Aakash Gupta published a thread on X compiling productivity data on neurodivergent workers. The numbers are so high they look like a typo.

90-140% more productive at JPMorgan. 150% more productive at UiPath. $40 million saved by a single person at SAP.

Two days later, Alex Karp — CEO of Palantir, a company worth $370 billion as of March 2026 — told Fortune there are only two paths to having a future: vocational skills or being neurodivergent.

Meanwhile, only 22% of autistic adults in the UK are employed.

The most productive talent pool in the labor market is also the most underemployed and the most hidden. AI is about to dismantle that broken system piece by piece.

The filter nobody designed on purpose

Your hiring process doesn’t exclude neurodivergent people out of malice. It excludes them by design.

Polished executive summaries. Behavioral interviews where the most socially fluid speaker wins. Group dynamics. “Culture fit” assessments that measure whether someone operates on the same social frequency as the team. None of these filters measure work capability. They measure neurotypical communication style.

The World Economic Forum estimates that 15 to 20% of the global population is neurodivergent. One in five people. That includes autism, ADHD, dyslexia, dyspraxia, among others.

And according to a Zety study with over 1,000 respondents, 94% of neurodivergent people have felt pressured to hide their condition during hiring.

The filter acts before anyone sees a single line of work. We recently wrote that every job posting written before 2025 assumes pre-AI productivity levels. This goes further: many job postings assume a single type of brain.

The data your HR department doesn’t have

Corporate neurodiversity programs are not new. What’s new is the volume of productivity data they’re generating. And the numbers are hard to dismiss.

Company	Program / Role	Measured result	Source
JPMorgan Chase	Autism at Work (tech roles)	90-140% more productive	JPMorgan Newsroom
UiPath + AutonomyWorks	AI data labeling	150% more productive	UiPath Blog
HPE	Software testing	30% more productive	Harvard Business Review
SAP	Complex invoice automation	$40M saved by one employee	Culture Amp
EY	Neurodiversity Centres of Excellence	~$1B cumulative value, 92% retention	EY

Outside the corporate world, the entrepreneurship data tells the same story. A BBC2 and Tulip Financial Group study found that 40% of self-made millionaires in the UK are dyslexic. Research published in Small Business Economics (Springer) shows that people with ADHD are roughly 5 times more likely to become entrepreneurs, based on ~29% of entrepreneurs having ADHD versus ~5% of the general population.

Let’s be clear: this isn’t a feel-good story. The numbers reveal a brutal business reality. When you match the right mind to the right problem, performance doesn’t just get better — it multiplies. And what’s missing in most companies isn’t talent — it’s the ability to make that match.

What AI eliminates (and who it frees)

AI is removing the exact tasks that neurotypical hiring processes were designed to filter for.

Polished cover letters? An AI writing assistant generates them in seconds. Formal written communication? Language models already produce text that exceeds the human average in clarity and structure. Processing long sequential instructions? AI agents decompose complex tasks into manageable steps. Organization, planning, time management? Context-aware AI tools directly reduce executive function load.

What’s left when you strip all that away is the actual work: pattern recognition, hyperfocus on complex problems, lateral thinking, the ability to spot anomalies everyone else misses. Exactly where neurodivergent profiles tend to excel.

The SAP case illustrates this well. Nico Neumann, who joined through the Autism at Work program in 2016, designed a tool that automates the posting of complex invoices with multiple cost allocations. It reduced processing time from 2-3 days to 20 minutes for American Express statements with over 20,000 accounting lines. He won the Hasso Plattner Founders’ Award — SAP’s highest internal honor.

Neumann wasn’t “better” in any general sense. His brain found a problem made for it — and the result speaks for itself.

The bifurcation we analyzed in engineering roles isn’t just senior versus junior. It’s also between those who think in standard patterns and those who think in divergent ones. AI accelerates both, but it removes artificial barriers for the latter.

Karp, Gartner, and the $370 billion signal

Alex Karp is dyslexic. He’s said it publicly several times. In December 2025 he told Fortune: “There is no playbook a dyslexic can master… therefore we learn to think freely.”

He built a $370 billion company with that kind of thinking.

And now he’s recruiting more people like him. The Palantir Neurodivergent Fellowship pays $110,000 to $200,000 per year. Karp personally conducts the final interviews. Not corporate social responsibility. A recruiting pipeline built to capture talent that conventional hiring throws away.

He’s not alone. Gartner predicts that by 2027, 20% of Fortune 500 sales organizations will actively recruit neurodivergent talent. Palantir is two years ahead of that curve.

The roster of neurodivergent founders reads like a hall of fame. Richard Branson built Virgin with dyslexia. Ingvar Kamprad founded IKEA and invented the naming system — furniture named after Scandinavian places — because his dyslexia made it impossible to remember numeric product codes. Elon Musk disclosed his Asperger’s on Saturday Night Live in 2021.

Don’t dismiss these as lucky anecdotes. They’re evidence of what kind of cognition builds empires.

The question for your company: would your current hiring process have screened any of these founders out?

I’m not asking from a distance. My wife says I’m “a bit like Sheldon” — the Big Bang Theory character. I struggle to read sarcasm. In meetings where everyone picks up on social subtext, I’m processing the words literally. That has cost me more than a few awkward moments in my career. But it’s also what lets me see things in data and systems that other people miss — because I’m not distracted by social signals.

When Karp says there’s no playbook a dyslexic can master, and that’s why he learned to think freely, I know exactly what he means. Every brain that doesn’t fit the standard mold develops alternate routes. Some of those routes turn out to be shortcuts.

Redesigning the pipeline is not a favor

Forget adding a “neurodiversity initiative” to your CSR report. This is about fixing a structural flaw in your talent pipeline that’s costing you access to 15-20% of the market — including the segment that data shows can be 90-150% more productive in specific roles.

Three changes that work:

Assess skill, not interview performance. If you need someone who finds patterns in data, give them a dataset and measure what they find. If you need someone who codes, give them a problem and measure how they solve it. Behavioral interviews measure how well someone tells stories about their work. They don’t measure the work itself.

Use AI to remove the barrier, not the person. AI writing tools in the application process. Instructions in multiple formats (text, audio, video). Structured onboarding with clear steps instead of “you’ll figure it out.” These adaptations don’t just benefit neurodivergent candidates — they improve the experience for everyone.

Design roles around cognitive profiles. Nico Neumann’s case at SAP wasn’t luck. Someone identified that an invoice pattern-recognition problem needed exactly the kind of brain he has. That requires your organization to understand which tasks benefit from hyperfocus, divergent thinking, or extreme attention to detail — and to design positions around those strengths instead of asking everyone to fit the same mold.

At IQ Source we evaluate by demonstrated capability, not interview performance. We use AI tools internally to reduce written communication barriers and structure clear processes. Not because we’re a perfect case study — but because we’ve seen the data, and because the person running the company knows firsthand what the filter feels like.

The same principle applies to AI fluency: the gap isn’t the tool, it’s how you connect people with what they do best.

The cost of doing nothing already has a number

If 15-20% of your potential candidates are neurodivergent and your hiring process filters them out before seeing their work, how much talent are you never seeing?

If the data from JPMorgan, UiPath, and HPE is accurate — and it comes from programs with years of operation, not three-month pilots — the performance gap between your current team and one that includes the people your process filtered out isn’t a minor adjustment. It’s the difference between an average team and an exceptional one.

If Gartner is right and 20% of Fortune 500 sales organizations will actively recruit this talent by 2027, companies that don’t move now are ceding a talent advantage to competitors that do.

The same logic from the 100x employee applies here: companies that assumed the old hiring math was fine are discovering it wasn’t.

---

I know the filter because I’ve lived it. And I know it can be redesigned.

If your last job description asked for “excellent verbal and written communication skills” without defining what that actually means in practice, your process has an invisible filter. We don’t know if it’s costing you access to 15% or 40% of your available talent — but we know how to measure it.

Send us the description from your last open position and we’ll show you exactly where the language filters by neurological pattern instead of actual capability.

Audit my job description