Prasad Kawthekar is the Co-Founder & CEO of Dashworks, an all-in-one AI search assistant. It enables users to find any document, message, or email and to accelerate a team's productivity. With over 40+ integrations, Dashworks enables teams to find and organize their internal knowledge across apps from one place and leverage their collective expertise.
What initially attracted you to computer science and machine learning?
When I started college, I expected to major in Math and Physics and enter academia. I took an intro Java programming course in my second semester and spent time building a few games. I was addicted. I loved getting real-time feedback on what I was building. In my second year, I took a Machine Learning course and became hooked on the question of how machines think. I then did a couple of research projects on active learning, which is the concept of machines asking questions so that they can learn faster. I also experimented with a few different areas of machine learning such as robotics, computer vision, and drug discovery before finally settling on Natural Language Processing for my graduate degree at Stanford. I love NLP because language is the root of intelligence and it’s been the hardest to crack.
You were one of the Founding Engineers at Cresta, what were some of your key takeaways from this experience?
I joined Cresta when there were only two people on the team. Now, of course, it’s a Series C company with a $1.7B valuation. Perhaps the most important thing I learned from my time there was the importance of customer obsession. They once flew out the entire team to a customer’s call center and made us work from there for a week. That experience shaped how we thought about the solutions we were designing. Customer centricity is now one of our guiding principles at Dashworks.
Could you share the genesis story behind Dashworks?
After Cresta, I began working at a company called Blue River. They’d been recently acquired by a Fortune 500 company, and I was asked to start working on a new project related to credit scores for lending. It took me seven email threads and eight weeks to find the right documents!
That’s when I realized the true cost of scattered information at a company. Soon after, I convinced my co-founder, Praty Sharma, to jump on board. We met at Stanford during grad school. At the time, he was working at Facebook on their Marketplace search engine.
The pain of trying to find information at work just seemed in stark contrast to how easy it is to find information on the web, where you can practically find answers to anything in the public domain in less than a second.
Dashworks enables teams to find and organize their internal knowledge across apps from one place and leverage their collective expertise. What are some of the technical challenges behind this?
Enterprise search has always been a hard problem to solve at scale. Compared to web search, you don’t have as much training data. This creates a bit of a cold start problem. You need data points to refine the model but companies don’t want to use you unless you’re guaranteeing accuracy. You’re also integrating a lot of different apps, each of which store their data in different formats. This makes indexing much harder and more time consuming. Finally, there is a lot of sensitivity around data security and permissions.
Traditional approaches to solving this problem have relied on indexing all of a company’s data, but we’ve been experimenting with different approaches that have only become recently possible because of the availability of fast and large context AI models that allow us to completely bypass indexing.
The Dashwork vision is making humanity omniscient, could you describe what this means for you?
We envision a world in which a human being is able to find any piece of information they need, no matter how specific or obscure. At the end of the day, information is power and we’re hoping to make humanity more powerful.
Web search engines are focused on democratizing publicly available information. But 96% of the world’s information is actually on intranets. And some unknown percentage of that information is behind the intranets of smaller companies. Until now, entrepreneurs have been focusing on solving the enterprise search problem for large companies with big budgets. But we’re focused on starting with smaller companies and working upwards.
On Twitter, you recently stated, “The simple text box is the interface of the future. Dashworks is designed entirely around it.” Could you elaborate on this statement on how you view this future, and how Dashworks is designing around it?
Graphical User Interfaces or GUIs came into existence because non-developers needed a way to interact with computers. So you saw the rise of big buttons, toggles, checkboxes, icons etc. The key constraint was needing to know a programming language to interact directly with a machine. But with the rise of AI, machines can understand natural language. And natural language is still the most intuitive way for humans to communicate. Which means that we envision a future in which you can execute pretty much any action – no matter how complex – with a simple natural language prompt.
That’s what we’re building toward at Dashworks. In the immediate future, it means reimagining things like user onboarding. What would it look like if onboarding happened entirely via a conversation with a machine instead of buttons and checklists?
What are the different types of integrations that Dashworks offers?
We support over 50 integrations across categories like Wikis, CRM, customer support, calendars, HRIS, developer tools, and project management. We integrate with applications including Notion, Google Suite, Confluence, Airtable, Asana, Intercom, Github, Microsoft Teams and more.
What are the types of productivity gains are seen from enterprises that integrate the Dashworks platform?
Our customers tend to see big gains in productivity and efficiency in a few ways:
- New hires get up to speed faster because they have all the information they need in a single place
- Employees spend less time answering repetitive questions, especially on Slack
- Projects move faster and collaboration is easier because all employees have access to the information they need and there are fewer silos
- Employees get more done because they don’t have to context switch as much and they spend less time searching for information
- The quality of decision-making improves because employees are able to gain context and a full picture on matters
Thank you for the great interview, readers who wish to learn more should visit Dashworks.
- Vianai’s New Open-Source Solution Tackles AI’s Hallucination Problem
- AI & AR are Driving Data Demand – Open Source Hardware is Meeting the Challenge
- What is a ChatGPT Persona?
- PyCharm vs. Spyder: Choosing the Right Python IDE
- “Brainless” Soft Robot Navigates Complex Environments in Robotics Breakthrough