Why Major Consulting Firms Are Failing the AI Implementation Test

The Consulting Industry Faces a Reckoning

The consulting industry faces an uncomfortable reckoning. After years of positioning themselves as indispensable guides through digital transformation, firms like PwC, Deloitte, McKinsey, and Boston Consulting Group now confront a harsh reality: when it comes to generative AI, their traditional playbooks are failing.

The Wall Street Journal's recent investigation into the consulting industry's AI struggles reveals a pattern that should concern both consultants and their clients. Despite collective investments exceeding billions of dollars and marketing campaigns promising results, major consulting firms are discovering that expertise in cloud migration and ERP implementations does not translate into AI deployment competence.

This disconnect matters because it signals a fundamental shift in the value proposition of management consulting and raises critical questions about how organizations should approach transformative technology adoption.

The Credibility Crisis

The article by Isabelle Bousquette and Mark Maurer documents what many technology executives have experienced firsthand but hesitated to articulate publicly: consulting firms often lack meaningful advantages over internal teams when deploying cutting-edge AI applications. Greg Meyers, chief digital and technology officer at Bristol-Myers Squibb, crystallized this reality with brutal clarity: hiring a Big Four partner to implement Google's Gemini or Anthropic's Claude offers no more value than hiring a college student experimenting with the same tools.

This assessment is not hyperbole but a measured evaluation from a senior technology executive at a Fortune 500 pharmaceutical company. Bristol-Myers Squibb terminated a yearlong consulting engagement focused on using generative AI to produce educational content for physicians, choosing instead to pursue the initiative internally.

The implications extend beyond a single failed engagement. When Dave Williams, Chief Information and Digital Officer at Merck, states that consultants are "learning on our dime," he articulates a fundamental breach of the consulting value proposition. Clients engage external advisors expecting expertise that exceeds internal capabilities. When consultants arrive with comparable or inferior knowledge, they transform from value creators into expensive overhead.

Magesh Sarma, chief information and strategy officer at AmeriSave Mortgage, experienced this firsthand. His organization discovered that consultants "really also had no idea how to do these things" and performed no better than internal teams would have. Even more damaging, Pat Petitti, CEO of Catalant, reports hearing repeatedly from clients who spent $20 million only to receive "a very long report on where AI is going without any real practical application."

These are not isolated complaints but a pattern indicating systemic problems in how consulting firms approached the AI opportunity.

The Root Causes of Consultant Underperformance

Several factors explain why consulting firms are struggling with AI implementation, and understanding these causes offers insights into both the limitations of traditional consulting models and the unique challenges of generative AI adoption.

First, generative AI represents a genuinely novel technology category. Unlike previous enterprise technology waves where consultants could develop expertise through early adopter engagements and then replicate successful approaches, generative AI remains experimental even for technology companies building the underlying models. The technology evolves weekly, best practices remain unclear, and successful proof-of-concept projects often fail when scaled.

Consulting firms built their business models on knowledge transfer and implementation of established practices. They excel at taking proven approaches and adapting them to new organizational contexts. Generative AI undermines this model because proven approaches barely exist. As Michael Mische, a former KPMG principal now teaching at the University of Southern California, observes, consulting firms occupy a "position of great vulnerability" because they hired too slowly and now lag behind rather than lead AI adoption.

Second, the consulting industry faces a more capable client base than in previous technology cycles. Tilak Mandadi, CVS Health executive vice president, explicitly stated that his internal team is "best equipped to come up with those use cases" for AI in healthcare's complex environment. His organization deliberately chose not to "hire a bunch of consultants to tell us what to do with GenAI."

This represents a significant shift. Modern enterprises employ sophisticated technology leaders with deep technical backgrounds. These executives recognize that consultants cannot offer meaningful advantages in areas where genuine expertise remains scarce. The democratization of AI tools through accessible APIs and improved documentation means that technical teams can experiment and learn as effectively as external consultants.

Third, consulting firms may have over-invested in marketing relative to capability building. PwC's campaign claiming "Nobody makes AI work for your business like PwC" and promising "We don't just bring promises. We bring results" created expectations that delivery teams could not meet. This gap between marketing rhetoric and implementation reality damages client relationships and erodes trust.

The revenue numbers tell an interesting story. While Gartner estimates global spending on generative AI consulting reached $3.75 billion in 2024, up from $1.34 billion in 2023, this growth masks underlying problems. KPMG reported $1.4 billion in potential AI-related projects as of July, up from $500 million two years earlier, but potential pipeline differs from delivered value. Accenture's generative AI bookings increased $100 million quarter-over-quarter in their most recent period, down from $200 million increases in previous quarters, suggesting momentum may be slowing.

The Scaling Problem

One of the most significant criticisms emerging from client experiences involves the inability of consulting firms to scale AI implementations beyond proof-of-concept stages. This represents a particularly damaging failure because scaling is precisely where consultants should excel.

Consulting firms built their reputations on implementation expertise. They understand organizational change management, have experience navigating enterprise complexity, and can deploy large teams to execute at scale. Yet multiple clients report that consultants successfully build proofs of concept but fail to scale them across the business to create meaningful value.

This failure suggests that the challenges of AI scaling differ fundamentally from previous technology implementations. Scaling generative AI requires not just technical deployment but also continuous refinement of prompts, management of model drift, integration with constantly evolving APIs, and navigation of emerging governance and compliance requirements. These challenges demand ongoing technical sophistication rather than the project-based implementation approach that consulting firms typically employ.

The scaling problem also reveals tensions in how consulting engagements are structured. Consultants typically work on fixed-scope projects with defined deliverables and timelines. AI implementation requires iterative experimentation, continuous learning, and willingness to abandon approaches that prove ineffective. These contrasting models create misalignment between how consultants sell and deliver work and what AI implementation actually requires.

The In-House Advantage

The trend toward internal AI capability development represents more than consultant underperformance. It reflects a strategic calculation by enterprises that AI competence constitutes a competitive advantage that should be built rather than bought.

When CVS Health, Bristol-Myers Squibb, and Merck choose internal development over consultant engagements, they signal that AI expertise is core to their business strategy rather than a commodity service. This mirrors how leading technology companies approached earlier innovations. Amazon did not hire consultants to build AWS. Google did not outsource search algorithm development. Netflix built its recommendation engine internally.

If AI delivers the transformative impact that both consultants and technology vendors promise, then organizations that develop genuine internal expertise will outperform those that rely on external advisors. This creates a challenging dynamic for consulting firms: the more transformative and valuable AI becomes, the less likely sophisticated clients will outsource it.

The counterargument holds that most organizations lack the resources and talent to build world-class AI capabilities internally. Not every company can recruit and retain the specialized talent that AI development requires. For these organizations, consulting firms should offer valuable support.

However, the client complaints documented in the Journal article suggest that consulting firms themselves lack this specialized talent. If consultants are learning on client engagements rather than bringing established expertise, clients might reasonably conclude they are better off learning themselves while retaining the knowledge internally.

What Consulting Firms Get Right

Balanced analysis requires acknowledging where consulting firms do provide value in AI initiatives. Several clients noted that consultants offer useful cross-industry perspectives, showing what works in other sectors and providing additional implementation capacity.

These contributions matter, particularly for mid-sized organizations that need temporary expertise without permanent headcount. Consulting firms can provide pattern recognition across multiple implementations, helping clients avoid obvious mistakes and accelerate learning curves.

The consulting firms themselves maintain they are seeing strong demand and delivering client value. KPMG, EY, McKinsey, Bain, and BCG all told the Journal that AI service demand is increasing. McKinsey Senior Partner Eric Kutcher continues telling CEOs that effective generative AI leverage can double share prices within five years, suggesting sustained confidence in the technology's potential.

This optimism may prove justified over longer time horizons. Fiona Czerniawska, CEO of Source Global Research, which tracks the consulting industry, argues that while the current generation of CIOs may be skeptical about consultant AI capabilities, a "second wave" will emerge in four to five years when the technology matures and becomes more predictable. At that point, AI implementation may become the kind of reliable, process-driven work where consulting firms traditionally excel.

Czerniawska's analysis includes an important insight: "The problem at the moment is that consulting firms have tried to put themselves at the cutting edge, and it's not really where they belong." This suggests consulting firms made a strategic error by positioning themselves as AI innovators rather than AI implementors.

Implications for Business Leaders

The consulting industry's AI struggles offer several lessons for business leaders navigating their own AI transformations.

• Treat consultant expertise claims with appropriate skepticism. When evaluating consulting proposals for AI initiatives, demand specific evidence of successful implementations, not just proof-of-concept projects. Ask detailed questions about the actual team members who will work on your engagement and their hands-on experience with the specific AI tools and use cases you want to pursue.
• Consider a hybrid approach that combines internal capability building with selective external support. Rather than outsourcing AI strategy and implementation to consultants, organizations might engage them for specific components where they offer genuine advantages, such as change management or process redesign, while building core AI technical capabilities internally.
• Recognize that AI implementation timelines and success metrics may differ from previous technology initiatives. The iterative, experimental nature of AI deployment means that traditional project management approaches may prove inadequate. Organizations need to build internal cultures that embrace experimentation, tolerate failure, and learn continuously.
• Invest in internal talent development. If the consulting industry's talent gap reflects broader market shortages, organizations that successfully recruit, train, and retain AI-capable employees will gain significant advantages. This may require rethinking compensation structures, career paths, and organizational cultures to compete for scarce technical talent.
• Be realistic about AI's current limitations. Both consultants and technology vendors have incentives to oversell AI's near-term capabilities. The gap between proof-of-concept and scaled production deployment remains substantial. Organizations should pursue AI initiatives with appropriate expectations, focusing on specific use cases with clear business value rather than seeking wholesale transformation.

The Broader Context

The consulting industry's AI challenges reflect broader questions about how organizations should approach transformative technology adoption. The traditional model of hiring external experts to guide implementation assumes that expertise exists and can be transferred. This model works well for mature technologies with established best practices.

Generative AI does not fit this pattern. The technology remains immature, best practices continue evolving, and even the companies building foundational models cannot predict how capabilities will develop. In this environment, the advantages of external consultants diminish while the value of internal experimentation increases.

This situation parallels earlier technology transitions where initial hype exceeded delivery capabilities. The first wave of cloud computing, big data analytics, and mobile application development all saw similar patterns: aggressive vendor and consultant promises followed by implementation challenges and client disappointment.

What distinguishes the current AI wave is the pace of underlying technology development. Cloud computing standards stabilized relatively quickly. Generative AI capabilities continue advancing rapidly, with new models, techniques, and applications emerging continuously. This creates ongoing challenges for any organization trying to build stable expertise, whether consulting firms or internal teams.

Looking Forward

The consulting industry will likely adapt to AI implementation challenges, but the adaptation may look different than their responses to previous technology waves. Rather than becoming AI implementation leaders, consulting firms may focus on helping organizations with the organizational and strategic challenges that AI creates: workforce transformation, business model innovation, governance frameworks, and ethical considerations.

These areas play to consulting firms' traditional strengths in strategy and organizational change while avoiding direct competition with technology companies and internal technical teams in implementation work. A consulting firm may not offer advantages in fine-tuning large language models, but it can provide valuable guidance on restructuring customer service operations to leverage AI-powered tools effectively.

The four-to-five-year timeline that industry observers cite for consulting firms to develop effective AI playbooks seems realistic. By that point, generative AI may transition from cutting-edge experimentation to established enterprise technology with proven implementation patterns. Consulting firms can then apply their traditional capabilities to helping mainstream organizations adopt what early movers have already figured out.

However, this timeline also means that leading organizations pursuing AI advantages today should not expect consultants to provide those advantages. The companies that will lead their industries in AI capabilities five years from now are those building internal expertise today through direct experimentation rather than outsourced consulting engagements.

Conclusion

The consulting industry's struggles with AI implementation reveal important truths about both the consulting business model and the nature of transformative technology adoption. Consulting firms built successful practices by developing expertise in established technologies and processes, then helping clients implement proven approaches. This model breaks down when the technology itself remains experimental and expertise is scarce everywhere.

For business leaders, the key insight is that external consultants cannot substitute for internal capability building when technologies are genuinely novel and strategically important. Organizations serious about AI transformation must invest in developing internal expertise, even if this requires higher upfront costs and steeper learning curves.

The $20 million consulting engagements that produce reports rather than implementations represent not just wasted money but missed opportunities to build organizational capabilities. The companies abandoning consultant-led AI projects to pursue internal development are making strategically sound decisions that will likely serve them well as AI capabilities mature.

Consulting firms will eventually find their place in the AI ecosystem, but that place may be more modest than their initial ambitions suggested. The cutting edge of technology innovation is not where traditional consulting firms belong. Their value lies in helping organizations implement proven practices at scale, manage complex organizational change, and navigate strategic decisions.

As AI transitions from experimental technology to enterprise standard, consulting firms will have opportunities to apply these capabilities. But for now, the AI boom is indeed leaving consultants behind, and business leaders should plan accordingly. The path to AI advantage runs through internal capability building, direct experimentation, and willingness to learn through doing rather than through hiring external advisors who are learning alongside you.

The organizations that recognize this reality and act on it will develop sustainable competitive advantages. Those that continue relying on consultants to deliver AI transformation risk falling behind competitors who are building genuine internal expertise. In the race to leverage AI effectively, there are no shortcuts, and expensive consulting engagements often represent detours rather than acceleration.