Red Hat is the world’s leading provider of enterprise open source software solutions, helping organizations, including 90% of the Fortune 500, run mission-critical workloads across hybrid cloud, automation, Linux and AI. But even in a highly data-driven marketing organization, teams trying to answer performance questions and find insights were slowed by fragmented information across dashboards, documents, spreadsheets and internal tools. To solve this, Red Hat built the Marketing Insights and Navigation Engine (MINE). Powered by an open hybrid cloud foundation and integrating Databricks for data intelligence, MINE is an advanced agentic marketing platform that gives marketers a conversational way to access trusted answers, navigate performance data and move from insight to action faster.
From fragmented data hunting to precision marketing
Red Hat’s marketing team needed to move faster, but the information marketers relied on was distributed across many systems. Campaign performance lived in dashboards. Definitions and process guidance lived in documentation. Pipeline, account engagement, and attribution context often had to be stitched together manually. The challenge was not simply finding data, but finding trusted, contextualized marketing intelligence that teams could interpret consistently and act on quickly.
"Before MINE, decision-making was slow and fragmented. Marketers had to navigate multiple dashboards and tools to find what they needed, then manually combine campaign performance, account engagement and pipeline impact across systems. Without a centralized place to clarify terms like marketing-sourced pipeline, leads and campaign attribution logic, teams had to rely on documentation or reach out to others for answers. "— Rutanshu Desai, Principal Data Scientist, Marketing Analytics, Red Hat
To solve this, Red Hat’s data science team worked directly with marketing stakeholders to identify the highest-friction workflows. After running 20 focus groups, the team identified common pain points around dashboard navigation, metric definitions, campaign analysis, and frequently asked questions.
"We deliberately avoided the AI-for-the-sake-of-AI trap. We started with real pain points from our stakeholders and built around the friction of finding accurate information across siloed locations."— James Noe, Manager of Data Science, Marketing Analytics, Red Hat
Those insights shaped the creation of MINE, Red Hat’s Marketing Insights and Navigation Engine: an internal GenAI-powered platform that gives marketers a conversational way to find trusted answers, navigate dashboards, understand metrics and access the right resources without deep technical expertise. Because MINE was built around real user needs, it became more than an AI experiment. It became a trusted tool designed around how marketers actually work.
To drive adoption, MINE also had to feel trusted, familiar and transparent from day one. The data science team integrated it with internal SSO and VPN, hosted the frontend on Red Hat OpenShift as the core hybrid cloud foundation to deploy and scale the platform frontend. Red Hat OpenShift provided the proven application platform Red Hat’s teams already knew and trusted, while Databricks was integrated to deliver governed, transparent data tables that showed marketers exactly where answers came from.
“We overcame AI skepticism with transparency by demonstrating repeatable, governed results. With Databricks, we could show marketing stakeholders exactly where an answer came from and why it was consistent. We didn’t give them a black box; we gave them an engine they could trust.”— Preston MacDonald, Senior AI Engineer, Marketing Analytics, Red Hat
For Red Hat’s marketing organization, the value became clear in day-to-day decision-making. MINE gives marketers real-time access to pipeline trends, segment performance, top-performing offers and lower-performing engagements, helping teams identify patterns and adjust strategy faster instead of waiting for quarterly reviews or relying on ad hoc requests.
“The beauty of MINE is that it’s real time. I can quickly see where we are with pipeline, how segments are performing, what offers are driving results and where we need to change our investment strategy.”— Leigh Day, Senior Vice President and Chief Marketing Officer, Red Hat
Building MINE on Databricks: a trusted agentic intelligence layer for marketers
Red Hat built MINE using Databricks as part of a multi-layered hybrid cloud environment where data, documents, models, tools, evaluation and serving could come together in one production-ready environment. The architecture aligned with Red Hat’s open source principles, giving the team the flexibility to prioritize Apache 2.0-licensed models, use approved models like IBM Granite for embeddings, and integrate other approved models through External Model Endpoints without slowing development.
“Databricks provided the data intelligence toolkit we needed to support the underlying data infrastructure and foundation for MINE. From Unity Catalog for model versioning, data tables and UDF tools to unstructured data in Volumes, vector stores and online tables, we had the foundation to build, support and store data within the MINE architecture.”— Preston MacDonald, Senior AI Engineer, Marketing Analytics, Red Hat
Unity Catalog serves as the governance layer for MINE, helping the team manage the structured and unstructured assets that power the platform. Red Hat uses Unity Catalog for governed data tables, model versioning and UDFs that act as agentic tools for structured data retrieval. This gave the team a consistent way to expose trusted marketing data to MINE, supporting a model where users only access authorized information and trace outputs back to reliable sources.
For unstructured marketing knowledge, the team uses Databricks Volumes and Mosaic AI Vector Search to retrieve relevant documents, internal web pages and source metadata. When a marketer asks a question, MINE retrieves the right context, generates a grounded response and provides links to source documents or dashboards for validation.
“Unity Catalog is a key foundation of trust for MINE. It provides centralized governance, verifying that when MINE generates an answer, it only draws from data that the specific user is authorized to see. Beyond access control, Unity Catalog enables lineage and versioning, which allows us to trace any insight back to its origin.”— Shraddha Sutar, Data Scientist, Marketing Analytics, Red Hat
Databricks Genie helps marketers explore structured data in natural language, such as finding campaign IDs, programs, event series, offers and account information. For marketers, this means they can start with a broad business question, refine it with filters, ask follow-ups and receive summaries or tables they can use in planning and reporting.
To support production deployment and continuous improvement, Red Hat used Mosaic AI Model Serving for scalable endpoints, along with MLflow, traces, inference tables and LLM-as-a-judge evaluation to monitor agent behavior and improve answer quality over time. With Databricks handling the governed data, GenAI development, model serving and evaluation layer, the lean data science team could focus less on infrastructure and more on the marketer-facing workflows, tools and user experience. That flexible foundation allowed the team to test new capabilities and improve answer quality without having to rebuild the architecture each time.
“Databricks provided the unified governance and data intelligence we needed to turn fragmented data into a trusted, scalable AI system, successfully moving AI from a laboratory experiment into a production-ready reality that our marketers actually trust.”— Shraddha Sutar, Data Scientist, Marketing Analytics, Red Hat
Business impact: faster decisions, more self-service and stronger collaboration
MINE has delivered measurable productivity gains for Red Hat’s marketing organization. Using real query volumes, user engagement and time-in-motion studies, Red Hat calculated that MINE improves time-to-insight by 70%, saving an estimated 34,000 hours annually by reducing the time marketers spend searching across dashboards, documents and internal systems. Applying a conservative junior marketer hourly rate, the team estimated $1.3 million in annual productivity value. Because MINE is used by marketers at multiple levels, including senior leaders making high-stakes decisions, the team believes the full business value is likely even higher.
“A key driver of the 70% improvement was giving users one centralized, conversational interface to access both data insights and documents. This significantly reduced the time spent switching between tools, tabs and multiple resources, helping users find the right information much faster.”— Rutanshu Desai, Principal Data Scientist, Marketing Analytics, Red Hat
For Red Hat’s CMO, the broader impact is organizational agility. MINE gives teams self-service access to performance patterns, helping them continuously identify what is working, spot underperformance and adjust campaigns, nurture programs or investment decisions faster. That has reduced repetitive requests to marketing operations and data science, while giving global teams a shared view of performance.
“MINE levels the playing field. Everyone is looking at the same information, pulled from consistent sources, refreshed in real time.”— Leigh Day, Senior Vice President and Chief Marketing Officer, Red Hat
Just as importantly, MINE has strengthened the partnership between marketing and data science. Using in-app feedback, Jira workflows and ongoing stakeholder reviews, Red Hat’s data science team continuously improves MINE based on real marketer needs. That collaboration has helped the team move from responding to one-off data requests to building a shared marketing intelligence layer that scales knowledge across the organization.
“We’ve moved from being a data-only team to living in the room with our users. We’re not just building for marketers; we’re building with them.”— James Noe, Manager of Data Science, Marketing Analytics, Red Hat
Beyond productivity, MINE is strengthening marketing’s credibility inside the business. By making trusted intelligence accessible to every marketer, Red Hat is reinforcing that modern marketing is measurable, technical and grounded in data. As MINE becomes part of everyday decision-making, marketing teams can move with more speed, precision and confidence.
“Marketing is a science. It is formulaic, technical and drawing from data just as much as other parts of the business. MINE is helping prove that.”— Leigh Day, Senior Vice President and Chief Marketing Officer, Red Hat
What’s next for Red Hat
To expand on the current success of MINE, Red Hat is connecting the platform through Model Context Protocol (MCP) so its governed marketing intelligence can be embedded in other tools and agentic workflows. By championing open standards like MCP, Red Hat is designing a blueprint for how enterprises can scale interconnected AI agents across different platforms without vendor lock-in. Future integrations could include Slack, internal marketing hubs and campaign management interfaces, extending trusted insights into the places marketers already work.
As MINE becomes more connected to execution, Red Hat sees new opportunities to measure downstream business impact, including click rates, revenue influence and campaign performance improvements. For CMO Leigh Day, the goal is not AI for its own sake. It is to improve customer experience, make better investment decisions and help Red Hat’s marketing organization operate with greater precision.
“We want to make sure we are always putting our customers first: finding what our they need, understanding their challenges and serving them the best way possible.”— Leigh Day, Senior Vice President and Chief Marketing Officer, Red Hat
Acknowledging the team behind MINE
MINE was built by Red Hat’s data science and marketing analytics team in close collaboration with marketing stakeholders across the business.
Core contributors: James Noe, Preston MacDonald, Rutanshu Desai, Shraddha Sutar, Greer Homer and Rylee Sweeney