Qlik has placed its Model Context Protocol server in the AWS and Databricks marketplaces, letting AI assistants and agents draw on governed data, established business metrics, and existing analytics logic instead of raw tables. The move targets one of enterprise AI's most expensive weak spots, where agents that lack business context return answers that sound authoritative and quietly miss.

[For more news, click here]

Ask a corporate AI assistant for last quarter's revenue and it will answer in seconds, with the calm confidence of a tool that never second-guesses itself. Whether the figure is correct is a separate question. If the assistant pulled from raw database tables and skipped the definitions, filters, and calculations a finance team spent years refining, the number can look authoritative and still be wrong. Polished output floating free of trusted context has hardened into one of the central problems facing companies as they move artificial intelligence into daily work.

Qlik, the analytics and data-integration company that says its software is used by 75 percent of the Fortune 500, is making its move to fix this problem. The company has expanded availability of its Model Context Protocol server across Amazon Web Services and Databricks, placing it inside the two marketplaces where a large share of enterprises already build and run their AI. The practical effect is that assistants and agents operating in those environments can reach Qlik's governed data, lineage, business metrics, and analytical logic directly, rather than starting cold with unmanaged data.

Why Business Context is the Hard Part

The Model Context Protocol began in late 2024 as an open-source project from Anthropic, a way to connect AI models to external tools and data without custom wiring for every pairing. By the end of 2025 Anthropic had handed it to a vendor-neutral foundation under the Linux Foundation, and through 2026 it has been adopted across the major AI platforms, with tens of millions of monthly software downloads and more than ten thousand public servers. Analysts often describe it as a universal connector for AI, the equivalent of a standard port that any tool can plug into.

Standardizing the connection, however, does not standardize the meaning of what flows through it. An agent can now reach a database with little effort. Knowing that revenue excludes intercompany transfers, or that a regional sales figure follows a particular fiscal calendar, is a different kind of knowledge, and it usually lives inside an analytics layer rather than the raw data. Qlik is working on exposing its engine, tools, and agents so assistants inherit the business logic instead of guessing at it.

“Customers are not looking for another standalone AI surface,” said Josh Good, VP, Ecosystems, Strategy, and Product Leadership at Qlik. “They want the agent frameworks they are already adopting to work with trusted business data, existing metrics, and the logic already embedded in their analytics and data workflows. By expanding Qlik MCP across AWS and Databricks, we are making it easier for customers to bring Qlik’s transformation and analytics strengths into those environments so AI can work with better context, greater efficiency, and more reliable outputs.”

A Pattern the Whole Industry is Chasing

Enterprises poured money into generative AI across the last two years, yet a large share of pilots never reached production, undone by weak governance, messy data, and unclear returns. A widely cited 2025 study from MIT found that the large majority of enterprise generative AI pilots had produced no measurable financial return. Gartner has forecast that forty percent of enterprise applications will include task-specific AI agents by the end of 2026, up from almost none a year earlier, while separately warning that a comparable share of agentic projects could be scrapped by 2027 for lack of measurable value.

The question now is which vendor can feed those models reliable, governed context. Data and analytics companies are racing to position themselves as that trusted layer, and the marketplace listings matter nearly as much as the underlying technology. On Databricks, customers can now buy Qlik through their existing committed spend, removing a procurement hurdle that often stalls adoption. On AWS, the server extends into AI environments including Amazon Bedrock, where many companies assemble their agents.

Qlik describes three patterns it expects to see. Some teams will use the protocol to build data pipelines faster while others will treat Qlik as a trusted source feeding larger agent workflows that stretch across several systems. And some will simply query Qlik in plain language through an assistant, pulling a governed answer without rebuilding the underlying logic in yet another tool. The common thread is reuse, letting work already done in one place travel to the AI experiences a company is adopting elsewhere.

What it Means for Decision-Makers

For business and technology leaders, reusing governed logic lowers the risk of an agent inventing its own version of a metric, and it trims the duplicated effort of rebuilding analytics inside every new assistant. An open protocol that lets agents read and act on enterprise data which widens the surface that security and compliance teams must watch. Analysts tracking the space have been blunt that the speed of adoption and the quality of deployment are not the same thing, and that controls such as access gates, audit logs, and human approval steps stop being optional the moment agents touch sensitive data.

Qlik’s expansion is an attempt to sit wherever its customers are building, on the premise that the winning position is not a new destination but a trusted layer inside the tools teams already rely on.

Related Articles

ServiceNow and Aramco Digital Partner to Expand AI Across Enterprise Workflows

VAST Data and CrowdStrike Extend Cybersecurity to Enterprise AI Data

Accenture and AWS Deepen Middle East Cloud Push as AI Demands Grow