How a 19-Year-Old in Karachi Built an AI Tool That’s Outperforming McKinsey Consultants

A certain type of ambition is one that doesn’t show itself. It doesn’t arrive in a pitch deck created by a Manhattan design firm or a Stanford hoodie. It occasionally appears in a ground-floor apartment in the Gulshan-e-Iqbal neighborhood of Karachi, operating on a used laptop with a fan that rattles when the GPU gets hot.

That’s essentially the environment in which Zain Ahmed, a nineteen-year-old without an MBA or Silicon Valley experience, created the product that is quietly unsettling consulting veterans.

ConsultIQ, his AI tool, helps more than just consultants. It performs better than junior analysts at companies like McKinsey on a number of practical measures, including competitive benchmarking, proposal drafting, market research, and the kind of document synthesis that used to take teams of people two full weeks to produce.

To grasp the scope of what Zain appears to be competing with, even if unintentionally, it’s worth taking a moment to consider what McKinsey has created. Over 100,000 documents and interview transcripts representing almost a century of intellectual property were used to train the company’s internal AI platform, Lilli. Currently, more than 75% of McKinsey’s 43,000 workers use it on a monthly basis.

It can create presentations, modify tone, draft proposals, and bring up internal experts in 70 different nations. McKinsey used years of iteration and practically limitless resources to build that. Zain reportedly used nearly none of those resources to create something that functions similarly well in a fraction of the time.

How? That is the area that is still a little challenging to square. His architecture, which included large language models, retrieval-augmented generation, and a layer of fine-tuning on consulting-specific documents scraped from public sources, wasn’t wholly original. However, his accuracy was more instinctive than technical. Without anyone telling him, he realized that thinking wasn’t the true bottleneck in consulting. It was the work that surrounded the thought.

Before they are even hired for a position, the typical consulting firm spends more than twenty hours putting together a proposal. When describing what his own platform was attempting to solve, Alibek Dostiyarov, the founder of Perceptis, a VC-backed startup with a former McKinsey consultant and former Apple engineer on its founding team, cited that figure.

Without the background or the $3.6 million that Perceptis raised from Streamlined Ventures and a group of angel investors, including a Meta board member, Zain independently came to the same conclusion.

There is a perception that the consulting industry has been sluggish to take into account the talent that comes from sources other than its typical talent pipelines. The structural reality that AI has subtly leveled a playing field that previously required decades and elite credentials to access is more significant than just the Zains of the world.

McKinsey’s own senior leadership is essentially confirming that the intellectual moat surrounding Big Consulting is shallower than it once seemed when they acknowledge that armies of business analysts producing PowerPoints are no longer strictly necessary and that technology can handle it.

In contrast to Lilli, what Zain created is not a final product. It’s less cohesive, rougher, and sometimes unpredictable in ways that a well-funded engineering team would swiftly resolve. However, those who have tested it say it holds up in the one area that is arguably most important: creating an engaging, personalized client proposal that gets the firm hired.

Some claim to be surprised by it. It feels less like automation and more like instinct because of the way it builds a story around a client’s particular issue, matching their experience and knowledge to the precise shape of an opportunity.

This is not really a story about a single adolescent in Karachi. It concerns the consequences of democratizing access to tools that previously required institutional scale through generative AI. Because of platforms like ConsultIQ and Perceptis, small and mid-sized consultancies—those that couldn’t afford to create their own Lilli—are now engaged in more serious competition.

The goliaths won’t go away, but smaller, quicker, more specialized players might be the main source of growth. Dostiyarov has been presenting investors with this argument. Without a thesis at all, Zain stumbled into this reality.

It’s difficult to ignore the fact that the most intriguing AI disruptions at the moment aren’t coming from the most obvious sources. They originate from individuals who were not informed that specific issues were the responsibility of specific organizations.

It was not Zain Ahmed’s intention to take on McKinsey. He simply needed to solve a problem, was capable of doing so, and didn’t know enough to believe it couldn’t be done. Competence, necessity, and the blissful lack of imposter syndrome may be the most underappreciated force in technology at the moment.

It’s difficult to predict whether ConsultIQ will grow into a legitimate business, be acquired, or simply disappear as the market becomes more competitive. Zain is 19 years old. From here, it’s hard to find the ceiling.