A debate broke out on LinkedIn this week between two of procurement technology’s most credible independent voices. Nico Bac argued that orchestration layers are not delivering the integration outcomes anticipated, because best-of-breed solutions cannot share data unless they share a data model. Jason Busch responded that layering exists because enterprises have fragmented technology environments that no single platform architecture can resolve, and that the fastest-growing segment of procurement technology is precisely the I/O and orchestration layer that Nico is questioning.
Both observations are accurate. The growth in orchestration layers Jason cited is real. The data-model fragility Nico identified is also real. The disagreement between them is not about whether either observation is true — it is about what the growth actually means. Is it evidence that layering works, or is it evidence of an unresolved problem driving repeat purchases? That question cannot be answered at the orchestration layer. It can only be answered at the substrate.
Layering works for the reason Jason describes. Layering also fails for the reason Nico describes. The question worth examining is when each happens, and what determines the difference.
The answer is not in the orchestration layer at all.
The current debate in procurement technology is taking place almost entirely at what I have called Wave 4 — the AI, orchestration, and autonomous coordination layer that has dominated investment cycles since roughly 2023. This is the visible tip of the architecture. It is the layer where vendor roadmaps are positioned, where consulting practices have built service lines, where analyst firms run their Magic Quadrants, and where the operational arguments between practitioners like Nico and Jason actually occur.
What sits beneath Wave 4 is not part of the visible debate, but it determines whether the debate has any practical resolution at all.
The illustration shows the same five-layer stack twice. On the left, the stack is drawn the way it is usually described: a broad foundation of inherited systems supporting a narrowing apex of current AI investment. On the right, the same stack is drawn the way it actually behaves under load. The wave layers occupy the same vertical positions in both views — AI at the top, substrate at the bottom — but the pyramid is inverted. Current AI investment occupies an enormous surface area at the top of the structure, and the substrate beneath it has narrowed to a point.
The contrast between the two views is the entire diagnostic argument. Vendors and consultants describe the architecture using the first image. The architecture behaves like the second image. The gap between description and behavior is where AI initiatives quietly accumulate the conditions for failure.
What the Substrate Actually Contains
Each prior technology wave left an operational inheritance that the next wave was expected to absorb but did not.
Wave 1, from the 1990s and inherited from earlier, deposited local databases, departmental spreadsheets, email-based approval flows, and the operating logic encoded in individual analysts’ working files. These are still operating today inside most enterprise procurement functions. They have not been retired. They were simply layered over.
Wave 2, the ERP customization cycle of the 2000s, deposited a layer of bespoke configurations, custom workflow logic, and process re-engineering decisions that were specific to the implementing organization. ERP was deployed on the assumption that it would be the terminal platform. Most of the customizations made during that wave are still load-bearing today, even where the organization has since migrated to cloud-native SaaS.
Wave 3, the SaaS and best-of-breed cycle of the 2010s, deposited a fragmented ecosystem of cloud applications, each of which was deployed under the assumption that it would be the platform around which the rest of the stack reorganized. That reorganization did not happen. Instead, the SaaS layer accreted alongside the Wave 2 ERP layer and the Wave 1 inheritance, producing the integration problem that both Nico and Jason are now debating.
Wave 4 — the AI, orchestration, and autonomous coordination layer currently being deployed — is landing on top of all three prior waves simultaneously. It is being asked to coordinate across data models that were never aligned, process logic that was never documented, and operational substrate that the organization itself cannot fully describe.
Beneath all four waves sits a fifth layer that almost nobody names: the substrate itself. The shadow workflows, the human-encoded operating knowledge, the undocumented exception handling, the relationships between analysts and suppliers that exist outside any system at all. This is the layer where, in my work on the 1998 Department of National Defence engagement, the diagnostic question “what time of day do orders come in?” produced the performance shift from 51% to 97.3% delivery accuracy. The substrate is where actual operational truth lives. It is also the layer that AI initiatives are categorically least equipped to recognize, because the substrate exists in human practice rather than in system data.
Why Layering Works Sometimes and Fails Other Times
Jason is correct that the I/O and orchestration layer is the fastest-growing segment of procurement technology. Nico is correct that orchestration alone does not resolve the underlying problem. Both are correct because the same architectural move produces different outcomes depending on the substrate conditions beneath it.
Where the prior waves have been retired, consolidated, or documented to the point of operational legibility, orchestration delivers what its proponents promise. Where the prior waves remain unresolved — with shadow workflows still active, customization logic undocumented, data models inconsistent across systems — orchestration compounds the problem rather than resolving it. The orchestration layer cannot impose coherence on substrate that lacks it. It can only surface the inconsistency at a higher altitude, where it then manifests as AI initiative underperformance.
The variable is not the orchestration architecture. The variable is the substrate condition the organization was carrying before it deployed the orchestration layer at all.
This is the structural insight that the Wave 4 debate cannot reach from inside Wave 4. The orchestration question is downstream of a question that almost no organization asks itself before committing to the orchestration investment.
Implementation Physics™
Procurement decisions do not fail inside procurement. They fail when real-world conditions hit them. The conditions that determine whether an AI or orchestration initiative succeeds are not located in the AI or orchestration layer at all. They are located in the substrate beneath it, in the prior-wave inheritance that the organization has either resolved or has not.
This is the operating principle I have begun describing as Implementation Physics™. The frameworks, technologies, and strategies that organizations select for deployment are subject to the structural conditions of the environment they are deployed into, in the same way that any physical system is subject to the conditions of its operating environment. A correctly chosen technology landing into a structurally fragile substrate does not produce the outcomes the technology was designed to produce. It produces the outcomes the substrate permits.
The current AI investment cycle in procurement is, in aggregate, betting that the substrate beneath it will hold the load. That bet is being placed largely without diagnostic verification of whether the substrate can actually hold the load. The MIT, BCG, McKinsey, and Stanford HAI research on AI implementation outcomes is now converging on a single explanation for why enterprise AI initiatives are underperforming relative to expectation, and the explanation is structural: organizational readiness, not technology selection, determines the outcome.
The Compounding Technology Shadow Wave™ framework names what the readiness research is pointing to. The substrate is where AI initiatives succeed or fail.
The Diagnostic Question Worth Asking
The Nico and Jason exchange is the kind of debate that recurs every technology cycle, because it is the visible surface of a deeper question that the industry has not developed adequate diagnostic language for. The Wave 4 conversation will not produce a resolution to itself. The resolution, when it comes, will come from organizations that examine their substrate conditions before committing further AI investment, and from frameworks that can name what the substrate examination is actually looking at.
The question worth asking, before committing to any further orchestration, layering, or AI initiative, is not which architecture to deploy. The question is whether the substrate conditions in the organization can support the load that the planned architecture will place on them.
That is the question Phase 0™ exists to answer. It is the question the 1998 DND case was, in retrospect, the foundational evidence for. It is the question that the entire MIT/BCG/McKinsey/Stanford readiness convergence is pointing at without yet having a name for. And it is the question that the Nico/Jason debate, conducted entirely at the Wave 4 layer, cannot reach from where the debate is being held.
The conversation about orchestration layers will continue. The data-model question will continue. The platform-versus-layering question will continue. None of it will resolve, at the architectural level, until the prior question is asked first.
The substrate is where the answer lives. The substrate is where the diagnostic begins.
The Compounding Technology Shadow Wave™ trilogy executive summaries are available at procureinsights.com. The Phase 0™ Diagnostic — for organizations preparing to commit further AI or orchestration investment — is at hansenprocurement.com/where-does-your-organization-sit-right-now/.
Hansen Models™ — Implementation Physics™ — Procurement Insights archive 2007–2026
-30-
Procurement Insights
The Nico Bac and Jason Busch Debate — and the Structural Question Both Left Unanswered
Posted on May 15, 2026
0
A debate broke out on LinkedIn this week between two of procurement technology’s most credible independent voices. Nico Bac argued that orchestration layers are not delivering the integration outcomes anticipated, because best-of-breed solutions cannot share data unless they share a data model. Jason Busch responded that layering exists because enterprises have fragmented technology environments that no single platform architecture can resolve, and that the fastest-growing segment of procurement technology is precisely the I/O and orchestration layer that Nico is questioning.
Both observations are accurate. The growth in orchestration layers Jason cited is real. The data-model fragility Nico identified is also real. The disagreement between them is not about whether either observation is true — it is about what the growth actually means. Is it evidence that layering works, or is it evidence of an unresolved problem driving repeat purchases? That question cannot be answered at the orchestration layer. It can only be answered at the substrate.
Layering works for the reason Jason describes. Layering also fails for the reason Nico describes. The question worth examining is when each happens, and what determines the difference.
The answer is not in the orchestration layer at all.
The current debate in procurement technology is taking place almost entirely at what I have called Wave 4 — the AI, orchestration, and autonomous coordination layer that has dominated investment cycles since roughly 2023. This is the visible tip of the architecture. It is the layer where vendor roadmaps are positioned, where consulting practices have built service lines, where analyst firms run their Magic Quadrants, and where the operational arguments between practitioners like Nico and Jason actually occur.
What sits beneath Wave 4 is not part of the visible debate, but it determines whether the debate has any practical resolution at all.
The illustration shows the same five-layer stack twice. On the left, the stack is drawn the way it is usually described: a broad foundation of inherited systems supporting a narrowing apex of current AI investment. On the right, the same stack is drawn the way it actually behaves under load. The wave layers occupy the same vertical positions in both views — AI at the top, substrate at the bottom — but the pyramid is inverted. Current AI investment occupies an enormous surface area at the top of the structure, and the substrate beneath it has narrowed to a point.
The contrast between the two views is the entire diagnostic argument. Vendors and consultants describe the architecture using the first image. The architecture behaves like the second image. The gap between description and behavior is where AI initiatives quietly accumulate the conditions for failure.
What the Substrate Actually Contains
Each prior technology wave left an operational inheritance that the next wave was expected to absorb but did not.
Wave 1, from the 1990s and inherited from earlier, deposited local databases, departmental spreadsheets, email-based approval flows, and the operating logic encoded in individual analysts’ working files. These are still operating today inside most enterprise procurement functions. They have not been retired. They were simply layered over.
Wave 2, the ERP customization cycle of the 2000s, deposited a layer of bespoke configurations, custom workflow logic, and process re-engineering decisions that were specific to the implementing organization. ERP was deployed on the assumption that it would be the terminal platform. Most of the customizations made during that wave are still load-bearing today, even where the organization has since migrated to cloud-native SaaS.
Wave 3, the SaaS and best-of-breed cycle of the 2010s, deposited a fragmented ecosystem of cloud applications, each of which was deployed under the assumption that it would be the platform around which the rest of the stack reorganized. That reorganization did not happen. Instead, the SaaS layer accreted alongside the Wave 2 ERP layer and the Wave 1 inheritance, producing the integration problem that both Nico and Jason are now debating.
Wave 4 — the AI, orchestration, and autonomous coordination layer currently being deployed — is landing on top of all three prior waves simultaneously. It is being asked to coordinate across data models that were never aligned, process logic that was never documented, and operational substrate that the organization itself cannot fully describe.
Beneath all four waves sits a fifth layer that almost nobody names: the substrate itself. The shadow workflows, the human-encoded operating knowledge, the undocumented exception handling, the relationships between analysts and suppliers that exist outside any system at all. This is the layer where, in my work on the 1998 Department of National Defence engagement, the diagnostic question “what time of day do orders come in?” produced the performance shift from 51% to 97.3% delivery accuracy. The substrate is where actual operational truth lives. It is also the layer that AI initiatives are categorically least equipped to recognize, because the substrate exists in human practice rather than in system data.
Why Layering Works Sometimes and Fails Other Times
Jason is correct that the I/O and orchestration layer is the fastest-growing segment of procurement technology. Nico is correct that orchestration alone does not resolve the underlying problem. Both are correct because the same architectural move produces different outcomes depending on the substrate conditions beneath it.
Where the prior waves have been retired, consolidated, or documented to the point of operational legibility, orchestration delivers what its proponents promise. Where the prior waves remain unresolved — with shadow workflows still active, customization logic undocumented, data models inconsistent across systems — orchestration compounds the problem rather than resolving it. The orchestration layer cannot impose coherence on substrate that lacks it. It can only surface the inconsistency at a higher altitude, where it then manifests as AI initiative underperformance.
The variable is not the orchestration architecture. The variable is the substrate condition the organization was carrying before it deployed the orchestration layer at all.
This is the structural insight that the Wave 4 debate cannot reach from inside Wave 4. The orchestration question is downstream of a question that almost no organization asks itself before committing to the orchestration investment.
Implementation Physics™
Procurement decisions do not fail inside procurement. They fail when real-world conditions hit them. The conditions that determine whether an AI or orchestration initiative succeeds are not located in the AI or orchestration layer at all. They are located in the substrate beneath it, in the prior-wave inheritance that the organization has either resolved or has not.
This is the operating principle I have begun describing as Implementation Physics™. The frameworks, technologies, and strategies that organizations select for deployment are subject to the structural conditions of the environment they are deployed into, in the same way that any physical system is subject to the conditions of its operating environment. A correctly chosen technology landing into a structurally fragile substrate does not produce the outcomes the technology was designed to produce. It produces the outcomes the substrate permits.
The current AI investment cycle in procurement is, in aggregate, betting that the substrate beneath it will hold the load. That bet is being placed largely without diagnostic verification of whether the substrate can actually hold the load. The MIT, BCG, McKinsey, and Stanford HAI research on AI implementation outcomes is now converging on a single explanation for why enterprise AI initiatives are underperforming relative to expectation, and the explanation is structural: organizational readiness, not technology selection, determines the outcome.
The Compounding Technology Shadow Wave™ framework names what the readiness research is pointing to. The substrate is where AI initiatives succeed or fail.
The Diagnostic Question Worth Asking
The Nico and Jason exchange is the kind of debate that recurs every technology cycle, because it is the visible surface of a deeper question that the industry has not developed adequate diagnostic language for. The Wave 4 conversation will not produce a resolution to itself. The resolution, when it comes, will come from organizations that examine their substrate conditions before committing further AI investment, and from frameworks that can name what the substrate examination is actually looking at.
The question worth asking, before committing to any further orchestration, layering, or AI initiative, is not which architecture to deploy. The question is whether the substrate conditions in the organization can support the load that the planned architecture will place on them.
That is the question Phase 0™ exists to answer. It is the question the 1998 DND case was, in retrospect, the foundational evidence for. It is the question that the entire MIT/BCG/McKinsey/Stanford readiness convergence is pointing at without yet having a name for. And it is the question that the Nico/Jason debate, conducted entirely at the Wave 4 layer, cannot reach from where the debate is being held.
The conversation about orchestration layers will continue. The data-model question will continue. The platform-versus-layering question will continue. None of it will resolve, at the architectural level, until the prior question is asked first.
The substrate is where the answer lives. The substrate is where the diagnostic begins.
The Compounding Technology Shadow Wave™ trilogy executive summaries are available at procureinsights.com. The Phase 0™ Diagnostic — for organizations preparing to commit further AI or orchestration investment — is at hansenprocurement.com/where-does-your-organization-sit-right-now/.
Hansen Models™ — Implementation Physics™ — Procurement Insights archive 2007–2026
-30-
Share this:
Related