Five years ago, the response was clear:
There is no appetite for this.
Operational reporting still meant 80-slide decks built by hand from disconnected systems. The process worked, but it had reached its limits. We believed there was a better way: one place connecting scattered data, business logic, reporting, planning and action.
So we built it anyway.
Eighty slides became one dashboard. Five dedicated views, all live, all automated. Once people saw it working, the reaction changed:
Wait. You can do that?
What began as an idea nobody was asking for became a genuine differentiator. It helped URTM earn the trust of a Fortune 500 client, and we have since applied the same thinking across leading companies in multiple sectors.
Now dashboards are everywhere. AI has made them faster and easier to build, so internal teams are increasingly creating their own. That should be a good thing.
But somewhere along the way, the dashboard became the deliverable.
And that was never really the point.
A dashboard is only as good as its foundation.
The chart was never the hard part. Everything underneath it was. Four things get in the way, and they are the same four now as in 2021:
- The data does not tie.
- The logic lives in people's heads.
- The skills rarely sit together.
- Nobody owns the whole outcome.
Garbage in. Better-looking garbage out.
The data does not tie
Systems hold different records, different definitions and different timing. One reports 1,434 employees. Another reports 1,438. Four people apart, and neither system thinks it is wrong.
The difference could be a leaver processed in one system and not the other. Contractors counted as headcount in one place. A dual-role employee recorded twice. A new hire starting next month. Each a reasonable rule, applied inconsistently.
You cannot pick one number and move on. Every discrepancy has to be found, explained and resolved before anything built on top of it can be trusted.
Exhibit 01
Thirty files, six owners, one month-end. The state most projects actually start from, then and now.
PwC's 2026 survey of 767 US operations and supply-chain executives found 87% said data-quality issues had affected the value delivered by their technology investments.
The logic lives in people's heads
The people who know what each number means are usually the same people already extracting, reconciling and explaining it. That knowledge is rarely written down, and new hires need months to learn it before they can be trusted with it.
So this is not only a data problem. It is a knowledge and capacity problem, and the two compound: the person who could settle the definitions is the one with no time to do it.
The skills rarely sit together
A working solution needs one view of four things at once:
- The business and the decisions it has to make.
- The data, and how the systems connect.
- The code and the automation.
- The needs of every end user.
Developers may understand the technology without knowing the business. Business teams know what is needed but not how to build it. Designers can improve usability, but only once they understand how each role actually works. Something is only user-friendly if you understand the users, all of them: a CFO, a department head, a project manager and an analyst open the same system with different questions and different tolerance for detail. Design for one and the others export to Excel.
A dashboard should eliminate the Excel export, not become the reason for another one.
Nobody owns the whole outcome
A prototype can be built quickly. A functioning operational solution touches IT, finance, operations, security and governance, and each team owns one piece of it.
Every dependency brings decisions, approvals and competing priorities. None of those controls is unreasonable. Together they create drag, and the initiative slows down because nobody owns the complete outcome.
BCG describes one company that used AI to reduce a task from ten days to one. Customers still waited ten days, because the surrounding process had not changed. The task became faster. The outcome did not.
AI makes the build faster. It does not remove the blockers.
This is where the hype comes from. AI can build the interface, write the queries and stand up a working prototype in days, so it looks as though all four have been answered.
They have not:
- It cannot reconcile conflicting definitions without a business decision.
- It cannot recover organizational knowledge that was never documented.
- It cannot determine what every user needs without context.
- It cannot remove approvals, dependencies or accountability.
- And it can produce a convincing interface long before the foundation is ready.
AI can assist with each of those tasks. It cannot decide how the business should define, govern and use the information. A model given two definitions of headcount may choose one without surfacing the conflict, or explaining why it discarded the other, and Deloitte's 2026 study found only one in five companies had a mature governance model for autonomous agents.
AI does not fix a broken foundation.
It quotes it back to you, fluently.
What Operational Intelligence changes.
The value was never in the visual layer. It is in what has to be true before the visual layer means anything:
- Connected, reconciled data people can trust.
- Automated reporting and workflows that remove the manual work.
- Clear, usable intelligence that helps people act.
This is the distinction between having a dashboard and having Operational Intelligence.
At URTM, we connect data, business logic, reporting, planning and action across the systems a business already uses. The result might be a dashboard, an automated report, an alert, a planning workflow or AI working with governed information. Usually, it is a combination, because the point is the decision at the end, not the screen in the middle.
The real question.
AI can help almost anyone build a dashboard now.
The harder work is making sure the numbers are trusted, the process is connected and the result changes what happens next.
The question is no longer whether you can build a dashboard.
It is whether you are building another place to look, or a better way to run the business.
Sources
PwC, 2026 Digital Trends in Operations Survey. 767 US operations and supply chain executives at companies with revenues of $100M or more. 87% report data-quality issues affecting the value delivered by technology investments. www.pwc.com/us/en/services/consulting/supply-chain-operations/library/digital-trends-operations-survey.html
BCG, "Look Past Productivity to Get Real Value from AI", August 31, 2026, drawing on BCG's 2026 AI Radar survey. Includes the case of a task cut from ten days to one with no change in customer wait time. www.bcg.com/publications/2026/why-ai-pilots-rarely-deliver-value
Deloitte, State of AI in the Enterprise 2026. Survey of 3,235 senior leaders across 24 countries. Governance maturity for autonomous agents. www.deloitte.com/us/en/what-we-do/capabilities/applied-artificial-intelligence/content/state-of-ai-in-the-enterprise.html
Maya Terzieva
Managing Partner, URTM Solutions Inc.
Helping organizations move from manual spreadsheets and disconnected reporting to dynamic, automated, and trusted intelligence.
