The Consulting Firms Are Eating Their Own Lunch
How AI advisory transformed from zero to 40% of McKinsey's business in three years
What happens when the experts selling AI transformation deploy 12,000 agents to automate their own work
McKinsey’s AI practice generated approximately $6 billion in revenue in 2024, representing 40% of their total business. Three years ago, that number was effectively zero. Boston Consulting Group went from no AI revenue to $2.7 billion—20% of their $13.5 billion total—in the same timeframe. Accenture booked $3.6 billion in generative AI consulting specifically, a ninefold increase from the prior year.
These aren’t just impressive growth numbers. They’re evidence that something structural has shifted in how businesses buy strategy and how consultants sell it.
The market that appeared overnight
The global AI consulting market was valued at $8.75 billion in 2024. Projections put it at $49 billion by 2032, growing at 24% annually. That’s faster than cloud infrastructure services, faster than martech adoption, faster than most enterprise software categories.
But focusing on the total addressable market misses what’s actually happening. AI consulting isn’t a new category alongside strategy, operations, and technology consulting. It’s becoming embedded in all three. McKinsey expects eventually 40-50% of their engagements to involve AI in some capacity—not because clients specifically request AI consulting, but because it’s increasingly difficult to solve business problems without it.
That’s the shift consultants understand and most companies are still processing. AI isn’t a workstream. It’s the operating environment.
Why this happened so fast
Traditional enterprise technology adoption follows a predictable pattern. Pilot phase, proof of concept, limited rollout, scaled deployment. It takes 18-36 months from initial interest to meaningful budget allocation. AI consulting skipped several steps.
Part of that acceleration came from the obvious demo effect of ChatGPT. Executives saw capabilities that would have seemed science fiction 24 months earlier working in a consumer app. The gap between “that’s interesting” and “how do we use this” collapsed to weeks.
But the real driver is simpler: companies realized they didn’t know what they didn’t know. They couldn’t even frame the right questions about AI implementation because the technology was advancing faster than their internal understanding. That created immediate demand for advisory services.
Consider the economics. A mid-size financial services company budgets $2 million for an AI strategy engagement with BCG. Six months later, they have a deployment roadmap, identified use cases, quantified ROI expectations, and begun hiring for an AI center of excellence. That $2 million consulting spend unlocks $15-20 million in internal investment—and potentially millions more in operational savings or revenue upside.
From the consulting firm’s perspective, those economics are better than traditional strategy work. The engagements are shorter, the value is more measurable, and the follow-on work is almost guaranteed.
The talent arbitrage
McKinsey hired 1,000 additional staff specifically for AI services in 2024. BCG did similar. Accenture announced plans to reach 80,000 data and AI professionals by 2026, up from about 40,000 currently. These aren’t incremental hires to staff existing work—they’re building entirely new capabilities.
Here’s what makes that interesting: many of these hires are coming from tech companies, not other consulting firms. Data scientists and machine learning engineers who spent careers at Google, Meta, and Amazon are shifting to consulting because the problems are more varied and, frankly, because consulting firms can now pay competitive wages given the revenue these practices generate.
This creates a strange dynamic. The expertise companies need is migrating from tech companies that built AI capabilities to consulting firms that deploy them. If you’re a Fortune 500 company trying to hire in-house AI talent, you’re competing with offers from McKinsey and BCG that include variety, prestige, and increasingly comparable compensation.
The implication: companies that assumed they’d build internal AI capabilities are instead becoming dependent on external advisors for longer than planned.
What this means for marketing organizations
Marketing was always going to be an early AI adopter—the use cases are obvious, the ROI is measurable, and the data already exists. But the way marketing organizations are approaching AI implementation reflects broader consulting dynamics.
Finance and banking lead AI consulting demand, capturing 22% of the market. Marketing and advertising is smaller in absolute terms but growing faster—roughly 25-30% annually versus 23% overall market growth. That’s because marketing leaders face a unique challenge: they need AI capabilities before they understand AI strategy.
Consider email automation. According to recent surveys, 70% of marketing teams believe AI-driven hyper-personalization will significantly impact email performance. But only 17% have taken steps to address explainability and bias risks in their AI tools. That gap between adoption intention and implementation readiness is exactly where consulting firms are finding opportunity.
The engagement pattern is telling. Marketing-specific AI consulting starts with use case identification, moves to vendor selection, and increasingly extends to change management and performance optimization. The average engagement is 4-6 months and runs $500,000 to $2 million for mid-market companies. Enterprise deals run higher.
For consulting firms, marketing AI is attractive because it sits at the intersection of technology implementation and business transformation. It requires technical expertise they’re building and strategic thinking that’s always been their differentiator.
The consolidation you’re not noticing
Deloitte executed 700+ AI projects in 2024, contributing substantially to its $67 billion in revenue. PwC has committed $1 billion to generative AI capabilities. EY trained 70% of its global workforce in AI basics. These aren’t pilot programs—they’re fundamental capability investments.
What’s happening is that consulting firms are becoming AI deployment platforms. The engagements aren’t “help us think about AI” anymore. They’re “implement AI for these specific functions and train our people to maintain it.”
McKinsey’s approach is instructive: they deployed 12,000 internal AI agents—tools that automate specific consulting tasks like data analysis, presentation creation, and research synthesis. Those agents now do work that previously required teams of 14 people, accomplishing it with teams of 2-3. McKinsey isn’t just advising on AI; they’re demonstrating it by reorganizing their own workforce.
That proof of concept is powerful. When a CMO asks “how do we know this works,” the consultant can literally show their internal tools and workflows. It’s not theoretical. It’s operational.
The questions nobody’s asking
The AI consulting market growing from $8.75 billion to $49 billion in eight years means hundreds of billions in client-side investment. Companies don’t spend a quarter-million on consulting fees unless they’re planning to spend millions implementing the recommendations.
Here’s the uncomfortable math: if AI consulting is $9 billion in 2024 and the typical consulting-to-implementation ratio is 1:7, that implies $63 billion in AI-related enterprise technology spend influenced by consulting recommendations. By 2032, if consulting hits $49 billion, you’re looking at $340 billion in implementation spend.
That’s not happening in a vacuum. It’s reallocating budget from somewhere. Traditional software licenses, legacy system maintenance, and yes, human headcount.
The consulting firms are careful about this messaging. They emphasize augmentation, not replacement. But their internal operations tell a different story. When McKinsey can accomplish with AI agents and 2-3 people what previously required 14, that’s not augmentation. That’s substitution.
What to actually do about this
First, recognize that the decision isn’t whether to engage AI consulting firms, but when and for what. The capability gap between leading-edge AI implementation and typical enterprise understanding is too large for most companies to bridge independently.
Second, be specific about scope. AI strategy engagements that try to “assess readiness across the enterprise” tend to produce expensive reports. AI implementation engagements focused on specific use cases with defined success metrics produce actual capability.
Third, demand transfer of knowledge, not just recommendations. The best AI consulting engagements leave your team able to continue the work independently. The worst ones create dependency.
For marketing leaders specifically, the practical question is where AI capability needs to live. Some functions—like predictive analytics for customer lifetime value or AI-driven content generation—benefit from consulting-led implementation and internal ownership. Others, like real-time bidding optimization or programmatic creative, might stay with vendors or partners permanently.
The trap is assuming you can figure all this out internally before engaging outside expertise. By the time you’ve clarity on the questions, the competitive environment has shifted. The companies succeeding aren’t those with perfect AI strategies. They’re the ones shipping imperfect AI implementations quickly and iterating.
The forecast nobody wants
AI consulting growing at 24% annually while overall consulting markets grow at 3-5% means a fundamental reweighting of what consulting firms do and how they make money. By 2028, AI-related work will likely represent 50-60% of revenue at major firms. That’s no longer a practice area—it’s the business.
For clients, that means the power dynamic shifts. When AI becomes the majority of consulting work, firms gain leverage. They’re not pitching to win work—they’re selecting which clients to take based on likelihood of successful implementation.
Companies that wait to engage until they feel “ready” will find themselves at the back of the queue behind clients who’ve already established relationships and track records. This isn’t a technology problem. It’s a capacity allocation problem.
The market for experienced AI consulting talent is tighter than the market for marketing AI tools. Your strategy for accessing that expertise matters as much as which technologies you deploy.