A serial founder, Ulupreneur Pratik Verma, co-founder of Okahu, says he got the bug from his father who started his first tech company when their family immigrated to the US from India.Pratik, who was finishing his PhD in chemistry at Stanford, decided to get involved.
While still pursuing his doctorate, he founded two companies. Pratik described several pivots before hitting the right niche with the second company, BlueTalon—a data security firm. There, he met and worked with Prasad Mujumdar until the company was acquired by Microsoft in 2019. When the pair had finished their vesting with Microsoft post-acquisition, they wanted to create a company together that builds upon their experience in effective use of data & AI in the cloud at scale.
In September 2023, they co-founded Okahu to help developers put AI to work in production by making it easy to monitor, identify and resolve issues like hallucinations in agentic or large language based AI apps.
Okahu’s initial focus is on enterprise developers building AI-enabled automated experiences for employees in workplace communication tools such as Microsoft Teams and Slack or in customer communication apps such as Instagram & WhatsApp.
We caught up with Pratik as the company was about to make a big funding announcement and “officially” come out of stealth.
What were your passions around building Okahu?
I learned two things after building BlueTalon. One is that building a company is much more fun when you enjoy working with the folks around you. And secondly, I learned that there are lots of things that can be created, but you need to ask yourself the question: What is one small thing that you can solve that will make lives easier for everyone?
When we started Okahu, we discovered that a lot of the things people face in ensuring proper data governance, they are going to face in building AI applications as well—so, how do we make natural language AI responses reliable? Tools that might have been used with previous generations of AI are not effective with large language models. When engineers start putting AI into real-world applications, where things stand today, they run into a problem. They need to figure out which AI responses are problematic, what is the impact of an error on the task being automated by AI and how to fix it.
At Okahu, we make it very easy for people to figure out where the things are breaking apart and offer suggestions on how to solve it.
So at the end of the day, people can actually benefit from AI rather than getting biased answers or what we call “hallucinations.”
We also wanted to work with people that we had worked with before and respected. The whole point of the fundraising was to get enough of the team together to work on an even more impactful problem the second time around. Each person on the team brought their unique point of view along with trust built on mutual respect.
What are your tips for fundraising founders?
We learned a lot through trial and error from previous fundraising efforts. We’ve learned to communicate the problem that we’re trying to solve in a way that anyone can easily understand our impact if we’re successful.. And secondly, we learned to be intentional and run fundraising like a process.When considering the type of investors we were seeking, I asked myself: “Okay, what is the goal? What are we trying to do with these funds? What value will we create that we will share with our partners?” We searched for people copacetic with the answer to these questions and weed out people who were not the right fit. Clearly, setting that intentionality allows you to be more successful.
What have been the biggest challenges in building the company?
The biggest challenge has been the saturation of the term “AI.” Everybody is investing in AI, every enterprise IT team is looking to add AI to their stack and in response everybody’s trying to put “AI” in their marketing or fundraising material. So, being clear about what our association with AI is and why we’re doing what we are doing in AI matters. That’s the key differentiation. Our pitch wasn’t just that we’re using AI; we’re saying people are building AI apps already, so there’s a workflow already there and we are simplifying monitoring. Being able to distinguish and articulate what we are doing differently—and why our particular expertise is uniquely qualified to address that particular problem—helped with our fundraising efforts.
What’s it been like working with Ulu?
I’ve gotten a chance to spend a lot of time with Clint and Steve. I appreciate Clint’s methodical approach towards building a company and systematically figuring out what drives the leading indicators of success.
There aren’t many investors that I know that have that clear, systematic way of decision making and coaching founders. With many investors, it’s more like a “gut feeling.”
Steve has been very accessible and supportive in resolving operational challenges. . When I recently shared a problem with him about having a tough time setting up the legal infrastructure for the team in India, his answer was, “I know someone else who’s done it before.” And within a couple of days, he was able to get me to a trusted person I could have an open conversation with. There’s a lot of things that I have to think about in building Okahu that I’m not necessarily an expert on. Steve has been willing to jump in where he can and help think through the problem; those are two attributes I found to be uniquely valuable.
What are your dreams for Okahu?
The dream is to ensure that any AI out in the world is trustworthy and reliable. We want to make sure people have safe, secure AI and that it’s making everybody’s life easier, rather than more difficult. And while we’re doing that, I want to work with all my friends and make new friends as they come into the company. And ten or twelve years from now, hopefully the company will take on a self-sustaining life of its own … in whatever shape or form that takes.
