Keebo AI optimizes data warehouses with automated ‘learning’ platform, raises $10.5M

With the rise of Snowflake and other cloud data warehouses, enterprises finally have a simple way to mobilize their data assets at scale. They can easily connect data from different sources and start driving efficiencies while keeping upfront investments (or CapEx) on the lower side.

The benefits of the solutions are unparalleled, but cloud data services also come with the challenge of high operating expenses. Essentially, due to constantly growing datasets, companies have to deal with high compute costs and query performance latencies. Without a solution, their teams have to give about 30-40% of their time to manually develop features that could optimize the warehouse for the required performance and budget constraints.

The time invested in manual processes is engineering resources that could otherwise have been used in other critical areas with clear business value, such as DataOps and cloud infrastructure.

Keebo automates data warehouse optimization

To address the challenge, Michigan-based Keebo AI offers a data learning platform that makes the entire process of optimizing the cost and performance of data warehouses intelligent and automated. Today, the company announced $10.5 million in a series A round of funding.

“Keebo…builds a set of ‘smart models’ by learning how users and applications interact with data over time and uses those smart models to automate and accelerate the tedious aspects of those interactions,” said Barzan Mozafari, CEO and cofounder of Keebo. “… For example, instead of just answering the user’s query at hand, Keebo learns from each query how to accelerate the next queries, even if those queries are seemingly quite different in nature. And it does all this learning simply by looking at performance logs and meta-data instead of the customer’s actual data.”

The turn-key and drop-in solution first debuted in 2021 as a semi-automated on-premise beta offering. Now, with this round, the company has expanded to fully automated warehouse optimization. The hosted capability, the company claims, can quickly identify opportunities to make data processing more efficient with less cloud usage, enabling businesses to reduce the cost of their cloud warehousing on average by 30-60%. 

Keebo claims that its platform has already helped many companies accelerate their analytical queries by up to 100 times. Its customer base includes players like Allbirds, TUI, Barstool Sports, PayJoy, 14 West, HyperScience and Dr. Squatch.

“Keebo takes care of the things that I don’t want to think about or deal with,” said Trish Pham, head of analytics at PayJoy. “It requires no work on my end. Even if I loved manual optimization, I couldn’t possibly achieve what Keebo achieves automatically. I log in for a few minutes every few weeks just to see what Keebo is saving us. I used to spend hours on manual optimizations every week.” 

The entire implementation effort of the platform takes 30 minutes from start to finish, which mostly involves creating the user and pointing the platform to the workloads that need to be optimized. It’s “set it and forget it”, meaning after the initial onboarding, there is no additional maintenance of implementation or learning curve. Users can also go in and set their objectives, like reducing compute bills or speeding up queries on the data warehouse.

Plan ahead

With this round of funding, which was led by True Ventures, Keebo’s total capital raised now totals $15 million. Mozafari said the company will use the fresh round to grow its team and further improve the product with three core capabilities.

“First, we want to close the gap between optimized and optimal performance. Second, we plan to extend our learning platform to support a wider range of data warehouse stacks. And third, we plan on reducing the time-to-value for our customers from 24 hours down to three hours,” he added.

Globally, the data warehousing market is expected to grow from $21.18 billion in 2019 to $51.18 billion by 2028.