Inside a publicly traded company’s journey to becoming “AI native”

A long banner hangs in Amplitude’s San Francisco office. It says, “No magical thinking allowed.”

No, it’s not a Joan Didion rag. CEO Spencer Skates told Business Insider that it was a reminder that technology can never replace deep thought and hard work. In the AI ​​era, this reminder is more important than ever, so much so that employees need to check it every day.

Amplitude is a publicly traded analytics company with 800 employees that is undergoing an AI transformation to revitalize its business.

Amplitude went public in September 2021 at the height of the pandemic and rose to an all-time high of $84.80 per share a few weeks later, but has since fallen significantly and has largely peaked at around $10 in recent years. Friday’s closing price was $10.25.

The company has acquired five AI startup companies since October 2024. Amplitude hired an AI-savvy head of engineering and appointed one of its acquired founders to a new AI leadership position. We licensed Cursor and GitHub Copilot for our employees and conducted a thorough AI week.

This is a change that many companies are making. We are rapidly moving from using little or no AI to becoming “AI native,” a term that is strangely difficult to define precisely. Large-scale language models are popping up everywhere in white-collar work as companies pursue the promise of increased efficiency.

Amplitude’s case may be particularly instructive given how skeptical the company’s CEO was about AI. Skate said that in 2023 and parts of 2024, he sees the AI ​​industry as full of “con artists”, including visionaries who promise to end world hunger and salespeople who promise to automate everything.

“There were all kinds of problems,” Skate said. By mid-2024, he realized that “breakthroughs will probably occur in the field of analytics within the next two to three years.”

“We have to make it ourselves,” he said. “So we went all in.”

‘Tens of millions of dollars’ poured into AI overhaul

Skates had two opening moves for the AI ​​overhaul.

The first is hiring a new chief engineering officer with experience in AI. Wade Chambers has been advising the company since 2016, while holding leadership roles at Twitter and Included Health.

When Chambers joined the company in October 2024, only 1% of Amplitude’s engineering, product and design teams were using AI.

The second is the acquisition of chatbot startup Command AI. This was the first in a series of acquisitions including June, Craftful, and Inari. Amplitude announced its latest acquisition, InfiniGrow, on January 14th.

Jana Welinder was the CEO of Kraftful, one of Amplitude’s acquisition targets. Kraftful was able to identify power users of its product, one of whom was Amplitude’s CPO at the time. She reached out and we had a conversation in February. The deal closed in July, and Wellinder was named Amplitude’s head of AI. The company’s blog post, which included an introductory Q&A, called her an “AI guru.”

Wender’s first task was to bring the company up to speed. Craftful ships new products every week. Amplitude’s shipments were less than monthly.

“Shipping infrequently like this slows down your team, which is not appropriate in the age of AI,” she said.

By October, Skate sent an official email to staff. The title is “Introduction: AI Native Amplitude”.

“Analytics will look very different in six months,” Skate wrote in an email. “We have the opportunity to of We’re an AI-native company in analytics, and we’re going to use everything we have. ”

He also asked employees to share upcoming launches on X rather than LinkedIn. This is because X is “a place where AI natives are”.

How much did Amplitude spend on AI, from tools to acquisitions? “Certainly in the tens of millions,” Skate said. “I wouldn’t be surprised if it exceeded $100 million.”

How to change the mind of an AI skeptic

The next difficult thing is to convince employees Get hands-on with the tools.

While some engineers are excited about the potential of AI, others are skeptical about its usefulness or worried about potential job losses. Not all engineers are as passionate about AI as executives.

Skate said engineers are particularly sensitive to “frustration” happening with AI, and many are skeptical. A bottom-up approach erases that skepticism, he says.

Shortly after joining the company, Chambers began planning an “AI Week” for the first week of June. It took six months to prepare and borrowed a lot of money from Facebook’s mobile push. He took his entire engineering, product, and design teams offline for a week. First, Mr. Chambers required the leader to go on stage and code something in front of the entire company.

“It didn’t go well,” Chambers said of the live vibecoding demonstration. “They had to work through that. They had to re-prompt a few different ways.”

But the message stuck with him, he said. Even readers who hadn’t coded all day were able to build something “pretty cool” within an hour session, barring a few hiccups.

Adding to the momentum were “enthusiasts,” engineers passionate about exploring new technology (some of whom Chambers brought over from his previous job). These engineers lead by example, he said.

Amplitude shared internal data tracking the number of employees using AI tools. In the last week of March, 14 employees were actively using Cursor. This number peaked in the first week of December, after AI week and before the holiday break cycle, with 174 employees.

What about the thorny question of AI adoption in enterprises: ROI? After all, a 2025 MIT study showed that 95% of companies that publicly announced the use of AI pilots did not report measurable ROI.

After implementing these tools, developer productivity skyrocketed by 40% and stayed that way, Chambers said. For some specific engineering teams, these gains appear to be as high as 300 to 400 percent, he said.

“There are a lot of people who think they’re the world’s best expert on something,” Chambers says. “Increasingly, even the most cynical team members were coming in.”