McKinsey Faces AI Challenge, But Not Immediately

Navin Chaddha, managing director of the 55-year-old Silicon Valley venture firm Mayfield, is betting big on AI’s potential to transform people-heavy industries like consulting, law, and accounting. The veteran investor, who has backed successes including Lyft, Poshmark, and HashiCorp, recently explained at a TechCrunch StrictlyVC event in Menlo Park why he believes "AI teammates" can create software-like profitability in traditionally labor-intensive fields. He also discussed why startups should target underserved markets rather than compete directly with giants like Accenture — though he acknowledged that disrupting industries built on relationships and trust can be tougher than Silicon Valley expects. This conversation has been edited for length and clarity.

You argue that law firms, consulting companies, and accounting services — a combined $5 trillion market — will be completely reimagined by AI-first companies operating with software-like margins. How do you back that up? What evidence have you seen beyond conceptual presentations?

One advantage of a firm that's been operating for over 50 years is witnessing every major trend, from mainframes and minicomputers to PCs, the internet, mobile, cloud, social, and now AI. Consider the late '90s concept of e-business: physical stores realized they couldn't survive on brick-and-mortar alone; they needed a digital presence. Then outsourcing and offshoring became massive trends—you couldn't build a software services company without a presence in India or another emerging market. The same shift occurred in supply chains and manufacturing with the rise of China and Taiwan. So, what defines this new AI era? AI is a 100x force, augmenting human capabilities to achieve better outcomes. I believe it will fundamentally reshape business.

AI will handle many repetitive tasks… and two growth models will emerge. One is organic growth. The second is inorganic growth.

Can you provide a concrete example of how this will function?

What can an LLM or AI actually do? Take implementing Salesforce, for instance. Who wants to do that manual work? A human project manager might outline the requirements: ‘You need to implement Salesforce.’ It's a standardized process. Use AI as the primary engine to execute it, and bring in a human only for tasks AI can't handle.

Suddenly, by operating this way, you rely less on human labor and more on AI, and customers pay only for the AI resources they actually use.

Save $450 on your TechCrunch All Stage pass

Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and more for a day packed with actionable strategies, hands-on workshops, and valuable networking.

Save $200+ on your TechCrunch All Stage pass

Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and more for a day packed with actionable strategies, hands-on workshops, and valuable networking.

The market entry strategy shouldn't be to challenge giants like Accenture, Infosys, or TCS head-on. Instead, target the underserved majority. There are 30 million small businesses in the U.S. and 100 million globally that can't afford traditional knowledge workers. Provide them with software-as-a-service solutions. These businesses say, "I need a virtual receptionist, a scheduling assistant, someone to build my website…" AI can be used to automate tasks like preparing startup funding documents, with human involvement reserved for negotiation. Don't compete with the Accentures of the world. Go after fragmented markets where you charge per event or outcome, not by the hour or a monthly contractor fee.

So, it's outcome-based pricing instead of time-based billing.

Exactly, it's outcome-based. Think of cloud billing or electricity—you pay for what you consume. If AI handles 80% of the work, it can deliver 80% to 90% gross margins. Human-led services might yield 30% to 40% margins. This creates a blended margin of 60% to 70%, producing 20% to 30% net income. And remember, most services companies are profitable. Many tech companies aren't; they rely on venture capital and public market funding.

You recently led the Series A for an AI tech consulting startup called Gruve. What did you observe in its early customer pilots?

This demonstrates the blend of inorganic and organic growth. Gruve was founded by successful entrepreneurs who had previously bootstrapped two services companies to $500 million in revenue each, with $50 to $100 million in profits. This time, they leveraged their expertise in security. They acquired a $5 million managed security services firm and decided all future growth would be driven by AI. They scaled that business from $5 million to $15 million in revenue in just six months, achieving an 80% gross margin with an outcome-based model. Customers, including Cisco, love it. Instead of paying a fixed $10,000 monthly fee for security outsourcing, they pay Gruve nothing unless an actual security event occurs that Gruve handles.

Couldn't firms like McKinsey simply acquire these AI capabilities? They have large, established businesses to protect.

This is a classic innovator's dilemma. When perpetual-license enterprise software companies saw SaaS emerging, they were reluctant to adopt a subscription model because it meant moving away from large upfront payments and annual maintenance fees. The business model innovation was key, and many incumbents hesitated. Similarly, McKinsey and Accenture are preoccupied serving their large existing clients, which creates an opening. I advise founders to serve the neglected masses, develop a unique go-to-market strategy, and target customers these giants can't efficiently serve.

But the incumbents will also be forced to transform. The small companies not competing with them today will, mark my words, be their competitors in a decade. Those large firms—McKinsey, BCG, Accenture, TCS, Infosys—all face the innovator's dilemma: when do they make the painful shift to an outcome-based AI model? As public companies, transitioning from predictable revenue to utility-based revenue is a daunting prospect.

Last fall, you allocated $100 million from your latest fund specifically for "AI teammates." What distinguishes a true AI teammate from a mere AI tool?

The industry is full of buzzwords: copilots, tools, agents, teammates. Mayfield's thesis is that an AI teammate is a digital companion that collaborates with a human toward shared objectives to achieve better results. It may be built on agentic or copilot technologies, but its manifestation is as a specific role—an HR teammate, a sales engineering teammate. The goal isn't replacement; it's partnership and collaboration.

When people began discussing AI teammates and assistants, it sounded innovative. But as job losses mount, could this terminology seem insensitive? Does Silicon Valley have a messaging problem?

You're absolutely right, and we shouldn't sugarcoat it. We must address it directly. Yes, there will be job displacement. But humans are adaptable; we are the jockeys, and AI is the horse. We will reinvent ourselves. The immediate focus is on cost reduction, but we'll learn to expand markets and increase revenue. This pattern repeats with every technological wave. When Microsoft Word arrived, people thought executive assistants were obsolete. Excel threatened accountants. Uber and Lyft were supposed to erase taxi drivers. Instead, these markets expanded.

My view is similar to how emerging markets like India, China, and Africa skipped landlines and went straight to wireless cellular. AI will perform work in markets where human labor isn't available or affordable to serve the customer. Long-term, I'm extremely bullish. Short-term, there will be pain, but no pain, no gain.

On the topic of coding, there was a recent "vibe-coding" acquisition: a six-month-old Israeli company with 250,000 monthly users and $200,000 in monthly revenue was bought by Wix for $80 million in cash. Does that valuation make sense to you?

Honestly, in today's AI climate, traditional math often doesn't apply. It's unpredictable. I'm surprised a company with $2.4 million in annual recurring revenue sold for only $80 million—I might have expected $800 million! [Laughs.] In this market, it's hard to know. It's a dynamic marketplace.

How do you invest rationally in such a market?

That's where the experience of proven investors becomes the secret recipe. It's not pure science; it's an art. It's the 10,000-hour rule: the more you practice, the better you get. Firms that have been around for 50 or 60 years have seen all kinds of bubbles.

The number-one rule is to have your own North Star. Maintain discipline and avoid FOMO—fear of missing out is for followers. If you have your own strategy and operate without fear, you'll do well. Remember, for VCs in this audience, we are in the money management business. It's not about collecting logos; it's about turning small amounts of capital into larger sums.

In this phase of the cycle, significant wealth will be created. But I believe 80% of participants will lose money because they don't truly understand what they're doing.