Cluely has been a startup that began with controversy and has stayed there by design.
Founded by Roy Lee and Neel Shanmugam, the company first came into the spotlight after the duo were suspended from Columbia University for building an AI-powered tool that quietly helped engineers pass technical interviews. What many saw as a violation of academic integrity became the foundation of a business idea that has now attracted some of the biggest names in venture capital.
Credits: Columbia Spectator
Within months of the incident, Cluely closed a $5.3 million seed round, followed by a $15 million Series A funding round led by Andreessen Horowitz. According to two investors familiar with the deal, the company is now valued at approximately $120 million.
“The point of Cluely is to solve the problem of not having the right information when you need it,” said co-founder and CEO Roy Lee. “And to magically be able to have all the information you need at every single point. It feels like magic, and it feels like cheating.”
AIM Media House sat down with Roy Lee, the 21-year-old co-founder and CEO of Cluely, who says there are very few people in the world who have taken as much risk as he has. There’s probably no other 21-year-old in the world who has as interesting a life as he has. And the upside of risk, the potential benefit is much, much higher, as we can see.
Rather than let critics define the company, Lee has chosen to confront the label head-on. “It is inevitable that when you see someone use Cluely in an interview, a sales call, a meeting and they use AI in a way nobody else knows, someone’s going to think it’s cheating,” he said. “Rather than let the public paint that picture for us, we embraced it.”
This embrace of controversy has become part of Cluely’s identity and its strategy. The company’s tagline, “cheat on everything,” is both literal and deliberate.
Credits: Decrypt
“Everyone already thinks AI is cheating,” Lee said. “You use AI to transcribe a meeting. You use it to write an essay so you’re cheating. The only way to flip that narrative is to say it before they can.”
Despite its polarizing brand, Cluely is not a gimmick. The company is already profitable, with a business model that spans both consumer and enterprise segments. Individuals can access the tool for $20 per month or $100 per year, while enterprise customers are billed annually per seat with some contracts reaching seven figures, according to Lee.
Most of Cluely’s revenue today comes from enterprise clients, including teams that use the software in recruiting, sales, and customer service. The product runs quietly in the background, helping users answer questions, pitch products, or respond fluently during fast-moving conversations.
Credits: Cluely
Under the hood, Cluely doesn’t train its own large language model. Instead, the company uses existing models like OpenAI’s GPT while focusing its technical efforts on a problem Lee describes as “context stitching.”
“The actual challenge isn’t the model,” he explained. “It’s how you stitch all the information on the screen, in the audio, and in the prompts into a coherent input that the model can reason over.” For speech-to-text, Cluely uses Deepgram, customized for speed and accuracy. “With a few tweaks to the model,” Lee said, “it’s the fastest and most accurate speech-to-text provider out there.”
Roy added to the controversy this week by talking about how he believes that there is no work life balance. The team with just twelve full-time employees lives and works together under one roof. “We hang out together. We watch movies. We go to the gym. Everything we do is way more fun than what you do at your company,” Lee said. “There’s no work-life balance because no one here wants to stop working on Cluely.”
The company’s structure is lean by design. Most operational roles are outsourced. And Lee is blunt about the roles he doesn’t believe in. “Product manager is a fake role invented by a fake person,” he said. “If your engineers are aligned and know what they’re doing, you don’t need someone telling them what to build.” The only two roles Cluely hires for, according to Lee, are engineers and influencers. While the engineers focus on core product, the influencers are responsible for driving growth across platforms like TikTok, X, and Instagram.
“You can’t buy attention anymore,” Lee said. “Corporate commercials don’t work. What works is creators who are hot on the algorithm.”Cluely’s marketing strategy relies entirely on short-form video creators with built-in audiences. No paid ads. No traditional PR. Just content that performs.
For the future, Lee believes Cluely will become the default interface for AI. “Nobody is going to be using chatgpt.com in five years,” he said. “Everyone will be using something like Cluely. This will be the dominant user experience.”
His ambition doesn’t stop there. Lee wants Cluely to be the fastest startup to hit $100 million in ARR, and is already in discussions to expand globally. “We’re planning on going global as soon as possible,” he said. “We’re already in conversations with many people from the Middle East. Who knows, maybe you’ll see an Indian office next year.”
For Lee, risk has always been part of the equation. He dropped out of college, built a product that sparked academic backlash, and turned it into a venture-backed business in under a year. For learners trying to break into AI, Lee’s advice is equally direct: build something. “The best way to learn is by acting,” he said. “And the best way to act is by building. Don’t study AI, use it.”
As Lee sees it, the trajectory of AI is not a matter of speculation but of certainty. The models we use today will be the least intelligent we ever encounter. If AI is already capable of handling complex tasks now, it’s only logical to assume it will outperform humans in those same tasks within a few years. And in that world, anything AI can do today is something no human should be required to do tomorrow.
