Monday morning fix: The executives offered simple steps leaders can act on immediately to move their data culture from talk to transformation.

Culture beats dashboards: A PYMNTS panel led by Velera’s Jeremiah Lotz said real impact comes when business leaders, not just data teams, own and trust the numbers.

Watch more: Building the Culture Behind the Data Strategy

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions .

Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

When it comes to data usage and actions, most institutions can point to dashboards and reports. Far fewer can prove that insight actually shapes decisions.

The difference is culture. A strong data culture shows up when leaders discuss outcomes in data terms, when frontline teams feel responsible for data quality, and when decisions can be traced back to trusted information rather than instinct.

That was the common thread in a PYMNTS panel with Jeremiah Lotz, senior vice president of enterprise data and experience design at credit union service organization Velera; John Sahagian, chief data officer at Chicago-based credit union BCU; and Richie Cotton, senior data evangelist at data advocacy and education group DataCamp. All three encouraged banks and credit unions to go beyond data strategy to achieving real impact by developing a positive data culture.

“A healthy culture ties back to data being used in a way that leverages trusted data,” Lotz told PYMNTS. “We’re using the information and decisions to trace back and that we have data to support that. And it’s not one in which we’re just kind of a ‘report theater.’ You need to understand how to use the data for good decision making.”

From there, the panel’s consensus on “healthy” was straightforward: The data strategy should be inseparable from the business strategy, and the people who run the business (not just the data team) need to trust, understand and use the numbers. The red flags are equally plain: Using data to justify a decision already made, ignoring contradictory evidence or punishing messengers for inconvenient truths. None of that fosters the curiosity or accountability that durable data cultures require.

How BCU Made Culture Real

Sahagian described BCU’s reset as a leadership decision followed by a lot of listening — and even more shared work. He offered a candid account of where the credit union started and how it changed course. In 2018, BCU found itself awash in data but no strategy to use it. So, he led an effort to craft an intentional strategy that started with a trusted data culture. It started with conversations.

“We went around the organization and asked, ‘What’s going on in your business? What are your goals?’” Sahagian said. “Then our business leaders recognized that we’re not going to just hire a data team to do all the work. The business has to do the majority of the work. They understand the data. They’re the ones that have to take ownership. All of a sudden you saw business leaders taking responsibility to make sure their data was right. That’s when I knew we were on the right track.”

Lotz agreed that the mindset shift has to be backed by investment and by redefining the role of the data function. Data teams, he said, should act as guides and “dot connectors,” bringing education to all levels, lighting the path, and helping business owners link insights across units so the organization can scale wins beyond one team or channel. That orientation, in his view, is where companies begin to see return on their data spend.

The panel offered a practical health check any executive can run. Start by asking business leaders how they know they’re succeeding and which data they trust. In high-functioning cultures, leaders can answer plainly, explain their problem‑solving approach and articulate how insight travels into action. In weak ones, data is an afterthought, or worse, a prop to defend decisions already made. Lotz added that the strongest performers treat data as an asset tied to a clear business outcome, a pattern Velera sees repeatedly across the more than 4,000 credit unions it serves.

Foundation Before Speed

Sahagian cautioned that quick wins only compound value when a common foundation is in place: documented processes, curated knowledge, and governance that makes data understandable and safe to use. As generative artificial intelligence (AI) lowers the barrier to entry, he said, the risk of more people making faster mistakes rises unless data is well described and controlled.

Lotz’s ambassador model is designed to meet that moment. In it, a company recruits naturally curious operators, equips them with training (including prompt‑writing basics) and shared vocabulary, and lets them carry data habits into everyday work.

Cotton’s contribution: Keep the vocabulary simple and the first skills practical, such as reading line and bar charts and writing prompts that return useful results.

The panelists were pragmatic about building literacy across the enterprise. Cotton argued that democratization starts with a shared vocabulary, so business and technical teams can hold productive conversations. Upskilling, he added, is simply learning on the job; given how fast tools evolve, it never really stops.

Asked where data and AI are moving the needle, Lotz focused first on fraud. With billions lost annually in the sector, he sees “curious minds” inventing new ways to detect and deter schemes, which is necessary because fraud evolves as quickly as defenses. He also pointed to personalization, noting that consumer comfort with tailored experiences has become an expectation.

The opportunity, he said, is to use models to predict needs and change what a member sees in real time (across branch and digital) based on behavior and history.

Finally, he sees AI helping revive growth in products such as wallets and installment payments by tuning experiences to stickier segments.

BCU is pushing on three fronts. First, member service: A revamped chatbot routes roughly 40% of intents to a conversational large language model grounded in the credit union’s documented knowledge base, improving containment and satisfaction.

Second, personalization: A “model of models” approach to next-best-action has produced two-to-three-times higher engagement in early trials compared with rule-driven offers.

Third, risk: A propensity-to-roll model helps collections allocate finite outreach, distinguishing chronic slow payers from members who need immediate help, allowing BCU to manage rising delinquency with static staffing.

Adoption as a Sequence

Beyond fraud, the panel framed adoption as a sequence: map consequences, decide what can be automated with guardrails, and keep humans attached where stakes or uncertainty are high.

That framing also addresses AI hesitancy, according to the panelists. It focuses on small, safe wins, relentless fact-checking and open discussions about what stays human. Sahagian and Lotz added that “grounded” systems fed with high‑quality internal knowledge and monitored by capable people are far easier to trust and scale.

For the “lightning round” at the end of the panel, the participants were asked for three things the audience could do immediately to improve their respective data cultures. Their responses:

Lotz: One, appoint cross-functional “ambassadors” with natural curiosity, connect them and give them air cover to spread good practices. Two, pick one modest use case and ship it end to end so teams can see value “in real life” and gain creative momentum. Three , build a cadence of skilling and upskilling so more decisions are made well and the tech investment pays off.

Sahagian: One, document business knowledge and processess so AI and people have reliable , searchable context. Two, normalize data conversations in business meetings; reinforce that data isn’t the domain of IT but the fuel for every function. Three, make ownership explicit; business leaders should know what data they have, which they trust, and why it matters to their goals.

Cotton: One, use meetings as leverage: track progress to targets with numbers to nudge culture toward evidence. Two, start AI where it eases pain: automate the tasks your teams hate, even if the first wins are narrow. Three, treat “what do I hate at work?” as the prompt for your first automation pilots.

The through line from all three leaders is that culture turns tooling into outcomes. It’s done by aligning on business goals, teaching people how to talk about and use data and delivering visible wins that build trust.

As Lotz put it, the job of the data team is catalytic: “We’re not just here to be data experts. We’re here to be guides. We’re here to be transformers in some way where we can help ultimately the company.”