Ataccama targets business users with latest AI capabilities

New capabilities

With enterprises increasing their investments in AI projects since OpenAI's November 2022 launch of ChatGPT, which represented a significant improvement in generative AI (GenAI) technology, data quality is becoming increasingly important.

While always vital to the success of any analytics or AI project, there was more human involvement in decision-making before the proliferation of GenAI and the more recent emergence of fully autonomous agentic AI applications. With humans involved, decisions based on bad data could be prevented. But with AI tools sometimes entrusted to act independently, the fail-safe of human oversight is decreased and sometimes eliminated.

As a result, the data informing GenAI and agentic applications needs to be high-quality more than ever.

Ataccama's new AI-powered data lineage capabilities provide natural language descriptions of how data was transformed throughout its lifecycle. As part of that, it explains filters, joins and calculations so that business users can understand the logic behind the descriptions.

Because the latest Ataccama update is designed to make it easier for users to understand their data so that it can be trusted, it adds valuable capabilities, according to Matt Aslett, an analyst at ISG Software Research.

"Ataccama has significantly enhanced its data lineage capabilities, enabling users to better understand the properties of data assets and how they are used across the organization," he said. "The new capabilities will better enable business users and data leaders to understand the lineage of data assets without needing deep technical expertise."

Meanwhile, with enterprises investing more in AI development, transparent representations of data lineage can help AI and machine learning governance by showing how data lineage inputs -- data's original ingestion point -- relate to users, pipelines and model outputs, according to Petrie.

"This makes AI architectures more transparent and models more explainable to key stakeholders such as customers or auditors," he said.

Feedback provided part of the impetus for adding natural language descriptions of data lineage calculations, with customers requesting that business users have the same data lineage information as technical experts, according to Jessica Smith, Ataccama's vice president of data quality.

"Customers told us they needed those insights to be just as accessible to business users -- clear, fast and usable without technical help," she said. "It's about removing friction for nontechnical teams that need to quickly and confidently make sense of the data."

Beyond AI-powered natural language descriptions of data lineage characteristics, Ataccama's platform update includes the following:

• Lineage diagrams that provide a high-level view of data flows with details that users can drill into on demand.
• Secure lineage capabilities that enable metadata extraction from on-premises and restricted environments without forcing users to move sensitive information to the cloud.
• Connections to Google BigQuery and Microsoft Azure Synapse that enable data profiling and data quality workloads to be executed without uploading data to Ataccama, which can help organizations save on data egress expenses.

Perhaps most significant are the secure lineage capabilities, according to Aslett.

"The ability to extract metadata from sensitive data ... is an important capability that will enable users to better understand their data environments while maintaining compliance with data sovereignty regulations and policies," he said.

In addition, Aslett highlighted pushing workloads down into data warehouses.

Looking ahead

With the latest Ataccama update now generally available, the vendor plans to continue to focus on using AI to foster data quality and engender trust in data, according to Smith.

In particular, Ataccama aims to add new agentic AI capabilities that include input from its Model Context Protocol server -- a standard that puts in place protocols for how agents interact with other tools and systems -- to fuel smarter autonomous data interactions, she said.

In addition, Ataccama is looking into adding new capabilities that address unstructured data and reference data management, as well as expanding its data observability capabilities, Smith continued.

"We're exploring new capabilities ... to give customers even deeper insight and control over their data environments," she said.

Petrie, meanwhile, suggested that Ataccama partner with more AI and machine learning providers so that customers can integrate Ataccama's capabilities with AI tools to improve the performance of AI/ML models and applications.

"[Ataccama's] rich lineage metadata can help AI/ML platform vendors govern data alongside AI/ML models and agentic applications, contributing to an end-to-end view of inputs, logic and outputs," he said.

Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.