DataRobot aims AI Cloud 8.0 at 'unpredictable market'
The updated platform enables organizations to adapt to fast-changing market conditions and perform continuous AI operations. The platform also includes integrations with Snowflake.
Independent AI vendor DataRobot updated its AI cloud platform with features it said will enable businesses to better deal with changing market conditions and continuously optimize their AI systems.
Among the new capabilities in DataRobot's AI Cloud 8.0, released on March 17, is AI App Builder, which the vendor said will enable organizations to better adapt to what it called an "unpredictable market landscape."
For the past couple of years, organizations have had to deal with a pandemic, unreliable supply chains, and plans to bring workers back to the office or operate with hybrid workforces. All of these have contributed to "uncharted market challenges," the vendor said.
The updated platform also includes a continuous AI cloud capability that recommends the best models for organizations to use to predict desired business results.
Continuous AI
DataRobot is providing a mechanism for enterprises that have difficulties continuously training and updating models to keep models from going stale, said Andy Thurai, an analyst at Constellation Research.
"AI models skew over time," Thurai said.
AI models skew over time.
Andy ThuraiAnalyst, Constellation Research
When data scientists create models based on historical data, those models will need to be continuously updated or they will lose their value, he said.
"DataRobot AI Cloud 8.0 is aiming to address that problem," Thurai continued. "They may be successful given their past history, but it's hard to tell."
The process of generating AI models, evaluating them to determine which is best and then putting them into production is not only hard, but can also take years, said David Menninger, an analyst at Ventana Research.
"As recently as two years ago, companies were updating their models quarterly," he said. "[It's] such a time-consuming process."
So a capability that will help enterprises automatically generate challenger models and replace existing models easily with no downtime will enable enterprises to save time, Menninger said.
One challenge for DataRobot AI Cloud 8.0 is how data scientists might receive some of the upgrades.
Data scientists like to work with power tools, Menninger said, noting that AI Cloud 8.0 reinforces DataRobot's self-service approach that enables users at times to create AI models without needing to bring in data scientists.
"Will data scientists accept the work that DataRobot does? Will they resent what DataRobot does?" Menninger asked. "They have to find a way to walk that fine line of engaging the data science community while also making it easier to use data science."
However, DataRobot appears to be aware of that dilemma and recently has made efforts to engage data scientists. Last month, the vendor introduced two new offerings for data scientists: Core and DataRobot 7.3.
DataRobot AI Cloud 8.0 is available now for deployment on public clouds, on premises and at the edge. DataRobot did not provide pricing information.
