Best practices for building trusted data foundations: Flexera webinar recap

---

“Trusted data” is a big topic right now, and it’s easy to understand why. Every AI initiative, every automation project, every dashboard your executives glance at before making a major decision is based on one question: can you really trust the data behind it? That was the main theme of this webinar.

James Dalley (senior director, solutions success at Flexera) and Clayton Starko (director, solutions architecture at Flexera) hosted the session to explain why that trust gap exists and what organizations can do to close it. If you missed it, we have you covered. Here’s a full recap of everything covered. Every stat, every analogy, every practical takeaway from James and Clayton’s talk.

Why technology environments feel more complicated than ever

James opened with a comparison that has become somewhat of a cliché in tech circles, but it still resonates now: the internet took seven years to reach 100 million users. ChatGPT did it in just two months. That disparity tells you something about how fast AI adoption is moving, and how little time organizations have had to adjust their data practices to keep up.

He then walked through five eras of how the tech stack has shifted over the past few decades:

• Desktop era: dominated by companies such as Microsoft, Lenovo and Dell, where IT controlled practically everything
• Data center era: Oracle, SAP and VMware centralized workloads, still inside a fairly controllable environment
• Cloud era: AWS, Microsoft Azure and Google Cloud introduced agility, but also data sprawl, shadow IT and new governance headaches
• SaaS era: tools like Microsoft 365, Salesforce and Workday decentralized ownership even further. Dalley pointed out that this is the point where more than half of all IT spend now sits outside of traditional IT’s control
• AI era: ChatGPT, Claude, Gemini and similar tools have all but removed the barrier to entry. Anyone can generate code, content or analytics with a prompt

Each step added convenience and speed. Each step also chipped away at centralized control over data. And that’s exactly where the risk shows up. When anyone can spin up an AI-generated report in seconds, data accuracy, governance, compliance and cost management get harder to hold onto, not easier.

Four pressure points organizations are experiencing

1. Data silos and fragmentation. Disconnected systems and inconsistent definitions create understanding gaps and slow decisions down
2. Regulatory, risk and compliance pressure. New regulations and cybersecurity demands have turned trusted data into a business requirement rather than a technical nice-to-have
3. The rise of AI, automation and analytics. Modern systems depend on reliable data pipelines to feed machine learning models and automations
4. Cloud, SaaS and hybrid environments. Infrastructure has moved fast, but visibility and control across cloud, SaaS and on-premises assets often haven’t kept pace

When James asked Clayton which of the four has the greatest impact on enterprises, Clayton said there is no universal answer. Still, in his conversations with customers, data silos and fragmentation come up more than anything else. He compared giving an AI model access to ungoverned, inconsistent data to handing a child a pair of scissors and letting them run around the house. The analogy is simple, but the message is clear: bad inputs produce bad outputs, and hallucinations are often the result.

That set up the session’s first poll: did the audience fully trust their data to make the next big decision? Just shy of 93% of respondents said not quite yet. That was perhaps not surprising given what James and Clayton had just discussed. Nonetheless, it served as a valuable wake-up call for anyone who believes their organization is already ahead of the curve.

Trusted data, as James defined it, is the foundational layer everything else sits on. Trusted data builds confidence, technology intelligence turns that confidence into action, and outcome streams deliver the actual business impact. It’s a chain: trust feeds intelligence, and intelligence feeds outcomes.

Clayton then shared a surprising stat: 95% of AI projects fail. James connected that figure to the previous poll, in which 93% of respondents said they did not fully trust their data to make their next big decision. Put together, the two stats tell a fairly clear story. Many organizations are trying to squeeze value out of AI without a solid data foundation underneath it. The ones actually succeeding are almost always the ones that built trustworthy, well-governed data first.

95% of AI projects fail because organizations lack technology intelligence

The five components of technology intelligence aren’t a sequential checklist. They’re ingredients that work together, the way ingredients in a cake do; leave one out and the whole thing falls flat.

• Trusted data foundation creates a single, accurate source of truth across domains like IT asset management (ITAM), FinOps, SaaS and cloud
• Analytics and intelligence shifts organizations from reactive reporting to proactive insight
• Contextual insights and correlation turns complex data into strategic clarity for executives
• Automation and action drives consistency, scalability and measurable performance gains
• Outcome streams the tangible results, measured across speed, scale and value

James highlighted one word from that list that he believes gets overlooked: scalability. It’s often seen as an afterthought when teams are focused on immediate results, yet it’s important for long-term success.

He drew on his experience as a former ServiceNow platform owner running configuration management database (CMDB) programs. Early CMDB projects often tried to pull in as much data from as many sources as possible. The idea was simple: grab the data now, and figure out how to use it later. That approach later gave way to more structured models. One was ServiceNow’s Common Service Data Model (CSDM). Another was its Identification and Reconciliation Engine (IRE). Both of which forced teams to slow down and think harder about what data they actually needed and why.

The CMDB programs that worked shared one trait: they were built around clear business needs. Once a team hit one goal, they rolled that data out to the rest of the company, then moved on to the next goal. The question stopped being “when will the CMDB be done?” (it’s never “done”) and became “what’s the next use case?”

That use-case-driven approach is also what ties this work back to what CIOs and CEOs actually care about, which was James’s segue into the next slide, mapping each piece of the “recipe” (trusted data foundation, analytics and intelligence layer, contextual insights and correlation, automation and action, outcome streams) to C-suite priorities.

For CIOs:

• Trusted data foundation drives confidence in AI and analytics, and supports operational integrity and risk management
• The analytics and intelligence layer optimizes spend, performance and asset lifecycle, and informs data-driven IT strategy
• Contextual insights and correlation connect technology KPIs to business value and support transformation and stakeholder alignment
• Automation and action increase efficiency and resilience, and reduce technical debt and manual intervention
• Outcome streams demonstrate IT’s strategic contribution through time-to-value, risk reduction and innovation velocity

For CEOs:

• Trusted data foundation builds credibility in digital initiatives, and strengthens investor and customer trust
• The analytics and intelligence layer enables faster, more confident decisions on growth, acquisitions and strategic pivots
• Contextual insights and correlation clarify how technology investments directly impact revenue, margin and agility
• Automation and action deliver measurable productivity gains and cost savings that improve shareholder value
• Outcome streams realize enterprise agility and business outcomes aligned to growth and sustainability

Where trusted data actually comes from (and what bad data actually costs you)

James took a step back to see what the market thinks about data quality before diving into governance.

Why has trusted data become such a priority now?

The answer is not simply that organizations have more data than before. Poor data quality is increasingly viewed by industry analysts as both an economic and strategic risk. As more and more enterprises invest in AI, cloud modernization and digital transformation, the quality of the data underneath those programs has a direct bearing on whether they succeed.

To demo this transition, James highlighted research from numerous industry analysts and advisory firms, all of which come to the same conclusion. Organizations can’t treat data quality just as an operational issue anymore. It’s now a business issue.

The economic cost of poor data

Poor-quality data has measurable financial implications.

As mentioned in the research presented during the webinar:

Research	Finding
Forbes	Poor data costs the U.S. economy an estimated $3 trillion every year.
Gartner	Organizations lose an average of $12.9 million annually because of poor data quality.
Forrester	More than 25% of organizations report annual losses exceeding $5 million because of data quality issues.
McKinsey & Company	Organizations can lose 15% to 25% of annual revenue due to poor-quality data.

As organizations lean harder into AI, automate more processes and depend more on analytics, the strategic fallout from poor data gets more serious, not less.

The additional figures James referenced:

• Forbes reports that 84% of digital transformation initiatives fail because of poor data
• Gartner predicts that 30% of generative AI projects will be abandoned due to poor data quality and inadequate governance
• Forrester estimates that 60% to 73% of enterprise data is never used for strategic decision-making, despite the investment organizations make in collecting it
• McKinsey & Company notes that successful analytics transformations depend on trusted data foundations rather than analytics tools alone

Taken together, these findings reveal a common pattern: organizations rarely fail because they lack technology, they fail because the information feeding that technology isn’t complete, connected or trusted.

Catching a data error early is inexpensive, but fixing it after the fact costs about ten times more. Cleaning up the downstream damage caused by a wrong decision based on bad data costs about ten times more. Prevention beats cleanup by a significant margin.

Given all that, James and Clayton pulled three recommendations straight from what industry analysts are telling organizations to do:

• Make data health measurable. You can’t improve what you don’t benchmark
• Invest in governance. Skip it, and whatever quality you start with tends to erode, not improve
• Act now. Prevention costs a fraction of recovery, in both money and time

Clayton then summarized the philosophy underlying it all. Governance builds trust. Fusion creates context. Democratization improves value. Trusted data is not only accurate. It is comprehensive, contextual, and consistent across the whole enterprise. That is what enables executives to act with confidence since it is complete, connected and trusted.

Establishing data governance: the five-stage data maturity curve

This was one of the more useful, hands-on parts of the session. Clayton walked through a five-stage data maturity curve. He made one thing clear from the start. This is not applicable to your entire data estate at once. You can, and probably will, sit at different maturity stages for different data sets and use cases at the same time.

Stage 1—Ad hoc. Focus: awareness. Data is siloed, inconsistent and incomplete. There’s no defined ownership. Decisions run on instinct rather than insight. Clayton noted plenty of organizations are still stuck here for at least some of their data, and there’s no shame in that.

Stage 2—Managed. Focus: structure. Basic governance and ownership start to take shape. Initial data quality efforts kick off. Reporting becomes repeatable, even if it’s still mostly manual and slow going.

Stage 3—Governed. Focus: trust. This is the tipping point, and Clayton called it the most important stage for building real trust in the data. A formal governance framework and data stewardship get established. Metadata and lineage visibility improve. Data fusion begins here, connecting and enriching data across systems until it’s solid enough to support operational reporting.

Stage 4—Optimized. Focus: enablement. Data fusion matures, with automation and enrichment layered on top. Data democratization starts to emerge as governed self-service access. Data literacy begins spreading as a broader organizational trait rather than a specialist skill, and predictive analytics and orchestration start to scale. Clayton flagged an interesting debate here among industry practitioners. Does fusion need to happen before democratization, or the other way around? In practice, the two often run in parallel. What matters more than the order is that trust gets built through fusion around the same time the data gets opened up for people to actually use.

Stage 5—Transformational. Focus: transformation. Data democratization is fully realized, meaning data is accessible and actionable across the whole enterprise. AI, automation and orchestration get embedded directly into workflows, and agentic AI starts to take real shape. Data monetization and innovation drive new value streams, and the organization increasingly operates as a genuinely data-driven enterprise.

Stages three and four (governed and optimized) are the critical hinge points in the whole curve. Fusion and democratization don’t always happen in a strict order; sometimes fusion enables democratization directly, sometimes they run side by side. What matters is that trust gets built through fusion before, or alongside, the moment the consumption layer gets switched on.

James added a second call to action here. Evaluate your own organization honestly against these five stages. There’s no shame in sitting at stage 1 for some data sets. That’s the purpose of a baseline. You record where you are and then devise a strategy to move forward. As James put it, there is no cheat code that takes you straight from ad hoc to optimized. It is a systematic climb, not a shortcut.

Establishing data governance the right way

Clayton was cautious to reframe what governance actually means. It’s easy to hear “governance” and picture a compliance team slowing everything down. His point is quite the opposite. Governance done right is an enabler, not a constraint. It builds trust. It establishes the access that enables transformation in the first place.

He grounded this in the classic framework: people, process, technology. He did, however, offer one vital detail. Data itself is the fourth leg of that stool. All four have to line up together.

People. Define data owners, stewards and executive sponsors. Build governance into roles, key performance indicators (KPIs) and outcomes. Build a culture where data literacy and shared responsibility are the norm, not an afterthought.

Process. Establish a governance framework. Two industry standards came up by name: DAMA, short for the Data Management Association’s body of knowledge for data management practices (formally known as DAMA-DMBOK) and DCAM, the Data Management Capability Assessment Model published by the EDM Council. Both give organizations a structured way to define standards for quality, classification, lineage and access, and to operationalize the data lifecycle through actual workflows instead of one-off cleanup projects.

Platform. Lean on platforms built for visibility and control, which is where Flexera and ServiceNow come in together. Automate lineage tracking, quality checks and enrichment, and integrate fusion and democratization directly into the systems people already use day to day.

Clayton drew on his CMDB experience again here to explain the difference between ownership and stewardship. Early on, his configuration management team owned the CMDB data outright. As the program matured, he learned something important. It’s preferable to transition that team from owner to steward. Ownership moved to the people actually closest to the data (the principal class owners, in ServiceNow terms), while the central team focused on forms, reporting quality and serving the people consuming the data. Trying to own everything centrally leads to drowning in it. Set clear processes and standards instead. Let ownership be closer to where the data actually is.

When people, process, platform and data all line up, organizations are in a position to layer in agentic AI and orchestration on solid ground, rather than bolting AI onto a shaky foundation and hoping for the best.

The power of data fusion

Data fusion got its own section. Clayton described it as the shift from simply managing information to actively monetizing intelligence. Organizations that reach real data maturity aren’t just keeping their data tidy; they’re using it to create value.

He described a parallel maturity curve specifically for fusion: fragmented, standardized, enriched, intelligent and autonomous, moving left to right as an organization matures. Trusted data fuels technology intelligence. Technology intelligence powers business outcomes. And outcome streams, in the end, are what define digital leadership.

James offered a really helpful perspective here. Many organizations still treat “data” as a technical issue that belongs to IT. But data behaves more like a living system. It not only drives business decisions but also helps explain and support them. His suggestion was to change the vocabulary internally, especially at the C-suite or board level, where the word “data” can undersell what’s actually being discussed. Terms like intelligence, automation or transformation tend to land better with leadership because they connect more directly to business value. Clayton tied this back to the five-stage curve. At stage 1, it’s just data. By stage 5, it’s real intelligence, because that’s what’s guiding the business forward at that point.

Data democratization in action

Democratization is where governance and trust meet action. Once you’ve got a trusted, unified foundation in place, the goal shifts to making that intelligence available to the teams who actually need it: finance, HR, upstream and downstream partners and, yes, IT too. Flexera enables this through governed self-service access, typically through application programming interfaces (APIs) and direct integrations with systems like ServiceNow.

Clayton returned to the AI stance here. AI is only as good as the data supporting it. Without trusted, well-governed data, insights become unreliable and transformation efforts stall. Building that foundation isn’t purely a technology exercise. People set the vision and own the accountability, processes define how data gets managed and governed, and technology provides the scale to make it all work. This is where Flexera One comes in: a centralized platform that unifies data across hybrid IT environments (on-premises, cloud and SaaS) into a single source of truth organizations can actually trust.

That framing led into the session’s second poll: how much effort do you need to put in before a data output becomes actionable? 62% said medium effort (some cleaning or tweaking needed), 38% said high effort (significant work required for reliable insight), and notably, zero attendees said they could trust their data immediately. That’s consistent with the 93% distrust figure from the first poll and reinforces the whole thesis of the session.

How to connect trusted data to measurable business value

Trusted data. A complete inventory with ownership and lineage, normalization and quality checks for accuracy and timeliness, and security and privacy policies applied. The result: a source of truth executives can actually rely on.

Technology intelligence. Fusion across ITAM, CMDB, SaaS, cloud and FinOps data, patterns and benchmarks that enrich decisions and reusable definitions and metrics that keep teams aligned. The result: insight that connects spend, risk and utilization.

Business outcomes. Cost optimization and vendor leverage, risk mitigation and audit readiness, and faster delivery through automation and AI confidence. The result: value that’s actually realized and measurable, not theoretical.

Picture data as the engine, and your business goals as the destination. The point is to connect the two, so trusted data actually turns into the kind of insight that gets you where you’re trying to go.

Where Flexera One fits in

James wrapped up the core content by showing how everything discussed throughout the webinar connects to Flexera One. Instead of just walking through a feature list, he framed Flexera One as a platform for building technology intelligence out of trusted enterprise data.

At the product level, the platform brings together IT visibility and IT asset management (ITAM) under IT, SaaS management, cloud license management and cloud commitment management under cloud, cloud cost optimization under FinOps, and workload optimization under DevOps. These capabilities share a common data foundation instead of operating as isolated point solutions.

Underneath all of that sits a unified technology intelligence layer: a unified data model, APIs and connectors, discovery and an analytics and reporting engine that handles inventory, normalization, visualization and CMDB enrichment. AI-powered capabilities, analytics and reporting are then built on top of the trusted data foundation to generate insights across hybrid IT environments. Instead of treating inventory, CMDB data, cloud resources, and SaaS applications as separate datasets, Flexera One connects them into a single view of enterprise technology.

James kept this portion of the presentation relatively brief, but the broader message reflected the theme of the entire session. Organizations don’t need another isolated management tool. They need a trusted, unified data foundation that connects technology, financial and operational data, so they can make better decisions, support AI initiatives and actually improve business outcomes.

What’s next for Flexera customers

Before wrapping up, James spent a few minutes on something specific to current Flexera customers: Flexera’s user groups. These sessions are peer-led, meaning customers drive the conversation for other customers. A governance council made up of customers chooses the topics, so the agenda stays genuinely relevant instead of turning into a sales pitch.

James also announced something new: the Flexera and ServiceNow Collective Council, a space for peers to trade insights, share lessons learned and swap practical tips on what’s actually working in their organizations.