AI wants to take on the big consulting firms — but can it actually pull it off?
Artificial intelligence has its sights set on a market that has been moving billions for decades. And this time, the target is one of the most traditional and lucrative sectors in the corporate world: strategic consulting.
For years, major consulting firms like McKinsey, BCG, and Bain have sold something seemingly simple: good advice. Except that advice came wrapped in polished slide decks, proprietary methodologies, and a team of highly trained professionals — all billed by the hour, of course. The model worked perfectly for decades, creating a billion-dollar industry that influences everything from multinational CEO decisions to entire government policies.
Now, developers around the world are trying to replicate that process using AI agents, and the result is at least interesting: knockoff versions of McKinsey consultants running right in your browser. 🤖
The trigger for this trend was an open-source repository from Vercel, an AI startup valued at over 9 billion dollars, which now houses nearly 90,000 reusable skills for AI agents. Among copywriting, code review, and data analysis, there is a growing set of skills directly inspired by the work of major strategic consulting firms. But the big question remains: do these tools actually deliver on their promises? Or are they just imitating the surface of work that goes far beyond any framework?
What are these consulting skills for AI agents?
First things first — it helps to understand what is actually happening here. In the world of artificial intelligence, skills are capabilities that developers create or download and assign to a model or AI agent so it can execute a specific task — without needing to train the model from scratch. Think of them as ready-made recipes: instead of explaining everything you want the agent to do, you simply load a skill that is already configured and ready to go.
The concept gained traction after Anthropic introduced skills for its chatbot Claude back in October. Since then, developers have been building and sharing thousands of skills that can be connected to various AI systems. Vercel’s repository works as a kind of massive collaborative library, where these contributions are published, reviewed, and installed by anyone. The scale this repository has reached — nearly 90,000 contributions — puts this trend on an entirely different level from what we were seeing two years ago.
Business Insider analyzed Vercel’s skills library and found at least four labeled with the term mckinsey and 26 skills labeled with the term consultant. The most popular among the consulting-related ones is precisely the one labeled mckinsey-consultant. It was first published on January 25 and has been averaging 445 installs per week. That is a respectable number, although it is still far from the most popular agents in Vercel’s library, which reach hundreds of thousands of installs.
This skill also has 200 stars on GitHub, which indicates real popularity among developers, and has already gone through some security audits — a sign that it is viable and genuinely gaining traction. Bottom line: people are finding this thing useful.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
How it works in practice
Vercel’s library describes this skill as a prompt framework — originally built for Claude — that guides the AI through well-defined stages: problem definition, hypothesis generation, structured analysis, and slide creation. The stated goal is to replicate the classic workflow of a typical McKinsey consultant.
Within this space, skills inspired by strategic consulting stand out for their ambition. Some promise to run market analyses in the style of the big firms, structure organizational diagnostics, map risks in corporate projects, and even generate recommendations based on famous frameworks like the BCG Matrix or SWOT analysis supercharged with real-time data. The level of detail in some of these skills is impressive: there are instructions that guide the agent to ask clarifying questions before responding, to structure hypotheses before presenting conclusions, and to cite sources when available — behavior that closely resembles the way a junior analyst at McKinsey was trained to operate.
The development of these skills reflects a genuine attempt to codify consultative reasoning — that process of breaking down complex problems into smaller parts, identifying value levers, and communicating insights in a clear and actionable way. This is something that took decades to be systematized by the big consulting firms, and it is now being translated into machine language by an open-source community that grows every week.
What McKinsey has that an AI agent still cannot copy
This is where things get really interesting. To understand the limitations of these skills, Business Insider asked Arvind Vasudevan, a former McKinsey employee, to examine the McKinsey-style agent published in Vercel’s library and assess how it compares to the real thing.
His answer was straightforward and revealing.
It misses the point of how MBBs and strategy consultants add value, Vasudevan said, referring to the group of top consulting firms that includes McKinsey, BCG, and Bain. According to him, a huge part of the value lies in the questions consultants ask and the conversations they lead — interactions that help clarify thinking, uncover unstated assumptions, and ensure deep reflection. None of that is happening with this agent, which basically runs a set of standard analyses without the Socratic questioning and critical thinking that define real consulting work.
McKinsey and its direct competitors do not just sell analysis. They sell relationships, context, and credibility. A senior consultant who walks into a board meeting carries years of exposure to similar situations, the ability to read the political landscape inside an organization, and the judgment to know when data is being interpreted correctly or when it is being used to confirm a pre-existing bias. That kind of contextual intelligence is hard to capture in a skill, no matter how well built it is, because it does not live in text — it lives in the accumulated experience of someone who has been through similar situations over and over again.
On top of that, there is a trust component that goes beyond the output itself. When a company hires a major consulting firm, it is not just buying the final report. It is buying the assurance that if something goes wrong, there is a team of experts accountable for that work — and that team has enough reputation to stand behind the recommendations it made. An artificial intelligence agent, no matter how sophisticated, still operates in a gray area when it comes to accountability. If the analysis is wrong, who is responsible? The developer who published the skill? The company that used the agent? That question still does not have a clear answer, and it matters a lot to executives who need to justify strategic decisions to boards and shareholders.
There is also the human factor in the process of developing a strategic recommendation. Consultants conduct interviews, observe team dynamics, pick up on organizational tensions that do not show up in any report, and adjust their analyses based on that qualitative information. An AI agent, for now, only has access to what is explicitly shared with it. It cannot notice that the CFO hesitated before answering a question or that two directors avoided making eye contact during a presentation. Those signals matter — and they are still beyond the reach of any skill, no matter how well written it is.
AI is already generating real revenue in the consulting market
Despite the limitations, it would be a mistake to underestimate what is already happening. AI agents that mimic the work of consultants are already generating millions in revenue for companies like PromptQL, a corporate AI platform launched by Hasura, a unicorn in the open-source world.
The platform helps clients build custom AI analysts, integrating companies’ internal data with the foundation models they already use. Once deployed, these AI analysts can execute tasks that would normally be handled by data scientists or engineers — and they keep learning and adapting over time.
Tanmai Gopal, co-founder and CEO of PromptQL, told Business Insider that the biggest barrier — or moat, in business jargon — to selling analysis is understanding the relationships between people, data, and revenue.
According to Gopal, McKinsey teams spend weeks embedded inside a company, absorbing how it actually works: the exceptions, the tribal knowledge, the definitions that differ between departments. That company-specific context is what makes their advice worth millions.
Gopal also pointed out that enterprise AI tools frequently fail because they lack proper grounding. They tend to guess instead of asking questions, learning from feedback, or maintaining shared understanding across teams.
PromptQL’s bet on solving the problem
PromptQL tries to address these issues through a shared layer of understanding that adjusts with every new input. The concept is clever: when a team member corrects the AI, teaches it a definition, or resolves an ambiguity, that knowledge becomes permanent and available to everyone. It is not a semantic layer maintained by data engineers — it emerges from real conversations between humans and the tool.
AI models do not automatically know internal nuances, like pricing changes, team-specific terminology, or conflicting definitions of what counts as revenue. The real problem, according to Gopal, is not capability — it is missing context.
In other words, the consultant’s slide deck was never really the product. The product is the judgment — and that is the part AI is still learning.
Where AI skills truly shine in this context
All that said, it would be unfair to ignore what these tools do really well. For preliminary analysis, synthesizing large volumes of information, and structuring problems that have already been clearly defined, artificial intelligence agents with consulting skills deliver results that would be impossible to replicate manually in the same timeframe. An analysis that would take two junior analysts three days to complete can be done in minutes — with reasonable quality, especially when the problem is well scoped and the data is organized.
For startups and small businesses that would never have access to a McKinsey or similar firm, this represents a real democratization of analytical tools. It is like having a strategic assistant that does not charge 500 dollars an hour and is available around the clock.
Another strong point is consistency. An AI agent does not have a bad day, does not forget to include an important variable because it was in a rush, and does not deviate from a defined framework for emotional reasons. When properly configured, it executes the same analytical process with the same rigor every single time — which is valuable in contexts where standardization matters, like internal audits, periodic performance reviews, and recurring diagnostics.
The future is complementarity, not replacement
The trend taking shape, then, is not about replacement — it is about complementarity. Consulting skills for AI agents are most useful when they function as a first stage — accelerating the collection, structuring, and synthesis work that precedes deeper human analysis. They reduce the time spent on repetitive tasks and free up space for qualified professionals to focus on what truly sets great consulting work apart: judgment, relationships, and reading the room.
In this scenario, the AI agent does not replace the consultant — it changes what the consultant needs to do in order to add value. And that alone is already a significant transformation in the market. Professionals who learn to use these tools as allies, rather than ignoring or fearing them, will likely have a real competitive edge in the years ahead. 🚀
The numbers show this movement is irreversible. With nearly 90,000 skills available in Vercel’s repository and a global community of developers feeding this ecosystem daily, artificial intelligence is increasingly infiltrating processes that were once exclusively human. The consulting market is just one more territory being explored — and it certainly will not be the last.
The slide deck was never the real product of a consulting firm. Human judgment is. And as long as AI has not mastered that skill, what we will see is a race to define where the machine’s work ends and where the irreplaceable value of the professional begins.
