Procurement Insights · May 19, 2026
A Gartner webinar published in late March 2026 — 4 Essential Ingredients for Successful AI Literacy and Change Management — argues that AI adoption requires twenty-five percent more training effort and up to two hundred percent more change management effort than traditional technology. Underestimating that lift, Gartner observes, puts AI efforts at risk of stalled adoption and lost value. The webinar positions AI literacy and change management as the unlock for converting AI investment into enterprise value.
The framing is not wrong. AI literacy matters. Training matters. Change management matters. Organizations that ignore these variables produce predictably worse outcomes than organizations that invest in them. The Gartner observation is operationally true at the layer it operates on.
The problem is that the framing is also extraordinarily familiar. It is the same diagnosis the industry has offered for every technology wave since 1990. The vocabulary changes with each era. The structural answer does not. And the documented failure rate across all thirty-five years demonstrates that the answer — however thoroughly applied — is structurally insufficient to explain what determines implementation outcomes.
The Thirty-Five-Year Pattern
Walk back through the technology eras since 1990, and the pattern is striking enough that it deserves to be named.
In the early 1990s ERP and business process reengineering wave, the industry response to stalled implementations was train users, manage change, redesign processes. User adoption was treated as the primary variable. The empirical record from that era documents implementation failure rates in the sixty-five to seventy percent range despite extensive investment in literacy and change management programs.
In the 2000s e-procurement, portals, and early SaaS wave, the diagnosis shifted vocabulary but not structure. Supplier adoption, user adoption, process compliance became the explanatory frame. Failed deployments were attributed to insufficient stakeholder buy-in, inadequate training, or change resistance. Failure rates remained stuck in the same band. The Change Management Myth post in this archive — published in May 2007, in the first weeks of the Procurement Insights publication — named the change management framing as a myth at exactly the moment the broader industry was institutionalizing it as the answer. The Dangerous Supply Chain Myths series that followed extended the critique through additional documentation. The analytical position the Hansen Models™ framework holds in 2026 was being articulated contemporaneously when the e-procurement wave was at its peak.
The 2010s cloud and SaaS best-of-breed wave produced an even more emphatic version of the same answer. Digital culture, agile transformation, user experience became the analytical centerpiece. Organizations were urged to invest in cultural readiness, agile mindsets, and user-centered design. SaaS fragmentation accelerated. Shadow workflows multiplied. The failure rate held. The Procurement Insights archive was interrogating the corporate culture framing contemporaneously — the 2009 Six Sigma relevancy debate and the 2010 Survival of the Fittest post both documented that the culture variable was insufficient to explain technology adoption outcomes at the moment the broader industry was treating it as the load-bearing factor.
The late-2010s analytics, automation, and RPA wave doubled down. Data literacy, automation readiness, citizen developers entered the vocabulary. Bad process logic was automated rather than structurally corrected. The literacy investment was real and substantial. The realized outcomes did not improve materially.
The 2020-2023 pandemic digitization wave compressed the same diagnosis into emergency conditions. Digital maturity, collaboration culture, resilience became the framing. Workarounds that were designed as temporary became permanent operational substrate. The literacy and change management investment intensified. The structural condition deteriorated.
The current 2023-2025 generative AI, copilots, and agentic systems wave has produced the most extreme version of the historical pattern. AI literacy, prompt training, adoption, trust dominates the discourse. Gartner’s twenty-five percent and two hundred percent figures quantify the scale of the recommended investment. And the documented AI initiative failure rate — measured by MIT, BCG, McKinsey, Stanford HAI, and other major research institutions — sits in the same sixty-five to eighty percent band the industry has continued to cite across thirty-five years.
What the industry has never sustained is the empirical discipline that would have made its own diagnosis falsifiable. No major firm has run longitudinal research measuring whether organizations that invested heavily in change management produced measurably better implementation outcomes than organizations that did not. No major firm has correlated end-user literacy investment levels against realized technology value capture across a controlled cohort. No major firm has measured corporate culture maturity against adoption success in a way that would empirically validate the framing the firm itself is publishing. The prescription has been continuous. The validation has been absent. If the industry had been tracking the relationship between the three explanatory variables and the outcomes those variables are claimed to produce, the absence of improvement would have been visible as feedback, and the framing would have evolved to engage the structural variables actually determining the outcomes. That feedback loop has not operated. Which is why the partial diagnosis has persisted without empirical challenge for thirty-five consecutive years.
The 2025 post on the AI Change Management Myth in ProcureTech selection and implementation extended the 2007 Change Management Myth critique directly into the current AI wave — the same analytical position the framework held nineteen years earlier, applied to the technology era now in active deployment.
The Visual Argument
[GRAPHIC: thirty-five-year pattern bar chart showing per-era attention levels for end-user literacy, corporate culture investment, and change resistance management, with technology advancement and technology implementation outcome lines overlaid]
The pattern is not subtle when rendered visually. The industry’s analytical attention to end-user literacy, corporate culture, and change resistance has been consistently high across every era — and has intensified with each successive wave. The technological capability ceiling has risen continuously, with each wave more sophisticated than the last. The realized implementation outcomes have remained roughly flat, oscillating within a narrow band around sixty-five to eighty percent failure for thirty-five years.
The visual produces an argument that words alone struggle to land. If literacy, culture, and change resistance were the load-bearing variables determining implementation outcomes, the outcomes should have improved as the industry’s attention to those variables intensified. They did not. The flat outcome line in the presence of rising attention to the explanatory variables is the empirical evidence that those variables are not where the explanation lives.
Something else is doing the work — or rather, something else is failing to do the work. That something else is the operating environment the technology is being deployed into. The substrate.
What the Pattern Reveals
The substrate question is structurally different from the literacy question. Literacy asks do the people know how to use the technology. Substrate asks can the operating environment actually support what the technology is being asked to do. The two questions sound similar. They are not.
A workforce can be fully literate in a technology and still produce stalled adoption if the operating environment cannot support the technology’s requirements. Training people on ERP did not fix misaligned process assumptions in the underlying business operations. Training people on SaaS did not fix the fragmented operating models the SaaS portfolio was being asked to integrate. Training people on analytics did not fix the bad data lineage the analytics depended on. Training people on RPA did not fix the broken process logic the RPA was automating. Training people on AI agents will not fix the unclear authority, shadow workflows, incentive misalignment, and undocumented operating logic the agents are being asked to navigate. And no, this is not a technology integration problem between diverse systems — the SOA promise and the integration architectures that followed it confirm exactly this point. Successive waves of integration architecture have been deployed at enterprise scale across two decades. The substrate problems those architectures were positioned to resolve have persisted regardless.
The thirty-five-year pattern is not evidence that literacy and change management are unimportant. It is evidence that literacy and change management cannot resolve substrate inconsistency. Organizations that invest in literacy without addressing substrate produce trained workforces operating inside environments that cannot support what the training is teaching them to do. The training does not fail. The substrate does. But the failure gets attributed to the training, to the culture, to the change resistance — which produces the next wave’s expanded investment in the same explanatory variables and the same disappointing outcomes.
The cross-temporal evidence is documented in the Procurement Insights archive. The 1998 to 2007 to 2024 post — bridging the DND engagement that began in 1998, the Virginia eVA initiative, and the Colgate-Palmolive procurement transformation — demonstrates that the same underlying success factor operated across three technology eras and three different organizations. The variable that determined outcomes was not the era’s dominant literacy framework or change management methodology. It was whether the organization had asked the prior question about its operating environment before deploying the technology investment. The cases span twenty-six years and three technology waves. The success factor stayed structurally constant.
This is the recursive failure mode the substrate thesis names. Each wave expands the literacy and change management investment in response to the prior wave’s stalled outcomes, then attributes the new wave’s stalled outcomes to insufficient literacy and change management investment. The diagnostic loop runs continuously without engaging the variable that is actually determining the outcomes.
What Gartner Is and Is Not Saying
Read carefully, the Gartner webinar does not claim that AI literacy and change management are the only variables determining AI implementation success. The webinar says AI requires more training and change management effort than traditional technology. That observation is operationally accurate and worth taking seriously.
What the webinar does not say is the deeper structural claim — that the operating environment must be validated to support the load the AI is being asked to carry, before the literacy and change management investment can produce returns. That is the missing layer. Not because Gartner is wrong about the variables they name, but because the variables they name sit at the people-adoption layer rather than at the substrate layer beneath it. It is largely at the substrate alignment level that an initiative’s success in any technology era is determined.
The Gartner four-ingredients framework is the latest sophisticated version of a thirty-five-year-old answer. The framework will produce the same outcomes the prior wave’s frameworks produced — useful at the layer it operates on, structurally insufficient at the layer where the implementation actually fails or succeeds.
The Practitioner Implication
For procurement leaders, compliance officers, and AI governance professionals reading the Gartner webinar and deciding how to act on it, the practitioner implication is straightforward. The Gartner literacy and change management framework is necessary but not sufficient. Investing in it without first validating that the operating environment can support what the AI is being asked to do produces trained workforces operating inside unready substrate. The training will register as effective in the short term. The outcomes will register as disappointing in the medium term. The post-mortem will conclude that the literacy investment was insufficient, the culture work was incomplete, or the change resistance was unusually severe.
The Phase 0™ Diagnostic exists to ask the prior question — can our current operating environment actually support the load we are about to place on it through AI deployment? — before the literacy and change management investment is committed. The Phase 0™ work and the Gartner literacy work are not competing analytical frames. They operate at different layers. The diagnostic runs first. The literacy investment runs second. Substrate before training. Recognition before adoption. The sequence matters because reversing it produces the recursive failure mode the thirty-five-year pattern documents.
The Bottom Line
The Gartner webinar is correct that AI requires more training and change management effort than traditional technology. Organizations that ignore this observation will produce predictably worse outcomes than organizations that act on it.
The Gartner webinar is also operating on the same structural assumption that has produced the same explanatory framework — and the same disappointing outcomes — for thirty-five consecutive years. The framework is not wrong. It is incomplete. The missing piece is the substrate question that no major analyst firm has yet centered in its AI governance framework.
The recursive failure mode will continue until the diagnostic question is asked before the literacy investment is made. The pattern has been visible for three and a half decades. The Hansen Models™ archive has documented it contemporaneously since 1998. The Phase 0™ Diagnostic exists to break the recursion. The question is whether the current AI wave will be the one where the industry finally engages the prior question — or whether the same partial answer will produce the same outcomes for the seventh consecutive technology era.
This post was developed through the ARA™ RAM 2025™ multimodel validation framework. The historical pattern analysis draws on the Procurement Insights archive — nineteen years of contemporaneous documentation across more than 3,500 documents — and the Hansen Models™ Implementation Physics™ framework first introduced in 1998. The Phase 0™ Diagnostic, for organizations preparing for the August 2, 2026, EU AI Act enforcement deadline and beyond, is at hansenprocurement.com/where-does-your-organization-sit-right-now/.
Source: Gartner webinar — 4 Essential Ingredients for Successful AI Literacy and Change Management, recorded March 30, 2026.
Hansen Models™ · Implementation Physics™ · Compounding Technology Shadow Wave™ · Phase 0™ · Hansen Fit Score™ · Hansen Strand Commonality™ · AGR Index™ · RAM 2025™ · ARA™ · Hansen Deflator Formula™ · Hansen Optionality Loss Estimate™
-30-
Procurement Insights
The Same Answer for Thirty-Five Years: AI Literacy, Change Management, and the Substrate Question Gartner Is Not Asking
Posted on May 19, 2026
0
Procurement Insights · May 19, 2026
A Gartner webinar published in late March 2026 — 4 Essential Ingredients for Successful AI Literacy and Change Management — argues that AI adoption requires twenty-five percent more training effort and up to two hundred percent more change management effort than traditional technology. Underestimating that lift, Gartner observes, puts AI efforts at risk of stalled adoption and lost value. The webinar positions AI literacy and change management as the unlock for converting AI investment into enterprise value.
The framing is not wrong. AI literacy matters. Training matters. Change management matters. Organizations that ignore these variables produce predictably worse outcomes than organizations that invest in them. The Gartner observation is operationally true at the layer it operates on.
The problem is that the framing is also extraordinarily familiar. It is the same diagnosis the industry has offered for every technology wave since 1990. The vocabulary changes with each era. The structural answer does not. And the documented failure rate across all thirty-five years demonstrates that the answer — however thoroughly applied — is structurally insufficient to explain what determines implementation outcomes.
The Thirty-Five-Year Pattern
Walk back through the technology eras since 1990, and the pattern is striking enough that it deserves to be named.
In the early 1990s ERP and business process reengineering wave, the industry response to stalled implementations was train users, manage change, redesign processes. User adoption was treated as the primary variable. The empirical record from that era documents implementation failure rates in the sixty-five to seventy percent range despite extensive investment in literacy and change management programs.
In the 2000s e-procurement, portals, and early SaaS wave, the diagnosis shifted vocabulary but not structure. Supplier adoption, user adoption, process compliance became the explanatory frame. Failed deployments were attributed to insufficient stakeholder buy-in, inadequate training, or change resistance. Failure rates remained stuck in the same band. The Change Management Myth post in this archive — published in May 2007, in the first weeks of the Procurement Insights publication — named the change management framing as a myth at exactly the moment the broader industry was institutionalizing it as the answer. The Dangerous Supply Chain Myths series that followed extended the critique through additional documentation. The analytical position the Hansen Models™ framework holds in 2026 was being articulated contemporaneously when the e-procurement wave was at its peak.
The 2010s cloud and SaaS best-of-breed wave produced an even more emphatic version of the same answer. Digital culture, agile transformation, user experience became the analytical centerpiece. Organizations were urged to invest in cultural readiness, agile mindsets, and user-centered design. SaaS fragmentation accelerated. Shadow workflows multiplied. The failure rate held. The Procurement Insights archive was interrogating the corporate culture framing contemporaneously — the 2009 Six Sigma relevancy debate and the 2010 Survival of the Fittest post both documented that the culture variable was insufficient to explain technology adoption outcomes at the moment the broader industry was treating it as the load-bearing factor.
The late-2010s analytics, automation, and RPA wave doubled down. Data literacy, automation readiness, citizen developers entered the vocabulary. Bad process logic was automated rather than structurally corrected. The literacy investment was real and substantial. The realized outcomes did not improve materially.
The 2020-2023 pandemic digitization wave compressed the same diagnosis into emergency conditions. Digital maturity, collaboration culture, resilience became the framing. Workarounds that were designed as temporary became permanent operational substrate. The literacy and change management investment intensified. The structural condition deteriorated.
The current 2023-2025 generative AI, copilots, and agentic systems wave has produced the most extreme version of the historical pattern. AI literacy, prompt training, adoption, trust dominates the discourse. Gartner’s twenty-five percent and two hundred percent figures quantify the scale of the recommended investment. And the documented AI initiative failure rate — measured by MIT, BCG, McKinsey, Stanford HAI, and other major research institutions — sits in the same sixty-five to eighty percent band the industry has continued to cite across thirty-five years.
What the industry has never sustained is the empirical discipline that would have made its own diagnosis falsifiable. No major firm has run longitudinal research measuring whether organizations that invested heavily in change management produced measurably better implementation outcomes than organizations that did not. No major firm has correlated end-user literacy investment levels against realized technology value capture across a controlled cohort. No major firm has measured corporate culture maturity against adoption success in a way that would empirically validate the framing the firm itself is publishing. The prescription has been continuous. The validation has been absent. If the industry had been tracking the relationship between the three explanatory variables and the outcomes those variables are claimed to produce, the absence of improvement would have been visible as feedback, and the framing would have evolved to engage the structural variables actually determining the outcomes. That feedback loop has not operated. Which is why the partial diagnosis has persisted without empirical challenge for thirty-five consecutive years.
The 2025 post on the AI Change Management Myth in ProcureTech selection and implementation extended the 2007 Change Management Myth critique directly into the current AI wave — the same analytical position the framework held nineteen years earlier, applied to the technology era now in active deployment.
The Visual Argument
[GRAPHIC: thirty-five-year pattern bar chart showing per-era attention levels for end-user literacy, corporate culture investment, and change resistance management, with technology advancement and technology implementation outcome lines overlaid]
The pattern is not subtle when rendered visually. The industry’s analytical attention to end-user literacy, corporate culture, and change resistance has been consistently high across every era — and has intensified with each successive wave. The technological capability ceiling has risen continuously, with each wave more sophisticated than the last. The realized implementation outcomes have remained roughly flat, oscillating within a narrow band around sixty-five to eighty percent failure for thirty-five years.
The visual produces an argument that words alone struggle to land. If literacy, culture, and change resistance were the load-bearing variables determining implementation outcomes, the outcomes should have improved as the industry’s attention to those variables intensified. They did not. The flat outcome line in the presence of rising attention to the explanatory variables is the empirical evidence that those variables are not where the explanation lives.
Something else is doing the work — or rather, something else is failing to do the work. That something else is the operating environment the technology is being deployed into. The substrate.
What the Pattern Reveals
The substrate question is structurally different from the literacy question. Literacy asks do the people know how to use the technology. Substrate asks can the operating environment actually support what the technology is being asked to do. The two questions sound similar. They are not.
A workforce can be fully literate in a technology and still produce stalled adoption if the operating environment cannot support the technology’s requirements. Training people on ERP did not fix misaligned process assumptions in the underlying business operations. Training people on SaaS did not fix the fragmented operating models the SaaS portfolio was being asked to integrate. Training people on analytics did not fix the bad data lineage the analytics depended on. Training people on RPA did not fix the broken process logic the RPA was automating. Training people on AI agents will not fix the unclear authority, shadow workflows, incentive misalignment, and undocumented operating logic the agents are being asked to navigate. And no, this is not a technology integration problem between diverse systems — the SOA promise and the integration architectures that followed it confirm exactly this point. Successive waves of integration architecture have been deployed at enterprise scale across two decades. The substrate problems those architectures were positioned to resolve have persisted regardless.
The thirty-five-year pattern is not evidence that literacy and change management are unimportant. It is evidence that literacy and change management cannot resolve substrate inconsistency. Organizations that invest in literacy without addressing substrate produce trained workforces operating inside environments that cannot support what the training is teaching them to do. The training does not fail. The substrate does. But the failure gets attributed to the training, to the culture, to the change resistance — which produces the next wave’s expanded investment in the same explanatory variables and the same disappointing outcomes.
The cross-temporal evidence is documented in the Procurement Insights archive. The 1998 to 2007 to 2024 post — bridging the DND engagement that began in 1998, the Virginia eVA initiative, and the Colgate-Palmolive procurement transformation — demonstrates that the same underlying success factor operated across three technology eras and three different organizations. The variable that determined outcomes was not the era’s dominant literacy framework or change management methodology. It was whether the organization had asked the prior question about its operating environment before deploying the technology investment. The cases span twenty-six years and three technology waves. The success factor stayed structurally constant.
This is the recursive failure mode the substrate thesis names. Each wave expands the literacy and change management investment in response to the prior wave’s stalled outcomes, then attributes the new wave’s stalled outcomes to insufficient literacy and change management investment. The diagnostic loop runs continuously without engaging the variable that is actually determining the outcomes.
What Gartner Is and Is Not Saying
Read carefully, the Gartner webinar does not claim that AI literacy and change management are the only variables determining AI implementation success. The webinar says AI requires more training and change management effort than traditional technology. That observation is operationally accurate and worth taking seriously.
What the webinar does not say is the deeper structural claim — that the operating environment must be validated to support the load the AI is being asked to carry, before the literacy and change management investment can produce returns. That is the missing layer. Not because Gartner is wrong about the variables they name, but because the variables they name sit at the people-adoption layer rather than at the substrate layer beneath it. It is largely at the substrate alignment level that an initiative’s success in any technology era is determined.
The Gartner four-ingredients framework is the latest sophisticated version of a thirty-five-year-old answer. The framework will produce the same outcomes the prior wave’s frameworks produced — useful at the layer it operates on, structurally insufficient at the layer where the implementation actually fails or succeeds.
The Practitioner Implication
For procurement leaders, compliance officers, and AI governance professionals reading the Gartner webinar and deciding how to act on it, the practitioner implication is straightforward. The Gartner literacy and change management framework is necessary but not sufficient. Investing in it without first validating that the operating environment can support what the AI is being asked to do produces trained workforces operating inside unready substrate. The training will register as effective in the short term. The outcomes will register as disappointing in the medium term. The post-mortem will conclude that the literacy investment was insufficient, the culture work was incomplete, or the change resistance was unusually severe.
The Phase 0™ Diagnostic exists to ask the prior question — can our current operating environment actually support the load we are about to place on it through AI deployment? — before the literacy and change management investment is committed. The Phase 0™ work and the Gartner literacy work are not competing analytical frames. They operate at different layers. The diagnostic runs first. The literacy investment runs second. Substrate before training. Recognition before adoption. The sequence matters because reversing it produces the recursive failure mode the thirty-five-year pattern documents.
The Bottom Line
The Gartner webinar is correct that AI requires more training and change management effort than traditional technology. Organizations that ignore this observation will produce predictably worse outcomes than organizations that act on it.
The Gartner webinar is also operating on the same structural assumption that has produced the same explanatory framework — and the same disappointing outcomes — for thirty-five consecutive years. The framework is not wrong. It is incomplete. The missing piece is the substrate question that no major analyst firm has yet centered in its AI governance framework.
The recursive failure mode will continue until the diagnostic question is asked before the literacy investment is made. The pattern has been visible for three and a half decades. The Hansen Models™ archive has documented it contemporaneously since 1998. The Phase 0™ Diagnostic exists to break the recursion. The question is whether the current AI wave will be the one where the industry finally engages the prior question — or whether the same partial answer will produce the same outcomes for the seventh consecutive technology era.
This post was developed through the ARA™ RAM 2025™ multimodel validation framework. The historical pattern analysis draws on the Procurement Insights archive — nineteen years of contemporaneous documentation across more than 3,500 documents — and the Hansen Models™ Implementation Physics™ framework first introduced in 1998. The Phase 0™ Diagnostic, for organizations preparing for the August 2, 2026, EU AI Act enforcement deadline and beyond, is at hansenprocurement.com/where-does-your-organization-sit-right-now/.
Source: Gartner webinar — 4 Essential Ingredients for Successful AI Literacy and Change Management, recorded March 30, 2026.
Hansen Models™ · Implementation Physics™ · Compounding Technology Shadow Wave™ · Phase 0™ · Hansen Fit Score™ · Hansen Strand Commonality™ · AGR Index™ · RAM 2025™ · ARA™ · Hansen Deflator Formula™ · Hansen Optionality Loss Estimate™
-30-
Share this:
Related