Reltio targets trusted data with GenAI, prebuilt tools

New capabilities

Trusted data is imperative for the simple reason that analytics and AI models and applications are built on data.

If the data used to train the models and applications is of high quality, the outputs delivered by the models and applications will also be high quality. If data quality is poor, the outputs of analytics and AI models and applications trained on that data will also be poor.

But there's more now to trusted data than merely enabling users to have faith in the outputs delivered by their analytics and AI products, according to Stewart Bond, an analyst at IDC.

Data quality and the resulting trust -- or lack thereof -- in data has always been important. And it has always been an issue, whether used to inform a simple report or a complex machine learning application.

Now, however, many enterprises are developing AI models, including generative AI, that rely almost completely on automation, from data ingestion through curation and training to analysis. Both the volume and complexity of data needed to train generative AI tools to understand an organization's operations make it impossible for humans to be involved in every step of the process to check data quality as data progresses through pipelines.

While humans need to be involved to ensure accuracy, machines are far more efficient at making unstructured data usable and observing the data that feeds the retrieval-augmented generation pipelines used to inform AI models. As a result, the foundation of data must be high quality from the start for modern models and applications to be trustworthy.

"There is [always] a possibility that bad data won't be caught. [But it] will occur at a much higher rate with GenAI as tasks become automated and people start to get pulled out of the loop," Bond said. "The quality of the data that the model is trained on and used at time of inference via RAG or embedding can improve the accuracy and relevancy of the outcome."

Reltio's platform update addresses trusted data in three ways, according to Venki Subramanian, the vendor's senior vice president of product management.

Data quality, the completeness of data and the timeliness of data are all important to facilitating trust, he noted. And Reltio's update consists of capabilities aimed at quality, completeness and timeliness.

"The capabilities of this release help organizations ensure that high quality, comprehensive data across the most important dimensions of the business -- customer, product and supplier -- is quickly unified and ready to mobilize across different teams," Subramanian said.

The new Customer 360 features, including audience segmentation and activation capabilities, address the timeliness of data, aiming to provide Reltio users with the right data at the right time.

A new audience segmentation user interface enables customers to discover and curate relevant data, while Reltio Intelligent Assistant enables conversational interactions with data so users can explore customer data and create up-to-date segments within that data.

Velocity packs, which are prebuilt industry-specific master data management capabilities, similarly address timeliness while also targeting data quality.

Reltio first introduced velocity packs -- data models and other prebuilt components so users can develop a master data management program more quickly than with a general-purpose platform -- in February 2023. Numerous vendors such as Databricks, SAS, SAP and Snowflake have developed industry-specific offerings to save customers some of the work it takes to configure their data systems and Reltio is following suit.

The first two velocity packs were for life sciences and healthcare. Since then Reltio has added velocity packs for insurance, financial services and B2B data.

The latest version continues the company's strategy of providing predefined data models and data integration configurations as well as AI capabilities to accelerate adoption and facilitate business engagement with trusted data.

The vendor's 2024.2 update adds velocity packs for suppliers and product inventory.

Lastly, the Reltio Business Critical Edition targets real-time data quality. The edition is a premium offering for Reltio Multidomain MDM and 360 Data Product users that aims to ensure mission critical data is consistently delivered without delays. Included are features such as disaster recovery configurations, Reltio Shield for data encryption and AWS PrivateLink for secure data transfers.

Bond noted that Reltio has been repositioning itself in recent years, expanding beyond its original focus on master data management to include a focus on data as a product with tools such as the velocity packs and customer insights with its Customer 360 features.

The new capabilities, while still addressing master data management, advance that evolution, according to Bond.

"The updates Reltio is providing with this release are significant because they are making the repositioning of the Reltio capabilities real," he said. "That is, Reltio is known as a master data management solution. But it has recently split its capabilities into … pillars that provide customers with choice and a path of maturity."

Aslett, meanwhile, highlighted the velocity packs and Customer 360 additions.

The intent of Reltio's platform is to provide trusted data by unifying and cleansing complex data from numerous sources, he noted. In addition, the platform is built to unify and cleanse that data in real time using machine learning and automated data validation rules.

"The latest version continues the company's strategy of providing predefined data models and data integration configurations, as well as AI capabilities, to accelerate adoption and facilitate business engagement with trusted data," Aslett said.

The velocity packs accelerate return on investment while the new Customer 360 features reduce the time it takes to reach actionable insights, he added.

While Reltio's updated platform aims to better enable users to trust their data, the impetus for developing the new capabilities was a mix of customer feedback and the logical evolution of the vendor's pre-existing tools, according to Subramanian.

The Business Critical Edition, enabling users to securely run mission critical data on Reltio, was something customers requested while the customer segmentations capabilities and new velocity packs build on what the vendor already offered.

Next steps

Reltio provides three major platform updates per year.

Just as the 2024.2 release targeted trusted data with improved data unification and AI capabilities, the vendor's next platform update will add more unification and AI features to deliver trusted data in real time, according to Subramanian.

As Reltio adds that functionality, it would also be wise to do more to clearly market the tools it provides, according to Aslett.

The vendor's focus on areas such as master data management, customer intelligence and data unification -- each of which is often the single focus of a vendor -- makes Reltio somewhat unique and difficult to classify. As a result, clear messaging could be beneficial.

"While the depth and breadth of functionality offered by Reltio is impressive, it also arguably results in the company's platform not neatly fitting into traditional product categories," Aslett said. "As such, the onus is on Reltio to articulate the strategic business benefits of data unification to improve trust in data."

Bond, meanwhile, noted that just as Reltio expanded beyond master data management, there are further opportunities to add new functions.

Adding more generative AI capabilities is one means every vendor can expand. But given its base in master data, which exists in every B2B or B2C interaction, there are myriad additional ways Reltio could broaden its platform to attract new users.

"Continued development of GenAI capabilities within the platform is obvious. But there are also opportunities to continue to expand the platform into adjacent data intelligence, observability and metrics capabilities," Bond said.

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.