Teradata on Tuesday announced a new integration with DataRobot that enables customers to develop AI models and applications using DataRobot's AI capabilities within Teradata's VantageCloud platform.
San Diego-based Teradata is a longtime analytics and data management vendor that offers VantageCloud for data storage and preparation and ClearScape Analytics for business intelligence.
Meanwhile, DataRobot is a Boston-based AI vendor that offers both traditional and generative AI models that customers can use in conjunction with their own data.
According to David Menninger, analyst at ISG's Ventana Research, enterprise interest in AI, including generative AI, is exploding, and because DataRobot has a different approach to AI and machine learning analytics than Teradata, the integration between the vendors would be beneficial to their respective customers.
Teradata already supports AI and machine learning, but with a programmatic approach, he noted, while DataRobot offers a graphical user experience. As a result, the vendors' tools complement each other.
“Every data scientist has their preferred approach,” says Menninger. “This partnership enables Teradata to support a broader range of users and gives DataRobot customers tighter integration with a robust platform for running analytics.”
In addition to DataRobot, Teradata has similar integrations with AI vendors H2O, Dataiku, and open-source PMML and ONNX AI and machine learning models.
“There's more to come,” said Michael Riordan, senior director of data science and analytics product management at Teradata.
Teradata first released VantageCloud and ClearScape Analytics in August 2022, just three months after OpenAI's ChatGPT hit the market and sparked an ever-growing interest in both generative and traditional AI.
Since then, Teradata, like many other data management and analytics vendors, has made AI a key focus of its product development. For example, in April the vendor partnered with Anaconda to strengthen open source support for VantageCloud's AI innovations. Later that month, Teradata introduced support for the open table formats Apache Iceberg and Delta Lake, building an ecosystem for AI development.
Integration
Since ChatGPT's launch, data management and analytics vendors have made generative AI a focus of their product development, given its potential to empower non-technical people to process data while also improving the efficiency of data experts.
The use of analytics within organizations has been stagnant for nearly two decades, hovering around a quarter of all employees, simply because analytics and data management platforms are difficult to use, require coding to perform most tasks, and require data literacy training to interpret the results.
Natural language processing (NLP) and low-code/no-code tools promised to broaden the use of analytics, but their capabilities have been limited: NLP tools are limited by their small vocabulary, and low-code/no-code capabilities only allow users to perform basic tasks.
But with generative AI, large language models have vocabularies as large as dictionaries, enabling true natural language interaction and even inferring intent. This combination, combined with a company's proprietary data, allows nearly anyone to query and analyze the data to inform decisions.
Meanwhile, generative AI reduces coding requirements, freeing data experts from time-consuming, repetitive tasks so they can do more with their time.
Teradata and DataRobot's new integration enables organizations to simplify the integration of data with AI and machine learning by applying DataRobot pre-built models to data in VantageCloud. The integration is accessible through Teradata's ClearScape Analytics Bring Your Own Model (BYOM) capability, which allows customers to choose their preferred models for AI development.
According to Donald Farmer, founder and president of TreeHIve Strategy, selecting DataRobot’s model will enable users to import and operationalize DataRobot’s pre-built AI models to accelerate their own AI development efforts in a way that benefits existing Teradata users.
“This will give Teradata users a powerful technology to build models faster and score them more efficiently than they can today,” he said.
Specific benefits and features of the integration include:
- It gives data scientists more choice to use the platform of their choice when developing predictive AI, generative AI and machine learning models and applications.
- Model deployment in a secure and trusted environment, including all cloud providers and on-premise, allowing you to scale with your enterprise needs while providing a clear cost structure.
- BYOM allows companies to access DataRobot-provided models as well as models they have developed themselves with DataRobot from within Vantage Cloud.
Farmer said the combination will benefit existing Teradata customers, but the real significance of Teradata's move will be in whether it helps attract new customers.
He said Teradata has struggled to attract new users because it is overshadowed by Snowflake, which offers similar data management and analytics capabilities within its platform, but Snowflake replaced its CEO in March and fell victim to a cyberattack in late May.
Despite increased investments in AI under new CEO Sridhar Ramaswamy and a hack that was caused not by a security flaw but by insecure customer logins, trust in the vendor has taken a hit and the company's shares are well below levels before former CEO Frank Slootman stepped down.
Snowflake's struggles may present an opportunity for Teradata.
“The big question for Teradata is how to find new greenfield customers,” Farmer said.[The integration] This is a win for existing customers, but does this make Teradata an attractive data platform for DataRobot customers? If so, this is an important move.”
But otherwise, he continued, the integration provides ease of use for existing Teradata customers but nothing more.
“That's great on paper and a nice thing to have, but it's not going to change anything big,” Farmer said. “Ultimately, if this makes Teradata a deeply integrated and robust enterprise data platform for DataRobot users and helps it attract more new customers, that's potentially something different and better.”
Menninger similarly said that while Teradata offers powerful data management and analytics tools, the company has been overshadowed by Snowflake and its rival Databricks.
Teradata continues to invest in a modern architecture that is available on all major cloud providers, which means that from a technology perspective, Teradata stacks up well against competing vendors and has earned high marks in ISG's Data Platform and AI Platform Buyer's Guides.
“In terms of revenue, we've been overtaken by vendors like Snowflake and Databricks. [but] “They are still considered the leader in terms of functionality, manageability, scalability and reliability,” Menninger said.
Our goal
According to Riordan, one of Teradata's main goals is to remain open and connected by integrating with a variety of other vendors and operating in a variety of environments to give customers choice.
The integration with DataRobot further strengthens that openness and connectivity.
Going forward, Teradata, through VantageCloud and ClearScape Analytics, will continue to explore ways to help companies operationalize their AI and machine learning efforts, Riordan said. Specific efforts include improving ClearScape's deep learning and language model capabilities, as well as upgrading the platform's user experience for model development and deployment.
Menninger said the continued focus on AI, particularly developing generative AI tools such as the ask.ai assistant that simplify use of the platform, is key to keeping up with other vendors' offerings.
“All data platform providers are in the process of adding GenAI tools to make their products easier to use,” Menninger said. “Teradata has started down this path… but we expect the competitive bar will continue to rise as more data and analytical capabilities are supported by GenAI technology.”
Meanwhile, Farmer stressed that Teradata needs to do more to grow its customer base.
Farmer said the vendor has focused less on acquiring new customers in recent years and more on meeting the needs of existing customers — a strategy that's been sound for a while, but at some point it needs to change — and with Snowflake struggling, that time may be now.
“I'd like to see them move back into a greenfield strategy and start to become a more agile platform for people who need to scale enterprise-class data management,” Farmer said. “I think they have an opportunity because of the Snowflake issues, and if they can move quickly, they have an opportunity.”
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with over 25 years of experience, focusing on analysis and data management.
