“There is no AI without data, there is no AI without unstructured data, and there is no AI without unstructured data,” said Chet Kapur, chairman and CEO of data management company DataStax. There is no AI without it.”
Kapur started a conversation at TechCrunch Disrupt 2024 about “new data pipelines” in the context of modern AI applications, and was joined by Vanessa Larco, a partner at VC firm NEA. George Fraser, CEO of data integration platform Fivetran. The chat covered multiple topics, including the importance of data quality and the role of real-time data in generative AI, but one of the big takeaways was that it's still early days to prioritize product-market fit over scale. It was about the importance of prioritizing. A.I. The advice for companies looking to jump into the fast-paced world of generative AI is straightforward. Don't be overly ambitious at first, focus on practical, incremental progress. reason? We're still figuring it all out.
“The most important thing about generative AI is that it's all about humans,” says Kapur. “The SWAT teams that actually go out and build their first few projects aren't reading manuals. They're writing manuals on how to build generative AI apps.”
It's true that data and AI are closely related, but it's easy to feel overwhelmed by the sheer amount of data that companies have. Some of it may be sensitive, subject to strict protection, and may be stored in countless locations. Larco, who has worked with and served on the boards of numerous startups across the B2C and B2B sectors, proposed a simple but practical approach to unlocking true value in the early stages.
“Think backwards at what you're trying to accomplish. What are you trying to solve and what data do you need?” Larco said. “We will find that data wherever it is and use it for this purpose.”
This is in contrast to applying generative AI across your company from the start, throwing all your data into a large-scale language model (LLM) and hoping it spits out the right thing at the end. That's likely to lead to inaccurate and expensive confusion, Larco said. “Start small,” she said. “What we're seeing is companies starting small with internal applications with very specific goals and then finding data that matches what they're trying to accomplish. That’s it.”
Fraser, who has led data movement platform Fivetran since its founding 12 years ago and amassed major customers such as OpenAI and Salesforce along the way, believes companies should focus on the real problems they face today. He suggested that.
“Just solve the problems you have today. That's the mantra,” Fraser said. “99% of the cost of innovation is always in the things we build that don’t work, not the things that worked out that we had planned to scale upfront. These are the things we always look back on. However, it does not account for 99% of the costs incurred by customers.
Much like the early days of the web, and more recently the smartphone revolution, the early applications and use cases of generative AI offer glimpses of a powerful new future powered by AI. But so far, they aren't necessarily game-changing.
“I call this the Angry Birds era of generative AI,” Kapur said. “My life hasn't completely changed, but no one does my laundry yet. This year, every company I work with is putting something into production. Small scale. So it's internal, but the reason we're bringing it into production is because next year I'm solving the problem of how do we put together a team to actually make this happen? In what we're calling the year of transformation, people will start using apps to actually change the trajectory of the companies they work for.”