Starburst launches AIDA assistant for governed data
Starburst has launched an AI data assistant called AIDA, available in the Starburst Enterprise Platform.
AIDA is designed to help users explore and analyse enterprise data in natural language instead of relying on dashboards and spreadsheet exports. The tool works across distributed data sources and is aimed at organisations that want to make decisions using governed data without moving it into a central repository.
Starburst is framing the launch around a common frustration in large companies: business teams can wait months for dashboards, then do further work in spreadsheets while still questioning the quality of the numbers. AIDA is intended to sit between those users and the company's underlying data estate, returning responses that reflect existing controls and definitions.
Reasoning layer
A central part of the launch is what Starburst describes as advanced reasoning. AIDA uses a reason-act-observe framework that combines live data sampling with metadata analysis, with the aim of producing answers based on more than simple query generation or summarisation.
That approach reflects a broader shift in enterprise software, as vendors try to move beyond chat interfaces that translate prompts into database queries to systems that can handle more steps in the analytical process. In practice, that means assessing available data, examining context and returning an answer in a format suited to the user.
AIDA can tailor responses by role. Data practitioners can receive more technical explanations, while business leaders can get shorter summaries intended for decision-making.
Starburst has also added a white-labelling feature so customers can apply their own branding to the assistant inside their organisations. The system supports several large language models within the platform, including models from Anthropic, OpenAI and AWS Bedrock.
Federated approach
The launch also reinforces Starburst's longstanding argument that enterprises should analyse data where it sits rather than move all of it into a single vendor-controlled environment. The platform is built around federated access to data spread across lakes, warehouses, cloud object storage and operational systems.
That stance has become more relevant as companies look for ways to apply generative AI tools to internal data while maintaining governance rules. For many businesses, the challenge is less about choosing a model than controlling how systems access financial, operational and customer information held across different platforms.
"Most companies are still approaching AI the wrong way, focusing on models instead of the data those models depend on," said Justin Borgman, Co-founder and Chief Executive Officer, Starburst. "The real challenge is applying AI to business decisions without moving data or compromising governance. Starburst's AI Data Assistant is built to solve that by providing access to trusted, distributed data from across the enterprise."
Starburst says the platform applies governance, definitions and access controls across multiple data sources to create what it describes as an AI-ready context layer. In that setup, AIDA acts as the user interface, while the wider platform governs how information is accessed and interpreted.
Business uses
Starburst outlined several business cases for the assistant, including identifying billing discrepancies across contracts, usage and invoicing data; spotting signals linked to customer churn; and helping with fraud and compliance investigations by connecting transaction and customer data with broader context.
Those examples point to where software suppliers see demand for AI assistants heading: not just in reporting and search, but also in operational work where staff need to combine information from several systems quickly. Adoption will depend on whether companies trust the quality of the outputs and can apply internal controls to sensitive data.
Industry analysts say governed access to distributed data is becoming a key issue as companies try to widen access to analytics through AI agents.
"As enterprises seek to democratize analytics with agentic AI, they need governed access to distributed datasets," said Kevin Petrie, Vice President of Research, BARC US. "Starburst meets this requirement and goes further to enable intent- and persona-specific reasoning on federated inputs. This helps diverse stakeholders make smarter decisions in the context of the business."