
As technology markets evolve and sourcing decisions become increasingly complex, companies benefit from structured methodologies that combine data, market knowledge, and professional judgement to support consistent decision-making.
Revision 2: An IT Sourcing Strategy with a Human Touch
I started experimenting with the solution logic intentionally shaped not only by procurement knowledge bases but also by my professional viewpoints, sourcing philosophy, governance preferences, implementation skepticism, supplier management instincts, and category reasoning patterns. In practice, it starts behaving less like a robo-consultant and more like my way of thinking.
Initially, the solution operated on relatively simple switch logic structures, in which different branches triggered different scenario variables. Basically, if A — then B; if C — then D. Then I started layering AI on top to aggregate multiple data clusters into a more digestible sourcing strategy output.
Recently, I downloaded my entire ChatGPT conversational archive, split it into distinct knowledge chunks (procurement, automation, writing style, ideation, governance thinking, etc.), and started feeding them back into the environment as knowledge files.
Then the real work started.
Over time, I realized the problem wasn't the AI model. It was the reasoning model.
So we systematically removed generic consulting language and replaced it with implementation experience, governance trade-offs, procurement heuristics, and the communication style I've developed over more than two decades.
Eventually, the entire prompt architecture had to be split into five sections due to its size and complexity. Some sections are dynamic and fed by workflow outputs, while others operate more like persistent behavioral instruction layers. On top of that, additional knowledge chunks are dynamically injected based on the sourcing context.
What I genuinely do not yet know is whether AI-supported sourcing analysis actually improves once it starts absorbing the reasoning patterns, biases, judgment logic, communication style, and accumulated frustrations of the people doing procurement work.
That is the part I’m trying to test properly now with your unbiased and constructive feedback.
Therefore, I opened AI credit-limited access to the environment for testing and feedback.
https://www.goodspending.com/sandbox (my website section with the link to the data intake form)
The intake process takes about 5 minutes with generic or imaginary non-confidential data, and the sourcing strategy usually arrives within 40–45 seconds via email. If nothing arrives, then either I broke another workflow somewhere (I'll fix that immediately), or the email landed in junk.
The solution itself can be migrated into a corporate Microsoft 365 environment and adapted to different categories, governance structures, supplier ecosystems, sourcing methodologies, and operational realities.
It seems that the next stage of enterprise AI isn't teaching models more procurement knowledge. Perhaps it's teaching them expert procurement thinking.
Deep Dive into the SaaS Sourcing Strategy topic
P.S. As a side experiment, you can also generate a surprisingly realistic personal profile from your own ChatGPT conversational archive. Mine was disturbingly accurate and definitely not sugar-coated. Consider this a bonus deliverable.
Comments