Teradata ClearScape Analytics update targets ROI on AI, ML
The vendor's latest platform update adds features such as a PySpark conversion tool and AutoML aimed at helping customers improve returns on their investments in AI and ML models.
Teradata on Tuesday unveiled new features for ClearScape Analytics aimed at enabling customers to get better returns on their AI and machine learning investments.
Too often, AI and machine learning operations are inefficient, the vendor noted. As a result, enterprises are unable to derive valuable insights from their data while spending large amounts of money on expensive tools and processes.
To address that imbalance, among other new features, Teradata's latest platform update includes a new tool aimed at eliminating costly data movement; AutoML to enable data scientists to automate model training; and an integration with open source analytics vendor Knime that lets users more easily develop data science workflows.
Collectively, the new features address their intended purpose of simplifying AI and machine learning so users can derive more value from AI and ML initiatives, according to Donald Farmer, founder and principal of TreeHive Strategy. As a result, the features are significant additions for Teradata customers.
"This release is a good attempt to address some recurring issues such as complexity, cost, scalability and productivity," he said. "There are also tools for easier adoption and operationalization, so that seems good, too."
Teradata's customers can definitely expect to see accelerated time-to-value and improved ROI. Collectively, the new features and enhancements simplify AI/ML workflows and boost productivity throughout the AI lifecycle.
Mike LeoneAnalyst, Enterprise Strategy Group
Mike Leone, an analyst at TechTarget's Enterprise Strategy Group, similarly noted that the new ClearScape Analytics features position Teradata's update to help enterprises derive better outcomes from their AI and machine learning investments.
"Teradata's customers can definitely expect to see accelerated time-to-value and improved ROI," he said. "Collectively, the new features and enhancements simplify AI/ML workflows and boost productivity throughout the AI lifecycle, which can help virtually all enterprises operationalize AI at scale while engaging both technical and non-technical users."
Based in San Diego, Teradata is a longtime analytics and data management vendor. ClearScape Analytics, first launched in August 2022, is the vendor's main business intelligence platform, while VantageCloud is its platform for storing and preparing data.
As a data platform vendor that provides both analytics and data management suites, the vendor's competitors include Databricks and Snowflake.
Recently, Teradata introduced an integration with AI vendor DataRobot aimed at enabling customers to develop AI models and applications using DataRobot's tools in conjunction with VantageCloud. In May 2023, Teradata unveiled a similar integration with Google's Vertex AI platform also aimed at providing users with the tools to build AI models and applications.
New capabilities
Spurred by the explosion of interest in generative AI over the past two years, enterprise interest in traditional AI and machine learning is also on the rise.
Generative AI has the potential to make data management and analytics more efficient by automating repetitive tasks that previously had to be done manually. It also manages processes such as data integration and observability that otherwise need to be managed by humans.
In addition, due to its natural language processing capabilities, generative AI eliminates much of the coding previously required to work with data management and analytics tools, thus making them more available to nontechnical users.
As a result of those benefits, many enterprises are developing generative AI tools.
Simultaneously, investments in traditional AI and machine learning are increasing, fueled, in part, by generative AI, which makes traditional AI and machine learning easier to develop by generating code and taking on other development tasks. In addition, more compute power is now available, which is key for handling AI and machine learning workloads, and enterprises now have the requisite data to accurately train AI and machine learning models.
However, more than three quarters of all machine learning models reportedly never make it into production, resulting in lost time and money with no return on investment.
Teradata is attempting to change that with its ClearScape Analytics update by reducing the complexity generally associated with developing and deploying AI and machine learning models, according to Michael Riordan, the vendor's senior director of product management for data science and analytics.
"This year we've had a focus on helping customers leverage ClearScape functions to maximize their investments, reduce complexity and drive outcomes," he said.
Toward those ends, the ClearScape Analytics update includes pyspark2teradataml, a tool that eliminates the need for data movement by enabling users to convert PySpark code to Teradata machine learning.
Previously, many customers had to export data to Apache Spark platforms to write the Python code required to build models. By eliminating that requirement and enabling users to write code in ClearScape Analytics, Teradata is simplifying model development as well as lowering the costs associated with creating models, according to the vendor.
Beyond the PySpark conversion tool, AutoML aims to enable data scientists to automate the model training process. By doing so, the tool will save data scientists from time-consuming tasks, making them more efficient, while enabling some less technical users to build AI and machine learning models.
Similarly, an integration between ClearScape Analytics and VantageCloud with Knime is aimed at enabling more than data scientists and other experts to build AI and machine learning tools.
Knime's open source low-code/no-code data science platform is suitable for data scientists but can also be used by less technical workers, which should let enterprises develop AI and machine learning tools more quickly while also limiting costs.
Finally, Teradata's update includes improvements to the user experience to better enable self-service access as well as Teradata open source ML to use open source machine learning tools on VantageCloud.
While the new features collectively aim to better enable enterprises to reap the benefits of their investments in AI and machine learning, AutoML is perhaps the highlight of Teradata's update, according to Leone.
Vendors such as Qlik and Alteryx are among other analytics and data management vendors that provide automated machine learning capabilities. Teradata's entry into autoML nevertheless stands to benefit the vendor's customers by enabling more employees, not only experts, to build and train models.
"Even with AutoML features being widely available for some years, this feature stands out the most," Leone said. "This capability enables faster time to production and expands AI capabilities to nontechnical users. With businesses [emphasizing] operational efficiency and getting models into production, AutoML's ability to streamline workflows directly addresses these areas."
Farmer likewise called out AutoML as a key new feature while noting that the PySpark conversion tool also has significant potential.
By reducing complexity, the PySpark conversion tools meets the needs of users of both Spark and Teradata, he said.
Meanwhile, by reducing the technical requirements needed to build AI and machine learning models, AutoML is a change for Teradata. It has traditionally targeted its platform toward experts, while enabling more widespread use, he continued.
"By reducing the technical barriers to entry, it can potentially accelerate AI adoption and innovation across different business units, especially those with limited data science resources," Farmer said.
Next steps
As Teradata continues to try to help enterprises get better returns on AI and machine learning initiatives, adding vector search and storage is a focal point for the vendor, according to Riordan.
Data analysis has historically focused on structured data such as financial records and point-of-sale transactions. Now, however, the vast majority of data is unstructured, such as text, audio files and images.
To access unstructured data -- which is needed to get a comprehensive view of an organization -- and put it to use training models and applications, it needs some form of structure. Vectors are numerical representations of data that can be applied to unstructured data to give it that structure so it can be searched and discovered.
Once Teradata adds vector search and storage capabilities, it will better enable customers to combine structured data with unstructured data to develop more accurate analytics and AI tools.
"A key foundation for our AI plans is bringing high-performance vector data to Vantage," Riordan said.
One way to do that is through integrations with partners, he continued.
"Integration of vector data from unstructured data sources … and combining that with the rich data already available in Vantage will bring analytic workflows to new heights," Riordan said.
Integrations to add capabilities that enable AI development are a wise area of focus, according to Leone.
"Teradata should be focusing a good amount on enhancing its AI ecosystem," he said. "The Knime integration is a great step towards doing just that."
In addition, adding prebuilt tools to help new customers get started and accelerate time-to-value for existing customers could be another way for Teradata to improve its offerings, Leone added.
Farmer, meanwhile, also suggested that Teradata could do more to expand its ecosystem for AI development. In particular, more focus on the open source community could be beneficial.
The integration with Knime and addition of Open-source ML address Teradata's need to expand its open source ecosystem, according to Farmer. But competing vendors such as data platform vendor Databricks have closer relationships to the open source community and more extensive integrations with open source platforms, according to Farmer.
"Although Teradata has improved its support for open-source tools and languages, companies like Databricks have a stronger heritage in the open source big data ecosystem," he said. "Teradata could do more."
Likewise, it could do more to add pricing options for non-enterprise users, Farmer continued.
Teradata uses a consumption-based pricing model with a monthly subscription starting at $4,400 or an annual subscription starting at $52,600.
"Teradata has traditionally been seen as a premium, enterprise-focused solution," Farmer said. "While they've introduced more flexible pricing options, competitors are still perceived as more cost-effective for smaller-scale use cases."
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.
