Joe DosSantos, VP of Enterprise Data and Analytics at Workday, speaks with Clyde Gillard, North American AI GTM Leader at HPE, in a video interview about challenges and opportunities in managing data, approach to data access and security, the journey from generative to agentic AI, and bringing agents into analytics. 

Responsible for providing data within the Workday environment, DosSantos aims to modernize the technology stack so that decision-makers in areas like sales, purchasing, and engineering can access timely analytics. His work centers on creating reliable sources of truth, automating processes, and leveraging AI to boost productivity, improve decision-making, and increase profitability.

Data, definitions, and decision-making

DosSantos reflects on the challenges and opportunities in managing data within an organization, particularly in the context of Workday.

For him, Workday serves as the central hub for managing people and money. “We do our finance work there, and we do our HR work. So if there’s a question of recruiting, if there’s a question of job promotions… terminations, or facilities management, that’s where we  do our work.”

However, DosSantos points out that the business of a software company involves many moving parts:

• Engineering and improving the product
• Interpreting market signals and aligning them with the product
• Marketing effectively and building a sales pipeline
• Servicing customers to ensure satisfaction

The challenge, according to DosSantos, is blending all these signals. He explains that understanding customer usage patterns — such as identifying which product features drive repeat engagement — can be critical to enhancing the customer experience. Similarly, packaging products and services in certain ways can lead to better customer satisfaction (CSAT) scores.

But success, DosSantos emphasizes, is not just about collecting information. “The key for us is  not necessarily just amassing information, but being able to put it together in a way that we can correlate it with each other.”

The data sources vary, sometimes Salesforce, Telemetry, or Workday itself, but the goal is to “package it in a way that it can be bound together.”

One of the hardest parts, DosSantos notes, is not technical but linguistic. “The hard part is getting people to agree on what things mean. It’s that things haven’t changed in 20 years.”

He shares the example of defining churn: “First of all, can we have a definition of churn? Well, obviously it’s when people stop buying.” DosSantos walks through a scenario of a renewal that slips past the quarter deadline, posing the question: Is that churn or a delayed renewal?

“These are not technical considerations. These are the words and how we think across the business collectively,” he explains. Without alignment on such definitions, even the most advanced analytics and AI tools lose their value.

For DosSantos, the mission is to bring data together from multiple sources, make data interoperable, and establish consistent definitions across the organization. Only then, he believes, can companies truly “drive the right kind of decisions.”

Building a pharmacy for data access and security

Moving forward, DosSantos draws on his manufacturing background to explain how he approaches data governance and security. “For the first 10 years of my life, it was all about putting products on a shelf,” he says. That experience taught him valuable lessons about policy, clarity, and process that he now applies to data management.

He uses a vivid metaphor comparing data governance to a pharmacy, where products like shampoo, razor blades, and prescription drugs each have clear, predefined access rules.

According to him, this “pharmacy” model reflects how the supply chain works; policies are embedded in the process so mistakes cannot happen. He applies the same thinking to data:

• Some data is the “shampoo” of data — publicly available, low-risk information such as old earnings reports or content already on a public website.
• Other data carries more sensitivity, such as financials, or is subject to legal restrictions like GDPR and PII regulations.
• Data governance requires careful, column-by-column review to define, tag, and secure each element.

Instead of focusing on whether a person should have access to a table, DosSantos advocates for role-based access tied to data classifications. The process, as he describes it, involves creating a precise glossary of terms, tagging data consistently with those terms, and applying appropriate security policies on top.

The outcome is a controlled, policy-driven “pharmacy experience” for data, where different people entering the same store will see different things based on their role, ensuring security and compliance without ambiguity.

The shift from GenAI to intelligent agents

Furthermore, DosSantos highlights how rising expectations are reshaping what customers demand from their tools and how Workday is responding. That shift, he notes, has put pressure on software companies to “up their game.” What once felt innovative is now considered baseline functionality. Customers increasingly expect every piece of software they use to be getting smarter, he adds.

While Workday has integrated GenAI into its product, DosSantos states that the next frontier is agentic AI.

He offers examples:

• An agent sorting through all incoming resumes to identify the best candidates
• Assisting in evaluating promotion criteria
• Accelerating fraud detection in dashboards
• Managing operational processes like call center or help desk interactions

“We’re developing these agents that are built to accelerate the productivity of teams,” DosSantos says. “In some respects, being just like a person. They have roles, they have titles, they have positions in their organization, they have security constraints.”

Bringing agents into analytics

While operational AI agents are becoming more common, DosSantos is focused on applying similar principles to analytics. He envisions a world where querying business performance is as simple as asking Alexa about the weather.

Concluding, he aims to replace some of the manual work done by BI developers, such as gathering requirements and formatting reports, with AI agents that can create insights on demand.

CDO Magazine appreciates Joe DosSantos for sharing his insights with our global community.