In June, Superhuman acquired GPTZero, a New York-based startup best known for building one of the most popular tools for detecting artificial intelligence in student writing. Its young founders and team have since joined Superhuman’s newly formed “Authenticity” group, working to integrate detection capabilities across the company’s products, most notably Superhuman Go, an AI assistant designed to work through everything a user does online.
The deal marks the latest step in a rapid turnaround led by CEO Shishir Mehrotra, who took over in early 2025 after the company, then called Grammarly, acquired his productivity startup Coda. That summer, Grammarly purchased Superhuman Mail, a powerful user email tool built on the idea of “no inbox.” In October 2025, Mehrotra renamed the combined company Superhuman, signaling its ambition to build a suite of AI tools beyond the writing assistant.
Before founding Coda in 2014, Mehrotra built and launched products for YouTube and Microsoft. It replicates the familiar playbook of big tech companies, with Google becoming Alphabet in 2015; Facebook is meta in 2021, as a well-known product evolves into something more expansive.
“There have been enough instances of doing it in a way that maintains the core brand,” Mehrotra told the Observer at the time.
Grammarly, still Superhuman’s most popular product, is used by 40 million people daily. Students rely on it to improve essays. Editors use it to catch typos and enforce style consistency, sometimes across sprawling internal rulebooks that can run hundreds of pages.
The use of AI in writing – arguably one of the most human forms of work – has been controversial from the beginning: How much is AI too used? Is it better to use AI to create the first draft or do the final polish? Increasingly, can anyone reliably tell the difference?
Mehrotra had to answer these questions publicly. Earlier this year, the company faced backlash over its “Expert Review” feature, which generated suggestions in the style of popular writers and led to a class-action lawsuit over approval.
“There’s a clear line, depending on the mission, between the things you should use AI for and the things you shouldn’t,” he told the Observer in an interview earlier this month. “In the case of a writer, it can be a great tool for doing research or understanding a topic. But if it creeps into the actual writing and takes away from the human doing the reporting, it reflects poorly on both the writer and the executor.”
Educators and editors have long complained that AI detection tools are unreliable, while many AI researchers say The problem is basically unsolvable. However, Mehrotra is already looking at a bigger picture. “Think of Grammarly as a super English teacher sitting on your shoulder all the time, helping you everywhere you work,” he said. “But this is actually a huge underutilization of the infrastructure we’ve built.” “The core infrastructure is the ability to bring immersive AI into your workplace.”
This idea underpins the company’s rebranding and, soon after, the launch of Superhuman Go, which extends Grammarly’s real-time assistance to a more general system of agents.
“We enable anyone to build an agent that works just like Grammarly, but does things that go far beyond grammar,” the CEO explained. “For a salesperson, this might mean having a sales coach next to you, warning you that you’re about to recommend the wrong product while you’re emailing a customer. For a support agent, it might mean having a support coach reminding you that a customer had a great connection yesterday and suggesting you mention it.”
Searching for human qualities in a world saturated with artificial intelligence
Superhuman’s many departments — from engineering to sales to customer support — use almost every AI tool on the market, Mehrotra said. But when the information reaches the CEO level, he faces his own test in drawing boundaries for the use of AI.
“I often tell people: Use AI to help you form your opinions, but not to write your summaries and reports,” he said. “When people send me something that clearly looks like it was written by AI, I often say to them, ‘I’d rather just get the prompt.’” I’d rather see your unique, raw vision and go from there, rather than read polished AI output.
The spread of artificial intelligence has also changed the hiring process. Technology leaders like Bill Gates predicted early in the AI boom that integrating AI into work would be like having a digital personal assistant all the time. This has changed what executives look for in job candidates, even for entry-level positions.
“AI has added another standard to almost every role; we turn everyone into a manager very early in their experience. You will have a team working under you from day one, which is very different from the way it worked in the pre-AI world,” Mehrotra said. “Suddenly, we have to look for very similar skills.” [to managers] Even for entry-level candidates. Your ability to work with AI to achieve the right outcome is incredibly similar to your ability to work with a team of human employees.
What remains is the search for unique human qualities. Mehrotra and his co-founder Matt Hudson have promoted the idea of applying a system of “self-questioning” to the decision-making process — a metaphor from mathematics that reframes prioritization.
“We often find ourselves in situations where we have ten questions to answer and we have to prioritize them,” he explained. “People usually rank them by intensity or urgency. But if you use a self-question test, you look at that list and ask: ‘If you answered this one question, how many other questions on the list would you answer?’ It might be question number six on a list of ten.”
This way of thinking is especially important in making strategic decisions. As AI becomes more than capable of generating plausible answers for everything, it becomes a scarce resource that decides which problems are worth solving and which outcomes are worth trusting.
“When I think about the misuse of AI, I think people often try to replace human insight with AI,” Mehrotra said. “And when they do that, they end up, at best, embarrassing themselves, and at worst, with really wrong results.” “One incredibly human thing is to govern with very little information. Ultimately, knowing what problem to solve also seems, at least for the moment, to be uniquely human.”
