NYC CIO Executive Dinner hosted by DataBank

Every AI roadmap now runs through the same question: can the organization trust the data underneath it? Most cannot answer yes, because the majority of enterprise data is unstructured, scattered across paper, email, PDFs and legacy systems, and out of reach of the models meant to use it. Matt Charlson shares DataBank’s view of why the leaders who govern that data first will out-execute everyone else over the next decade. He makes the case that AI amplifies whatever it is fed, that trust rather than speed is what stalls executive decisions, and that “build or buy” is the wrong first question. He closes with what standing still actually costs once someone measures it, and what it takes to reverse the leak.

KEY TAKEAWAYS

• AI will not fix your data. It will scale it.

Agents move whatever you feed them faster. Without governed intake, “AI-powered” means bad decisions delivered at machine speed. Data readiness is the precondition for AI value, not a cleanup task to schedule afterward. (85% of failed AI projects cite poor data quality as a root cause, Gartner.)

• Trust, not speed, is the real bottleneck to faster decisions.

Leaders do not stall because the answer is slow. They stall because they do not trust what sits underneath it. (67% of organizations do not fully trust their data for decisions, up from 55% a year earlier, Precisely & Drexel University, 2025.)

• Measure the cost of standing still, then fix the inputs before choosing build or buy.

The cost of inaction never appears as a line item, so it is routinely underestimated. Both paths fail on ungoverned inputs (70% of digital transformations fall short regardless of path, McKinsey), and both are reversible once measured: a state agency cleared a 180,000-report backlog in 90 days and reached 90%+ automated indexing.

Boards and executive teams are asking the same question of every AI initiative: what will it return, and how will we know? In Hyland’s own customer research, proving ROI is the top concern about adding AI, well ahead of governance, integration or change management. This conversation takes a leadership view of that gap. It explores why value has to be defined in business terms before technology is chosen, why ROI alone undersells what AI changes, and why the organizations that connect and govern the content they already have are the ones turning pilots into measurable outcomes. Attendees leave with a practical way to frame a business case that their peers, their finance partners and their board will accept.

KEY TAKEAWAYS

• The hardest part of AI is proving it, not building it.

Technical complexity and change management are not what hold leaders back. Justifying the investment is. (54% of Hyland customers name cost and proving ROI as their biggest concern about adding AI, versus 21% for governance and risk, 20% for integration and 6% for change management.)

• Define value broadly, and define it early, with the business.

A defensible case starts with the outcomes that matter most to the people who own them, quantified against agreed metrics and assumptions, then communicated in executive terms. Look beyond ROI alone to cost and productivity, risk reduction, revenue growth and strategic enablement. Engaging early turns a vague ambition into a validated model in weeks, not quarters.

• Value compounds when AI builds on what you already have.

The strongest results come from connecting and governing existing content in place, without a disruptive migration. That is how organizations have scaled volume without scaling.