Zayd Enam and the trouble with locked doors

Before there was an AI company, there was a pair of trousers. Zayd Enam's grandfather had earned a place to study electrical engineering in Bihar, but he could not enter the laboratories in his kurta-pajamas. He needed different clothes. Friends helped him buy them. He went on to complete a doctorate and return as a professor. The obstacle had been absurdly small; its consequences could have been very large.

Years later, his grandson would discover that technology, too, sometimes arrives improperly dressed. An enterprise can admire a piece of software and still decline to buy it. The research may work. The procurement department may have other plans. Enam has spent much of his career trying to get past that particular doorway.

He co-founded Cresta, which brought AI coaching into customer conversations, and led it as CEO before handing over to Ping Wu in 2023. His next venture, Enam Co., carries his grandfather's name. Across the two companies sits a recurring question: how do you get useful knowledge into the hands of somebody who needs it now?

The route through community college

Enam's own education followed a less tidy route than the familiar Berkeley-to-Stanford shorthand suggests. He left high school in Pakistan to start a business. That first venture ultimately failed. He then moved to Southern California, worked at another startup, and took community college courses before transferring to Berkeley.

That sequence matters. A transfer student has already had to learn how an institution's doors work. There are prerequisites, rules, and somebody on the other side who decides whether you belong. By the time Enam reached Stanford's AI Lab, he had experience with both building something and finding a route into a place that could help him build it better.

At Berkeley he studied electrical engineering and computer science. His interests took him into research on unsupervised learning. At Stanford, his adviser was Sebastian Thrun, whose projects had already carried artificial intelligence out of academic settings and into the world. Enam wanted to do the same.

A useful machine has to survive Tuesday

His student blog offers a glimpse of the engineer beneath the founder. In an essay titled “Why is machine learning 'hard'?”, he argued that easier access to algorithms does not make implementation easy. Frameworks can spare you some mathematics. They cannot tell you why a model has failed to perform well enough.

He treated machine learning as a difficult debugging problem. When an ordinary program breaks, an engineer often has a fairly direct trail to follow. A model can disappoint for several interacting reasons, and the experiment that tests a proposed fix may take time. The waiting is part of the work.

It is an unfashionably practical starting point for an AI entrepreneur. A demonstration lasts a few minutes. A working system must keep producing useful answers after the demonstration ends. Enam's early writing shows an interest in that gap, before the business of selling across it became his full-time occupation.

First, help the people marking homework

The work that became Cresta began close to home. At Stanford, Enam worked on helping teaching assistants and graders give students more consistent feedback. The system looked for recurring mistakes and learned which responses helped. He later said it doubled the graders' speed.

A university classroom made a convenient place to test the idea. It also exposed its commercial limits. The team wanted to carry the approach into a wider range of knowledge work. Thrun's advice was to take it into the world, test it there, and bring those lessons back into the research.

The appeal was easy to understand. Experienced workers know things that are difficult to reduce to an instruction manual. A system that notices useful patterns could help a new colleague at the moment of uncertainty. Feedback would arrive while there was still time to use it.

The CEO with an intern's badge

Intuit supplied an unusually literal test of getting inside the customer. Enam's sales efforts had not gained much traction. After a presentation by Intuit co-founder Scott Cook, he sent Cook a cold email and secured a meeting with the company's CIO. The project fitted Intuit's AI strategy. A one-person supplier did not fit its requirements.

There was another offer: a summer internship. Enam accepted. Inside Intuit, he gained access to the data and deployed his software with a team in Tempe, Arizona. An experienced salesperson began using it and encouraged colleagues to follow. Enam recalled that the team's performance doubled within weeks.

When Intuit wanted to expand the project, he asked for an enterprise agreement. The contract also terminated his internship. He signed as Cresta's CEO and as Intuit's former intern. It is a pleasingly bureaucratic ending: the paperwork finally caught up with the work.

The episode gave a young company a customer and a case study. It also made the founder's method visible. When a buying process blocked the experiment, he took a job that let him conduct it.

A parrot, a Porsche, and a naming problem

The company's financing supplied another encounter with somebody else's conditions. Enam asked Andy Bechtolsheim to invest. Bechtolsheim wanted to meet before making a decision and was about to leave for Germany. They arranged to meet at Stanford's Gates Computer Science Building.

In Enam's recollection, Bechtolsheim arrived in a silver Porsche 911, wearing Birkenstocks with socks. The conversation took five minutes. He agreed to invest, then asked what the company was called.

Unfortunately, it was called Grey Parrot Inc. Enam had incorporated under a placeholder inspired by the bird's linguistic abilities. Bechtolsheim refused to wire the money until they found something better. Even a promising AI company can lose a round to a parrot.

Cresta emerged after more work on the name. Yoky Matsuoka helped connect the crest of a repeating wave with the idea of redirecting repetitive work toward creative activity. Bechtolsheim approved. The money followed. When Enam announced his next venture in 2024, he said Bechtolsheim had again written the first check.

The expertise already sitting in the room

Cresta's early proposition was grounded in the differences between people doing the same job. Some sales representatives ask better questions. Some handle objections more effectively. Their colleagues may have the same products and the same screen, yet come away with different results.

In a 2020 explanation of the business, Enam described reviewing a customer's previous conversations to identify responses associated with successful outcomes. Those examples trained a model that could prompt a representative during a live interaction. Coaching could also focus on the behaviors a particular worker needed to improve.

He paid attention to smaller conveniences, too: reducing keystrokes and making tedious tasks easier. Those details help explain the product's intended appeal to the person using it. A tool has to earn a place in somebody's working day, even when an executive has already bought it.

His public argument about AI followed the same line. At a DLD discussion in July 2021, alongside Snowflake's Jennifer Belissent and investor Ulrike Hoffmann-Burchardi, he challenged the idea that automation and job elimination exhausted the possibilities. He was interested in machines that could help people perform better.

Capital arrived; the title changed

Cresta's commercial progress brought larger investors into that experiment. A $50 million Series B led by Sequoia followed its emergence from stealth. In March 2022, the company announced an $80 million Series C led by Tiger Global, at a $1.6 billion valuation.

The latter round also included Zoom, Five9, and Genesys. Their participation placed companies already serving the contact center market among Cresta's backers. Enam had spent weeks courting those investors after Tiger offered to lead the round. The financing was also a chance to build relationships inside the industry.

In May 2023, Cresta announced Wu as CEO, succeeding Enam, who had stepped down into an advisory role. Wu had joined in 2021 and had been serving as interim chief executive. Enam expressed pride in the company's work and support for its next phase.

That handover belongs in his biography as much as the funding does. He had helped turn a research project into a business with enterprise customers. He then left the chief executive's seat and began another company. The Cresta co-founder remained; the job description changed.

A portrait ought to contain more than a face

Enam's curiosity also wandered outside company building. In February 2024, he proposed an art project about character. Photographs, he argued, concentrate on appearance while missing the events that reveal a person's nature. He wanted collaborators to help make work that could represent more of a life.

His reference point was Cubism: bringing several views together instead of settling for one flat representation. It is an interesting concern for somebody whose career is often compressed into a funding number and a university name. He was asking for the experiences behind the portrait.

The post was a proposal. Its appeal lies in the question itself. Which events deserve to be in the picture? For Enam, the unsuccessful first business, community college, and the Intuit internship all complicate the neat outline of a technology founder.

This time, buy the workplace

Enam Co. changes where he stands in relation to the problem. The company acquires services businesses and builds AI into their operations. Instead of selling a tool into a customer's organization, it takes responsibility for the organization delivering the service.

Its stated ambition is broader access to expertise. The company presents kindness, clarity, and intensity as its values, pairing technical work with a responsibility toward the people affected by it. Those are commitments to be tested through operating decisions, rather than a set of outcomes already achieved.

“We make expertise abundant.”

Enam Co.'s stated mission

There is a familiar cast around the new venture. Its public materials identify Bechtolsheim and Thrun among its backers, alongside Gil & Co. and former AT&T Communications CEO John Donovan. The people who helped Enam get started have remained part of the wider story.

In a March 2026 update from Pitztal, Enam described time in the mountains with his team and recruiting across San Francisco, Berlin, and Toronto. He emphasized learning from people closest to the work. The setting had changed, but the practical interest was recognizable.

The grandfather's trousers make an apt final detail. Talent had reached the laboratory door; a rule kept it outside until friends found a way through. Enam's companies have dealt with their own versions of that distance between capability and access. His next experiment puts him inside the businesses he wants to change, with considerably more to do than sign an internship form.