I learned how to think about decisions in a place most people never have to. Federal prison. Eighteen months of solitary confinement. No data. No dashboards. No AI. Just the consequences of the decisions I had already made and a long, quiet runway to figure out how to make better ones.
That experience is the reason I notice something now that most operators miss.
The companies I work with run between one and fifty million in revenue. They have never had more information available to them in human history. They have dashboards on three monitors. AI agents writing weekly summaries. Real-time KPIs they can pull up on their phones at dinner. By every measurable input, they should be making the sharpest decisions of any generation of operators that has ever existed.
They are not. They are making slower, more anxious, more expensive ones. And almost no one is willing to say out loud why.
The Lie the Industry Sold
For the last fifteen years, the business intelligence industry sold small and mid-sized businesses a story that sounded like progress. More data. More dashboards. More AI. Better decisions. Trust the stack. Let the numbers tell you what to do.
It hasn’t worked. Pricing calls get delayed by quarters. Hiring decisions ping-pong between candidates for weeks. Expansion plans get built on dashboard screenshots and a hunch. The stack got smarter. The decision-making got worse.
I often say it this way. We are swimming in answers, and somewhere along the line we forgot the question.
Dashboards Don’t Decide. AI Doesn’t Decide. People Do.
Here is what no one in this industry wants to say out loud. Nothing in the modern BI stack was actually built to decide. It was built to display, summarize, and recommend. Those are three very different jobs from the one operators actually need done.
Dashboards measure activity. They tell you what already happened. They are a rearview mirror with a great paint job. Useful for context. Not useful for choosing.
AI summarizes inputs. It compresses information and produces output that sounds confident. Ask it whether to expand into a new market and it will give you an answer. Ask it the same question with the opposite framing five minutes later and it will give you the opposite answer, just as confidently. AI is built to respond. It is not built to weigh tradeoffs against what you actually care about.
And here is the part that matters most. Neither dashboards nor AI are values-aligned. They don’t know what your culture is. They don’t know what your exit strategy is. They don’t know what kind of company you are actually trying to build. They produce outputs that look like decisions and feel like decisions, but underneath, they are still the same gut calls you were making before. Just dressed in a nicer suit.
The Patterns That Show Up Everywhere
A few patterns surface in nearly every mid-market business I look at.
The dashboard becomes the meeting. Leadership teams open the BI stack, stare at the numbers, agree the numbers are bad or good, and disperse. Nothing is decided. The dashboard becomes a place to discuss in circles instead of a tool that drives action.
The AI summary becomes the decision. A founder asks ChatGPT what to do about an underperforming general manager and gets a polite, structured five-paragraph response. They forward it to a partner. The partner reads it. Nothing happens. The summary quietly substitutes for the deliberation that should have happened.
Green charts create false confidence. The pricing dashboard is up. The margin line is green. The CEO assumes the pricing question is handled. But there is no model running scenarios, no thresholds defined, no pressure-testing of assumptions. Just a chart that looks healthy right up until the day it doesn’t.
Tools become rationalization engines. This is the most expensive pattern of all. The decision was already made in the founder’s head three weeks ago. The dashboard is doing PR. The AI summary is providing cover. What looks like analysis is actually confirmation dressed in slides.
In every one of those patterns, the operator believes they are using their data to think more clearly. They are doing the opposite. They are using their data to avoid thinking.
The Missing Piece Isn’t More Intelligence. It’s Structure.
When I say structure, I do not mean another tool. I do not mean a process document. I mean a system for making the decision itself. The frameworks, thresholds, and pressure-testing that turn information into action.
A real decision system blends two things almost every BI stack misses. The first is deterministic modeling. The math that says, given these inputs, this is what the outcome looks like. Clean, predictable, repeatable. The second is probabilistic modeling. The math that says, here is the range of outcomes, here are the odds, here is where the downside lives. Dashboards give you neither. AI fakes the second one badly.
On top of that, you layer the part nobody talks about. A culture-informed weighting matrix. What does your team value? What is your exit strategy? What kind of risk are you actually willing to absorb to grow? The same expansion decision is the right call for one company and a catastrophe for another, depending entirely on what they care about. No off-the-shelf tool will ever tell you that. It has to be built for you.
Then you pressure-test. You run the scenarios. You define the thresholds before you ever pull the trigger. You decide in advance what “wrong” looks like, so you can recognize it the moment it shows up. And only then do you give the decision back to the human being whose job it is to make it. With structure underneath. Not gut feel.
That is the thing dashboards and AI cannot do. That is the thing operators actually need.
What Wrong Looks Like
I tell operators something they don’t always want to hear. Most of your most important decisions are not being made. They are being deferred, dressed up, or delegated to a chart. The technology you bought to help you think clearly is doing the exact opposite. It is giving you a beautiful, real-time, AI-enhanced view of your indecision.
The fix is not a better dashboard. It is not a smarter AI agent. It is a structured way to look at the decision in front of you through the lens of your culture, your growth plan, and the actual scenarios of what could go right and what could go wrong. Knowing what wrong looks like before you commit. Pressure-testing your own assumptions before the market does it for you.
That work is not glamorous. It does not fit on a dashboard. It is the work of turning a question into a system. And it is the work that almost no operator is doing.
What Comes Next
Small and mid-sized businesses are not going to solve this problem by buying more software. They are going to solve it by recognizing that intelligence and decision-making are not the same thing. The gap between the two is exactly where most of their money is being lost.
The operators who win the next decade will not be the ones with the prettiest dashboards or the most AI subscriptions. They will be the ones who finally built a way to decide.
Everyone else will keep swimming.
About the Author
Christian Torres is the founder of Stark Analytics and a Decision Intelligence Speaker and Strategist. His work sits at the intersection of business intelligence, AI, scenario modeling, and human judgment under pressure. After nearly a decade in federal prison, including 18 months in solitary confinement, Christian rebuilt his life through systems thinking, Excel, and hard-earned decision discipline.
Today, he helps leadership teams stop confusing dashboards with decisions. Through keynotes, executive workshops, and custom decision systems, he helps organizations pressure-test major calls before they commit time, money, and people.
To bring Christian in for a keynote or facilitated workshop on decision intelligence, AI, dashboards, and better business decisions, connect with him through Apogee Global Speakers Bureau (https://apogeeglobalrms.io/book-top-speakers-at-apogee-executive-speaker-bureau/)
