Something interesting is happening in AI. A new bottleneck is emerging. It’s not the models or the hardware. It’s the context, and that discrepancy is showing up in production success for AI projects.
Context includes all of the information that AI needs to return an accurate, governed answer. Solving the business context problem, which has become the most important bottleneck in AI, means solving for governed access to data in addition to providing grounded business semantics. In one way or another, this single factor determines all of the other factors that make AI useful in enterprise settings, including the accuracy and applicability of results in enterprise production environments.
Data, data everywhere…
But accessing data is not always easy. Data sources are often spread out, and the data held in the–the context held in them–is diffuse. The impact of legacy data sources makes access even more important because it intensifies this problem. Data, and the business context that surrounds it, isn’t held in one place. It’s spread across the entire business. It exists in cloud data lakes, data warehouses, databases, SaaS systems, and on-premises deployments too. To make things worse, some context exists in unknown locations, held in documents that haven’t even been charted or mapped yet.
Because of this, the stakes couldn’t be higher. Business context is an absolute need, not an option. Without it, hallucination becomes a worsening problem because your AI is trained only on generic sample data, not your data, which is a foundational piece of your business context. So even when accessing data is critical for AI success in production, achieving that access isn’t easy.
Starburst and Qlik target the AI data access problem for business context
Accessing data wherever it lives has been Starburst’s mission from day one, and recently, we took another step in that direction through our partnership with Qlik, expanding our mission even further and improving access to business context.
The approach happens in two parts.
Starburst provides a universal context layer
Starburst is built to access data everywhere, including the contextual data needed by AI. So whether that data is held in an Iceberg data lakehouse, a data lake, a cloud data warehouse, or a legacy on-premises system, Starburst acts as a universal context layer capable of accessing their entire data foundation.
From there, our platform features like Starburst’s AIDA help pull value from that data easily using a conversational, natural language interface, and our MCP server allows data from across your entire data estate to be used easily. At the same time, the choice over how much or how little data you move is always yours. If it makes sense to move some data for certain workloads, you can. Meanwhile, you are always free to use data federation to access data wherever it lives.
Qlik provides transformational and CDC support
At the same time, Qlik helps by providing integration, replication, transformation, and operational support. There are times when data does need to move. A standard example of that is for change data capture purposes (CDC). When data needs to move, Qlik manages that movement through built-in CDC and data integration and transformation capabilities. In addition to that, Qlik’s robust set of visualization capabilities provides a path for BI users migrating to agentic AI workflows.
Qlik and Starburst work together by providing robust support for access and your data pipeline, while maintaining flexibility and trust.
Where Starburst and Qlik will support you in production
What are some concrete examples of this approach in practice? Let’s look at two key use cases.
Accelerating SQL by combining visual transformation with distributed compute engine performance
Writing and maintaining SQL has traditionally occupied the majority of data engineers’ time. Using Starburst and Qlik together, along with AI assistance, you can vastly improve that workflow and achieve efficiency gains by decoupling pipeline design from query execution mechanics.
In practice, this might see Qlik operating as a visual transformation layer, automatically generating standardized, Trino-compatible SQL from graphical data workflows so engineers do not have to write code from scratch. Starburst then acts as the distributed compute engine, parsing the generated SQL and applying advanced optimization techniques like query pushdown to execute the workloads directly at the data sources.
This combined architecture shifts the data engineering paradigm from manual query writing to automated workflow orchestration, enabling teams to scale their data operations without accruing technical debt.
Improving AI accuracy by providing universal access to context
Without proper access to business context, deploying AI-generated SQL into production environments introduces operational risk because standard language models lack the structural awareness to verify schemas or respect data access policies. Without architectural guardrails, autonomous agents frequently hallucinate queries, forcing engineering teams to maintain manual review gates or stall production AI deployments entirely.
The joint framework of Starburst and Qlik addresses this challenge by powering hybrid data for Qlik applications from Qlik Sense and Qlik Cloud Analytics. This integration allows users within Qlik Sense to leverage hybrid data environments seamlessly, safely exposing the entire distributed data footprint to conversational and autonomous workflows. Before generating any analytical queries, AI agents leverage the Starburst catalog for automated schema discovery, ensuring that all downstream requests are structurally grounded in actual enterprise data assets rather than generic assumptions.
Starburst and Qlik work together to accelerate enterprise AI
Together, Starburst and Qlik are working to create enterprise AI solutions built for the enterprises that use them. This means using Starburst to access the data foundation and context layer needed to make generic models operate on your datasets for the benefit of your business.
The objective of this partnership is to expand the complementary capabilities of Starburst and Qlik’s architectures, creating a model of enterprise AI success that is replicable and achieves the results that enterprise customers require.
Want to know more about Starburst? Start here.
