AI's Attempt to Clone McKinsey: A Mirage or Reality?

For decades, consulting firms like McKinsey have charged a premium for their expertise, delivered through detailed presentations and billed by the hour. Now, AI startups are trying to replicate this model, making digital consultants available at the click of a button.

The Rise of AI Consultant Skills

Vercel, an AI platform valued at over $9 billion, has introduced a library of nearly 90,000 open-source skills. These skills, ranging from copywriting to consultant-style problem solving, allow developers to imbue AI agents with specific capabilities without starting from scratch.

The concept gained popularity when Anthropic launched skills for its chatbot Claude last October. Since then, developers have actively created and shared these skills, including at least four labeled "mckinsey" and 26 tagged "consultant" in Vercel's library. The "mckinsey-consultant" skill, averaging 445 installs a week, is a notable example. But does it truly measure up to the real deal?

What AI Can't Imitate

Arvind Vasudevan, a former McKinsey engagement manager, tested these AI tools and concluded they fall short. "It misses the point of how MBB and strategy consultants add value," he noted. The AI lacks the critical questioning and deep thinking that human consultants provide, opting instead for standardized analysis.

So, if AI can't ask the right questions, how can it achieve the same insights? The reality is that these agents might execute tasks but fail to grasp the nuances and complexities that seasoned consultants do.

The Business of Digital Consultants

Despite these shortcomings, companies like PromptQL are already capitalizing on AI-driven revenue. By integrating client data with foundation models, PromptQL's platform creates custom AI analysts capable of assuming roles typically held by data scientists.

Tanmai Gopal, PromptQL's cofounder, argues that understanding the intricate relationships between data, people, and revenue is vital. The real value lies in the deep company-specific knowledge consultants acquire, something AI currently struggles to replicate.

These AI models aren't inherently flawed. rather, they lack the context and judgment that define successful consulting. So, when will AI truly step into the shoes of a McKinsey consultant? Color me skeptical, but that's still a distant goal.