Matt Waxman, Chief Product Officer At Precisely, On How AI-Driven Innovation Is Transforming Enterprise Data Management

“For most of my career, product development capacity was the bottleneck. Coding was a specialized skill; there was always more to build than the process allowed, and everything around it, from roadmaps to prioritization, was organized around that limit.

“AI is removing the limit. When building gets faster, everything upstream and downstream must keep pace, and that is where most of the change is landing for a CPO. Answers that once took a week of pulling data now take a prompt, and agents handle much of the gathering and analysis. Product teams spend less time assembling the picture and more time deciding what it means.

“At Precisely, we're moving toward agents that draft specifications, kick off prototypes, and support prioritization, with a human in the loop reviewing all of it. We're still building this.
“The judgment stays with people. The connective tissue is what's moving to agents.”

“Once someone has built their own dashboard with a prompt, they don't go back to filing a ticket. I see this in customer conversations and inside my own company.

“The tension is that this flies in the face of governed, clean data, which has always been delivered centrally with heavy process. Employees have seen the new way, but the control model hasn't caught up. What I'm hearing about is a shadow data layer: different AI tools pulling from different datasets, producing competing numbers for the same KPI, with no clear trail back to the source. That is a problem with the model, not the people. They are doing what the tools invite them to do.

“These two worlds will have to be bridged. The question is how governance and lineage move into the access layer itself, so speed and control stop being a trade-off. Many companies are trying to fly the plane before they've built the engine, and I don't think anyone has this fully worked out.”

“A prompt is only as good as the context you give it. A skill, meaning a prompt or set of instructions packaged so others can reuse it, is only as useful as the information it can reach. When the data fails, the AI fails, whether the cause is stale records, inconsistent formatting, or missing fields.

“Agents raise the stakes. They are triggered by data and take actions based on what comes back, so a wrong input doesn't just produce a wrong answer. It produces a wrong action, repeated at speed. That means operational disruption, financial loss, and compliance exposure, and in healthcare, an agent working from outdated data could even affect patient care. The institution might not know until later.

“This is why I keep putting data ahead of the AI conversation. Break down silos, fix quality, build governance in, and enrich the data with the context AI needs. Organizations with a formal data strategy report 70 percent trust levels in their data, compared with just 50 percent for those without one.

“The fix is not glamorous. It is the part that makes everything else work.”

“AI is changing the economics of that balance. Product leaders can now continue investing in established products at a much lower cost than was historically possible. Work such as reducing technical debt or refreshing a user interface can be completed far more efficiently, freeing capacity for emerging technologies and entirely new opportunities.

“That means organizations do not necessarily need to choose between maintaining the core business and innovating. They can do both within existing budgets by working more cost-effectively. I do not think there is a settled answer on the right investment split, and I would be wary of anyone who claims one. The priority is to use the efficiency AI creates without compromising the quality and reliability of the existing portfolio.”

“Data is only valuable if it can be relied on, which is why protection and governance can't sit apart from AI anymore. They need to be designed in from the start, in the architecture and the operating process, not added after launch.

“Agents act continuously and increasingly rely on personal data to do it. That raises the bar for accountability and transparency under GDPR, the DUAA, and the EU AI Act. I see responsible data use as part of what customers are buying, not only a compliance line item.”

“The most useful customer conversations start with a problem, not a product. What is different now is how quickly those conversations can move from discovery to something tangible. Product teams can generate working prototypes in minutes, allowing them to explore a problem with a customer and demonstrate a potential solution almost simultaneously.

“That shortens the feedback loop and gives customers a more direct role in shaping what gets built. Expectations are rising as quickly as the hype, however, and a single poor result can undermine confidence. Customer engagement helps product leaders distinguish between capabilities that address a meaningful need and those that are simply fast to announce.”