Why the consulting industry’s race to “own the AI stack” ignores 17 years of documented evidence that equation-based transformation has a 70-80% failure rate.
On October 4th, 2025, Hugo Raaijmakers, Global Head of AI Innovation at PA Consulting, posted on LinkedIn about the explosive growth of AI agents in consulting. His message was clear: “The firms that move from slides to systems — that can build, orchestrate, and scale AI agents — will lead the next era of the industry.“
The post included a CB Insights graphic showing that since 2023, the world’s largest consultancies—Deloitte, Accenture, BCG, KPMG, and EY—have executed 100+ AI-related partnerships, acquisitions, and investments.
Hugo’s conclusion? “Advisory alone won’t cut it.”
He’s half right.
But there’s a critical problem: the consulting industry is repeating the exact same technology-first implementation mistake that has produced 70-80% failure rates for the past three decades—from ERP in the 1990s, to e-procurement in the 2000s, to digital transformation in the 2010s, and now AI/Agentic transformation in the 2020s.
I know this pattern intimately. I’ve been documenting it since 2007 in the Procurement Insights archives. But the evidence goes back even further.
2008: I Warned About This Exact Pattern—With SAP
In 2008, I published a white paper for the CATA Alliance titled “SAP Procurement for Public Sector.” The paper analyzed why SAP implementations consistently failed despite the vendor’s technical sophistication and market dominance.
The findings were clear:
Technology-First (Equation-Based) Implementations Failed Because:
- Organizations led with technology capabilities instead of process understanding
- People and processes were bent around the tool rather than the reverse
- Customizations and workarounds became the norm, creating escalating costs
- Stakeholder resistance increased due to inherited systems they wouldn’t have chosen
- Success metrics focused on deployment, not operational outcomes
The white paper documented high-profile failures:
- Hershey Foods: $112M over 30 months, order fulfillment process broken
- HP: Lost $400M in revenue from failed SAP rollout
- Cadbury Schweppes: £12M hit from SAP supply chain project failure
- FoxMeyer Drug: Bankruptcy, $500M lawsuit against SAP
The brutal irony? These failures weren’t caused by bad technology. SAP’s software was (and is) technically sophisticated. The failures happened because consultancies led with technology before assessing organizational readiness.
Sound familiar?
2025: Consultancies Are Making the Same Mistake—Now With AI
Hugo’s LinkedIn post and the CB Insights graphic reveal that the consulting industry hasn’t learned from 17 years of documented failures.
Let’s examine what his post reveals:
Hugo’s Four Imperatives for AI Success:
- Orchestrate the fragmented AI agent stack
- Unlock proprietary data as fuel for intelligent agents
- Turn services into scalable AI products
- Build the human-AI workforce
Notice what’s missing?
Not a single mention of:
- ❌ Organizational readiness assessment
- ❌ People-first implementation methodology
- ❌ Process understanding before technology selection
- ❌ Behavioral alignment evaluation
- ❌ Foundation-building before tool deployment
This is pure equation-based thinking: Technology → Process → People
The same approach that produced:
- 70-80% ERP failure rates (1990s-2000s)
- 70-80% e-procurement failure rates (2000s-2010s)
- 70-80% digital transformation failure rates (2010s-2020s)
And now, the same approach being applied to AI agents in 2025.
The Consultancies’ Own Implementation Failures
Here’s the devastating validation: The consultancies acquiring these 100+ AI companies have struggled with their own technology implementations.
From my 2008 white paper:
Hewlett-Packard (HP)
HP, a supposed SAP integration expert, lost $400 million in revenue from a failed SAP rollout. As RedMonk analyst James Governor observed:
“HP is trying to build an application management business to rival IBM’s. What better case study in proving your R/3 and Netweaver capability than showing everyone how to merge two SAP systems? Who would want to go to HP now for large scale SAP integration? The CEO just publicly said HP can’t effectively manage such a project.“
If a technology company that positioned itself as an SAP integrator couldn’t succeed with its own implementation, why would clients trust them to implement AI agents?
Accenture’s Pattern
My white paper noted that even with Accenture’s implementation framework for Canada Post, success required the client taking full ownership and responsibility—not relying on the consultant’s methodology.
The Canada Post team told me directly: “Rather than relying on the software vendor, we took control of the program from the beginning using a collaborative process strategy.”
Translation: Success happened when the client ignored the technology-first approach and led with people/process understanding.
The 100+ AI Acquisitions: A Map of Future Failures
The CB Insights graphic showing Deloitte, Accenture, BCG, KPMG, and EY’s 100+ AI partnerships reveals a predictable pattern:
Each consultancy is acquiring point solutions across:
- Data as differentiator
- Embedding AI
- Workforce transformation
- Race to own the stack
This creates the exact “band-aid on brittle foundation” problem I warned about in 2008:
- Fragmentation: 100+ disconnected AI tools requiring integration
- No canonical architecture: Each client gets a Frankenstein stack
- Missing behavioral hooks: Technology deployed without governance/readiness
- Equation-based methodology: Technology → Process → People
The result? In 2028-2030, we’ll have the same conversations:
- “Why did our AI transformation fail?”
- “Change management was the problem”
- “We need better talent”
Reality: They led with technology again, ignoring the people-first foundation required for success.
The Hansen Model Alternative: Agent-Based Transformation
For 27 years—since building the RAM (Relational Acquisition Model) platform in 1998—I’ve championed a fundamentally different approach:
Agent-Based (People-First) Model:
People → Process → Technology
- Lead with people and process understanding (not technology capabilities)
- Technology becomes a problem-solving tool (not the driver)
- Assess organizational readiness BEFORE deployment (Hansen Fit Score)
- Build behavioral integration hooks (canonical architecture, not silos)
- Measure success by operational outcomes (not deployments)
Validation:
- 97.3% accuracy over 7 years (RAM 1998-2005)
- $12M acquisition by public company (market validation)
- 27 years of continued application (vs. 70-80% industry failure rate)
The Comparison: Equation-Based vs. Agent-Based
[HTML GRAPHIC TO BE INSERTED HERE]
What Hugo’s Post Misses: The Systemic Thinking Question
On the same day Hugo posted about AI agents, Sheena Smith asked a more fundamental question at DPW Amsterdam:
“How are we looking at systemic thinking now, to blow up traditional processes and create new operating models that fit today’s world? How are we developing the talent who knows how to do that?“
Hugo’s answer (implicit): Deploy 12,000 AI agents, acquire 100+ tech companies, turn services into products.
The problem? You cannot develop systemic thinking by replacing human judgment with AI agents.
Systemic thinking develops through:
- Deep domain expertise (my “time of day” procurement insight from 1998)
- Agent-based problem-solving (people/process first, technology second)
- Pattern recognition across disparate systems (strand commonality)
- Curiosity to challenge assumptions (behavioral hooks framework)
Hugo’s model eliminates the exact conditions that create systemic thinkers.
The Critical Questions for 2025
If you’re a CPO, CIO, or executive evaluating AI agent offerings from consultancies, ask:
About Organizational Readiness:
- What is our Hansen Fit Score (organizational readiness for AI)?
- Do we have the people/process foundation to absorb this technology?
- What behavioral hooks prevent this from becoming another silo?
About the Consultancy:
- What was your own implementation success rate with similar technologies?
- Can you show validated operational outcomes (not just deployments)?
- How do you assess client readiness before recommending technology?
About the Approach:
- Are you leading technology-first (equation-based) or people-first (agent-based)?
- What happens if our readiness score is below threshold—do you still sell us the solution?
- How do you build capability before deploying technology?
If the consultancy cannot answer these questions convincingly, you’re about to become another failure statistic.
The Bottom Line
Hugo Raaijmakers is right that advisory alone won’t cut it.
But he’s wrong about the solution.
The future of consulting isn’t about:
- Acquiring 100+ AI companies
- Building autonomous agent stacks
- Turning services into scalable products
The future of consulting is about:
- Assessing readiness before recommending technology
- Building people/process capability before deployment
- Creating behavioral integration architecture (not point solution silos)
- Measuring success by operational outcomes (not revenue from AI agent sales)
The consulting industry has a choice:
- Repeat the pattern: Lead with technology, achieve 70-80% failure rates, blame “change management”
- Learn from 27 years of evidence: Lead with people/process, build foundations, achieve sustainable transformation
My 2008 SAP white paper warned about this exact pattern.
Hugo’s 2025 LinkedIn post proves the industry hasn’t learned.
The 100+ AI acquisitions aren’t a sign of progress—they’re a map of future implementation failures.
Unless consultancies fundamentally change their approach from equation-based (technology-first) to agent-based (people-first), we’ll be having this same conversation in 2030.
The technology isn’t the problem. It never was.
The problem is leading with technology before building the foundation to absorb it.
30
BONUS COVERAGE – THE ABOVE IN PICTURES
Procurement Insights
100+ AI Acquisitions Since 2023: Consultancies Repeat the Same Technology-First Mistake (Again)
Posted on October 4, 2025
0
Why the consulting industry’s race to “own the AI stack” ignores 17 years of documented evidence that equation-based transformation has a 70-80% failure rate.
On October 4th, 2025, Hugo Raaijmakers, Global Head of AI Innovation at PA Consulting, posted on LinkedIn about the explosive growth of AI agents in consulting. His message was clear: “The firms that move from slides to systems — that can build, orchestrate, and scale AI agents — will lead the next era of the industry.“
The post included a CB Insights graphic showing that since 2023, the world’s largest consultancies—Deloitte, Accenture, BCG, KPMG, and EY—have executed 100+ AI-related partnerships, acquisitions, and investments.
Hugo’s conclusion? “Advisory alone won’t cut it.”
He’s half right.
But there’s a critical problem: the consulting industry is repeating the exact same technology-first implementation mistake that has produced 70-80% failure rates for the past three decades—from ERP in the 1990s, to e-procurement in the 2000s, to digital transformation in the 2010s, and now AI/Agentic transformation in the 2020s.
I know this pattern intimately. I’ve been documenting it since 2007 in the Procurement Insights archives. But the evidence goes back even further.
2008: I Warned About This Exact Pattern—With SAP
In 2008, I published a white paper for the CATA Alliance titled “SAP Procurement for Public Sector.” The paper analyzed why SAP implementations consistently failed despite the vendor’s technical sophistication and market dominance.
The findings were clear:
Technology-First (Equation-Based) Implementations Failed Because:
The white paper documented high-profile failures:
The brutal irony? These failures weren’t caused by bad technology. SAP’s software was (and is) technically sophisticated. The failures happened because consultancies led with technology before assessing organizational readiness.
Sound familiar?
2025: Consultancies Are Making the Same Mistake—Now With AI
Hugo’s LinkedIn post and the CB Insights graphic reveal that the consulting industry hasn’t learned from 17 years of documented failures.
Let’s examine what his post reveals:
Hugo’s Four Imperatives for AI Success:
Notice what’s missing?
Not a single mention of:
This is pure equation-based thinking: Technology → Process → People
The same approach that produced:
And now, the same approach being applied to AI agents in 2025.
The Consultancies’ Own Implementation Failures
Here’s the devastating validation: The consultancies acquiring these 100+ AI companies have struggled with their own technology implementations.
From my 2008 white paper:
Hewlett-Packard (HP)
HP, a supposed SAP integration expert, lost $400 million in revenue from a failed SAP rollout. As RedMonk analyst James Governor observed:
If a technology company that positioned itself as an SAP integrator couldn’t succeed with its own implementation, why would clients trust them to implement AI agents?
Accenture’s Pattern
My white paper noted that even with Accenture’s implementation framework for Canada Post, success required the client taking full ownership and responsibility—not relying on the consultant’s methodology.
The Canada Post team told me directly: “Rather than relying on the software vendor, we took control of the program from the beginning using a collaborative process strategy.”
Translation: Success happened when the client ignored the technology-first approach and led with people/process understanding.
The 100+ AI Acquisitions: A Map of Future Failures
The CB Insights graphic showing Deloitte, Accenture, BCG, KPMG, and EY’s 100+ AI partnerships reveals a predictable pattern:
Each consultancy is acquiring point solutions across:
This creates the exact “band-aid on brittle foundation” problem I warned about in 2008:
The result? In 2028-2030, we’ll have the same conversations:
Reality: They led with technology again, ignoring the people-first foundation required for success.
The Hansen Model Alternative: Agent-Based Transformation
For 27 years—since building the RAM (Relational Acquisition Model) platform in 1998—I’ve championed a fundamentally different approach:
Agent-Based (People-First) Model:
People → Process → Technology
Validation:
The Comparison: Equation-Based vs. Agent-Based
[HTML GRAPHIC TO BE INSERTED HERE]
What Hugo’s Post Misses: The Systemic Thinking Question
On the same day Hugo posted about AI agents, Sheena Smith asked a more fundamental question at DPW Amsterdam:
Hugo’s answer (implicit): Deploy 12,000 AI agents, acquire 100+ tech companies, turn services into products.
The problem? You cannot develop systemic thinking by replacing human judgment with AI agents.
Systemic thinking develops through:
Hugo’s model eliminates the exact conditions that create systemic thinkers.
The Critical Questions for 2025
If you’re a CPO, CIO, or executive evaluating AI agent offerings from consultancies, ask:
About Organizational Readiness:
About the Consultancy:
About the Approach:
If the consultancy cannot answer these questions convincingly, you’re about to become another failure statistic.
The Bottom Line
Hugo Raaijmakers is right that advisory alone won’t cut it.
But he’s wrong about the solution.
The future of consulting isn’t about:
The future of consulting is about:
The consulting industry has a choice:
My 2008 SAP white paper warned about this exact pattern.
Hugo’s 2025 LinkedIn post proves the industry hasn’t learned.
The 100+ AI acquisitions aren’t a sign of progress—they’re a map of future implementation failures.
Unless consultancies fundamentally change their approach from equation-based (technology-first) to agent-based (people-first), we’ll be having this same conversation in 2030.
The technology isn’t the problem. It never was.
The problem is leading with technology before building the foundation to absorb it.
30
BONUS COVERAGE – THE ABOVE IN PICTURES
Share this:
Related