# Less Than 1% of Candidates Pass This Hiring Bar > Matt Swulinski built Wispr Flow's growth from pre-launch to millions of users. He says fewer than 1% of candidates are AI-native systems thinkers, and the hiring data says that's exactly who the market now pays for. Published: 2026-08-19. Last updated: 2026-08-19. Matt Swulinski was asked on [20VC](https://www.thetwentyminutevc.com/matt-swulinski) how what he looks for in talent has changed in the last year or two. A year ago, he said, he wanted ten years of experience. Now? **"If they're AI native or a systems thinker and can deconstruct what makes their job hum, that person plus experience will out-compete someone that just has experience."** And of the candidates he actually meets, fewer than 1% clear that bar. This isn't a hot take from the sidelines. Matt was marketing hire #1 at Wispr Flow and built its growth function from pre-launch to millions of users. Before that, Superhuman. He scaled Wispr Flow by thinking in systems, so his hiring test is whether you can too. ## What systems thinking actually means It has nothing to do with which AI tools you've used. It's a loop: 1. Take one step back from the task you're doing. 2. Identify how that task connects to the rest of your job. 3. Ask whether some or all of it can be automated. 4. Automate it, then replace that time with higher-leverage work. Matt's interview version of this is brutal: ask a candidate what it would take to fully automate their job. Most people, he says, discover they don't really understand their job. They know the pedals and the route. They've never popped the hood. In the full conversation with Harry Stebbings, he walks through how he built AI-native systems to fuel Wispr Flow's growth. It starts with simple processes, then adds feedback loops that make those processes better over time. That second part is where most people stop short: a prompt with some context is a process. A system critiques its own output and feeds the critique back in. ## The market already prices this Matt's 1% is one hiring manager's funnel. The aggregate data says the same thing louder. As far back as May 2024, [71% of leaders told Microsoft and LinkedIn](https://www.microsoft.com/en-us/worklab/work-trend-index/ai-at-work-is-here-now-comes-the-hard-part) they'd rather hire a less experienced candidate with AI skills than a more experienced candidate without them. Two-thirds said they wouldn't hire someone without AI skills at all. Two years later the premium shows up in wages. [PwC's 2026 Global AI Jobs Barometer](https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-ai-jobs-barometer.html), built on a billion job ads across 27 countries, measures a **62% average wage premium for workers with AI skills**, up from 57% the year before. Jobs requiring those skills are growing roughly eight times faster than the market overall. Experience stopped being the tiebreaker. Legibility of your own job is. ## Why "become AI native" keeps failing Here's the uncomfortable half of Matt's argument: a lot of companies are trying to make their teams AI-native, and they're failing. His diagnosis is blunt. The people in those roles are not systems thinkers. You can't hand an agent to someone who can't name the parts of their own job. The failure is measurable. An [MIT NANDA report](https://finance.yahoo.com/news/mit-report-95-generative-ai-105412686.html) found 95% of enterprise gen-AI pilots showed no measurable P&L impact, despite $30-40 billion in spending. To be fair, it's not only the people. Microsoft's [2026 Work Trend Index](https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization) finds organizational factors drive 67% of AI's reported impact versus 32% from individual mindset. Readiness times capability. A team of systems thinkers inside a company that punishes experimentation stalls just as hard as a ready company staffed with people who can't name the parts of their own job. Either factor at zero zeroes the whole product. This is the same skill gap behind the [agent operator role Aaron Levie sees coming](/blog/agent-operator-million-jobs). The models are good enough. The scarce resource is people who can decompose work, because decomposed work is delegable work. ## There's never been a better time to become AI-native The vision for General Input is to make these AI-native systems much easier to build: batteries included, low friction, and easy to share across your team. The mapping is still yours to do. Nobody can deconstruct your job but you. So run Matt's audit on yourself before someone hiring you does. What would it take to fully automate your job? If you can't answer, that's not evidence your job is safe. It's evidence you haven't looked under the hood yet. The 1% who have are getting hired. ## FAQs ### What is an AI-native systems thinker? Someone who can step back from the task in front of them, map how their whole job fits together, and identify which parts an agent can run. The skill isn't knowing AI tools. It's understanding your own work well enough to delegate pieces of it. ### Does AI skill really beat experience in hiring now? Increasingly, yes. As far back as 2024, 71% of leaders told Microsoft and LinkedIn they'd rather hire a less experienced candidate with AI skills than a more experienced one without. PwC's 2026 AI Jobs Barometer measures a 62% average wage premium for workers with AI skills. ### How do I become a systems thinker in my current role? Run the automation audit on yourself: ask what it would take to fully automate your job. Write down every moving piece, the boring admin, the reporting, and how they connect. Then automate one piece and spend the reclaimed time on higher-leverage work. Repeat.