Every executive has heard the pitch: give employees access to dashboards, self-service analytics, and a few AI copilots, and the organization will magically become “data-driven.” Donna Medeiros, VP of AI and Data Advisory at Data Society, has watched that promise fail to deliver again and again, and her explanation for why is refreshingly direct. Access was never the bottleneck. Fluency is.
Handing someone a dashboard doesn’t make them capable of interpreting it correctly, questioning its assumptions, or knowing when the underlying data can’t support the conclusion they want to draw. A genuine data-driven culture in organizations requires something deeper than tool distribution. It requires skills, the right platforms, governance, and a way to prove the whole system is actually producing better outcomes.
Data Is The Foundation, Not A Side Input
It’s worth starting with a point so foundational that it’s easy to skip past, but Donna states it as the premise everything else rests on.
“Data underpins all of AI. Without data there isn’t AI capabilities.”
Executives investing heavily in AI capabilities sometimes treat data quality and data culture as a secondary workstream, something the data team handles quietly in the background while the “real” AI work happens elsewhere. That ordering is backward. Every AI model, every automated workflow, every agentic tool making recommendations inside your business is only as good as the data feeding it. An organization that hasn’t built genuine data fluency across its workforce is, in effect, building AI capability on a foundation it hasn’t actually secured. This is a big part of why data maturity work, understanding an organization’s own “DataDNA” profile before scaling AI initiatives, has to come before or alongside AI investment rather than after it: https://datasociety.com/in-the-quest-for-data-maturity-an-organizations-datadna-profile-matters/
The New Frontier: Unstructured Data
Here’s where the conversation about data culture gets genuinely current, rather than repeating advice from a decade ago. Most legacy governance practices, literacy programs, and even executive intuition about data were built around structured data: rows, columns, numbers in a spreadsheet. Donna points to a different reality now defining where the real opportunity, and the real challenge, sits.
“Unstructured data is the big frontier now for AI. That’s where so much of the content is, so much of the valuable information is.”
Video, audio, images, free-text customer feedback, call transcripts, internal documents. This is where a massive share of an organization’s actual institutional knowledge lives, and it’s precisely the category of data that traditional data culture initiatives were never designed to handle. If your organization’s idea of “data literacy” still centers entirely on reading a chart or building a pivot table, it’s addressing a shrinking slice of the problem. Building a data-driven culture today means expanding fluency to cover how unstructured content gets captured, tagged, governed, and used responsibly, which is a meaningfully different skill set than traditional business intelligence literacy. This shift also has direct implications for AI governance, since unstructured data introduces new categories of risk, from privacy exposure in call recordings to bias embedded in image or video training sets, that a governance framework built only for structured data won’t catch.
Fluency Is A System, Not A Skill
Perhaps the most useful reframe Donna offers is that fluency isn’t a single skill you train once and check off a list. It’s a system with several interlocking parts, and all of them have to function together.
“Fluency being the skills, using the tools, having the right platforms, having the governance that underpins and has some of the metrics to show that the fluency actually made outcomes, the companies worked.”
Unpack that sentence and you get a genuine checklist for any organization trying to assess where its data culture actually stands. Skills: do employees know how to work with data relevant to their role, not just in the abstract. Tools: do they have access to platforms suited to the kind of data and analysis their job requires. Governance: are there clear, well-understood guardrails for how data gets used and shared. And metrics: can the organization actually point to evidence that fluency translated into better outcomes, not just better attendance at a training session.
That last piece, metrics, is the one most organizations skip, and it’s exactly the same blind spot that shows up in AI training programs that never get benchmarked before rollout. A data culture initiative without measurable outcomes is really just an activity, not a strategy. Executives serious about building a data-driven culture need to define upfront what “working” looks like: faster decision cycles, fewer data-quality escalations, higher self-service adoption among analysts, more accurate forecasting. Whatever the specific measure, it has to exist before the initiative launches, not get invented retroactively to justify the budget that was already spent.
Why Governance And Training Can’t Be Separated Here Either
The same principle that applies to AI governance broadly applies just as directly to data culture specifically.
“Governance also has to be paired with the trainings to get that AI fluency needed.”
A data governance policy that employees don’t understand functions the same way as an AI acceptable-use policy nobody read: it exists on paper but does nothing in practice. If your organization has invested in a data governance framework but hasn’t paired it with active, ongoing training on what that framework means for day-to-day work, you likely have a compliance document, not a culture. Real data-driven cultures treat governance training as a continuous part of onboarding and skill development, not a one-time click-through at the start of employment. Data Society’s work with organizations like Discover Financial Services illustrates how tightly governance, literacy, and measurable outcomes need to be woven together for a data initiative to stick rather than fade after the initial rollout enthusiasm wears off: https://datasociety.com/case-study/discover-financial-services/
The ROI Problem That’s Stalling Data Culture Everywhere
None of this matters if the investment doesn’t produce results the organization can point to, and Donna is blunt about where the research currently stands.
“The research right now is showing organizations just aren’t getting a return on investment.”
That’s a sobering statement for any executive who has already sponsored a data literacy program, a new BI platform rollout, or a governance overhaul. It’s also, in a strange way, encouraging, because it means the problem isn’t unique to your organization, and it isn’t a sign that data-driven culture is an empty buzzword. It’s a sign that most organizations are building the pieces (tools, some training, some governance) without connecting them into the kind of integrated system Donna describes, and without measuring whether the system is actually working. Forbes’ reporting on how data literacy increases workforce productivity backs this up: the productivity gains are real, but only when literacy efforts are structured, sustained, and measured rather than treated as a one-off workshop: https://datasociety.com/forbes-why-data-literacy-increases-productivity-among-employees/
If you’re an executive sponsoring a data culture initiative this year, treat it the same way you’d treat any other capital investment: define the outcome you expect, benchmark where you stand today, build fluency across skills, tools, and governance together, and measure whether it worked. Skip any one of those steps and you’ll end up with exactly what the research describes: activity without return.
Frequently Asked Questions
Building a data-driven culture means embedding data fluency across skills, tools, governance, and measurable outcomes so that employees at every level can responsibly use data to inform decisions. It goes beyond simply granting dashboard access; it requires training people to interpret, question, and act on data correctly.
Access alone doesn’t teach employees how to interpret data correctly, recognize its limitations, or apply it responsibly to a decision. Without fluency, skills, governance, and proof that outcomes improve, dashboard access often produces false confidence rather than better decisions.
Unstructured data includes formats like video, audio, images, and free text, and it now holds a large share of an organization’s valuable information. Most traditional data literacy and governance practices were built for structured, tabular data, so organizations need updated fluency and governance approaches to handle unstructured content responsibly.
Data governance provides the guardrails for how data is used, shared, and protected, but it only works as intended when paired with ongoing training. Governance policies without training tend to exist on paper without changing actual behavior.
Organizations should benchmark relevant metrics, such as decision speed, data-quality issues, or adoption rates, before launching a data culture initiative, then track the same metrics afterward. Without this before-and-after measurement, it’s difficult to prove that fluency-building efforts produced a real return on investment.
Ready To Turn Data Access Into Real Data Fluency?
If your organization has invested in tools and dashboards but still can’t prove your data culture initiative is working, book time with Donna to map a fluency-first strategy for building a data-driven culture that actually shows results: https://meetings.hubspot.com/donna-medeiros/meet-with-data-societys-ai-and-data-advisor
