Pangram, a startup founded by former Tesla and Google engineers, raises $4M to detect AI-written content

Pangram, the startup building tools to detect AI-generated text, has raised $4 million in seed funding to improve the accuracy of its detection technology. The round was led by California-based VC firm ScOp, with participation from Script Capital, Cadenza, and a group of angel investors. Haystack VC, which backed the company early on, led the pre-seed round.

The latest $2.7 million extension adds to an earlier $1.25 million pre-seed round, bringing the company’s total funding to just under $4 million. The new funding will be used to grow the eight-person team and expand the product lineup, including tools for everyday users, not just enterprise clients.

Founded by former Tesla engineer Bradley Emi and Stanford alum Max Spero, Pangram is betting it can help sort out what’s written by humans as AI-generated text pours into classrooms and offices.

With $4M Funding, Pangram Aims to Battle AI Text Flood in Classrooms and Workplaces

“Identifying AI text with confidence is increasingly essential across education, business, and the media,” said CEO Max Spero. “We have the best technology to do that.”

As tools like ChatGPT and other AI chatbots continue to spread, companies and schools are racing to verify the authorship of written content. That growing concern is fueling demand for reliable detection software—something Pangram believes it can deliver with a smarter approach.

Unlike some of the bigger players in the space, Pangram says its tech is built to learn from edge cases. Its system relies on a feedback loop that generates and trains on confusing examples, helping it improve over time. It also leans on open-source models to keep infrastructure costs low.

“What makes great machine learning products in general is great data. And I think that’s like where we’ve kind of focused most of our effort,” Emi told Reuters in an interview.

The startup already counts Quora and NewsGuard as clients. It offers a $15/month plan for individuals with up to 600 scans, and has pro and developer options priced at $45 and $100 per month.

With deep experience from Tesla and Google, and a sharper focus on adaptability over scale, Pangram is betting that its approach will stand out as more organizations look for help separating machine-generated text from the real thing.