Welcome back to Connecting.The.Dots.
If you've been following this series, you know we've explored the Intelligence Triad, the Virtuoso Dynamic Model, Data Hedges, and why data readiness is the cipher key to AI success.
But there's something we haven't addressed yet. Something most organizations systematically avoid because it reveals uncomfortable truths.
Today, we're talking about Control Towers—and why the missing piece in your transformation isn't better design, it's the courage to monitor whether what you designed actually works.
Line of Sight
A few years ago, I was running global procurement operations deployments. Our office was located in the World Trade Center at Amsterdam's Schiphol Airport, and from my desk, I had a clear line of sight to the airport's control tower.
One evening, I found myself staring at that control tower. Watching the coordination. The constant monitoring. The way flights came in, went out, runways shifted based on wind conditions, gates reassigned in real-time.
And then it hit me.
That control tower wasn't just managing what was happening now. It was staying ahead. Continuously monitoring operations. Detecting problems before they cascaded. Adjusting in real-time to reality, not to the plan.
The plan was the schedule. Reality was the weather, the delays, the mechanical issues, the connecting passengers running late. The control tower existed to bridge the gap between what should happen and what actually happens.
That was my aha moment.
After that evening, I built my first project management control tower on Asana. The technology and the business need were a perfect fit. That became my proof of concept—a Project Management Control Tower that enabled parallel projects to run at global scale across multiple regions and teams simultaneously. The result? We achieved 30x more output than before. Not 30% more. Thirty times more.
That success opened my eyes to what becomes possible when you make the invisible visible.
From there, I explored different applications of the control tower concept, each teaching me something new about the gap between design and reality:
Transformation Control Towers became my second evolution. I used the tower for process design, documentation, and core model design, while simultaneously tracking gap remediation across 16 different teams globally. This taught me that the control tower isn't just for execution monitoring—it's equally powerful for design convergence, ensuring that distributed teams are building toward the same target model.
Process Oversight Control Towers became my latest application, focused on operations at scale. For our global Master Data Management operations, I leveraged process data for the first time to track cycle times, SLA compliance, right-first-time ratios, and volumetrics of input/output flows. This control tower revealed patterns we couldn't see before—where process debt was accumulating, where manual interventions were compensating for system gaps, where our design assumptions were breaking down under operational reality.
But here's what most people misunderstand: I'm not obsessed with building a product called "control tower." I'm obsessed with understanding the flow of inputs and outputs, measuring performance, and monitoring efficiency.
Because control towers reveal uncomfortable truths. And in business, we're remarkably good at avoiding those.
The Uncomfortable Dots We Ignore
Think about how we approach transformation initiatives. We design. We deploy. We celebrate go-live. Then we move on to the next project.
But we rarely build the monitoring that would reveal whether what we designed actually works as intended.
Why? Because control towers show us uncomfortable dots. They expose:
- Where our design assumptions were wrong
- Where execution drifts from design over time
- Where the environment changed but our process didn't
- Where manual workarounds emerged because the system couldn't handle reality
- Where people quietly stopped following the process because it didn't work
These uncomfortable dots threaten our narrative of successful delivery. They challenge our confidence. They require us to admit that what we built might not be working.
So we don't build the control tower. Or we build it but don't look at the data. Or we look at the data but only track metrics that confirm success.
The result? Long-term inefficiencies hidden in uncomfortable spaces we don't want to examine.
Process Debt: The Gap Between Design and Reality
In software development, we talk about technical debt—the accumulated cost of shortcuts and quick fixes that make future changes harder and more expensive.
I want to introduce a parallel concept: Process Debt.
Process debt is the gap between how you designed your process and how it actually operates. It's the difference between your theory and your reality. It accumulates when:
- People create manual workarounds
- Systems don't integrate as planned
- Data quality forces exception handling
- Volumes exceed design assumptions
- Business context changes but processes don't
Like technical debt, process debt compounds. Every workaround becomes embedded. Every exception becomes standard practice. Every gap between design and reality becomes normalized.
And just like technical debt, process debt is invisible without deliberate monitoring.
That's where control towers come in.
Process Mining: The Ultimate Control Tower
Remember Edition #2, where we discussed making the invisible visible? Process mining is the ultimate tool for this.
Traditional control towers rely on metrics we define in advance. We decide what to measure, build dashboards, track KPIs. But this approach has a fatal flaw: we can only monitor what we thought to measure.
Process mining works differently. It reconstructs your actual process from the digital footprints left in your systems—every transaction, every click, every status change. Then it shows you reality: not what should happen according to your design, but what actually happens in practice.
Process mining reveals:
The invisible variations - Your process doesn't run one way. It runs 47 different ways depending on circumstances you never anticipated in the design.
The hidden bottlenecks - The constraint isn't where you thought it was. It's three steps earlier, in a handoff you assumed would be instant.
The silent workarounds - People route around your designed process 23% of the time, using email and spreadsheets because the system can't handle their reality.
The process drift - Your process looked like your design six months ago. Today? Not even close.
This is the control tower that shows you every uncomfortable dot. Every deviation. Every inefficiency. Every gap between design and reality.
And here's what matters: you can't optimize what you can't see. You can't fix process debt you don't know exists. You can't achieve the perfect process without continuously monitoring the gap between design and execution.
The AI Trap We're All Walking Into
Now let's connect this to what's happening right now with AI.
Everyone's racing to implement AI. To automate. To reduce costs. To gain efficiency.
But here's the question nobody's asking: Automate what, exactly?
Are we using AI to automate broken processes at AI speed? Are we embedding intelligence into workflows that are fundamentally flawed? Are we applying cutting-edge technology to preserve yesterday's inefficiencies?
I see this pattern everywhere: organizations rushing to use AI to execute processes faster, without first asking whether those processes should exist in their current form at all.
This is the equivalent of automating the horse-drawn carriage instead of inventing the automobile.
The Real Opportunity: AI for the Perfect Process
Here's what I believe the real opportunity is—and it builds directly on the white canvas thinking from Edition #3.
AI should play two critical roles:
First: Design the Perfect Process
When you have a white canvas moment—when you're not reengineering but designing from scratch—you should use AI in the design phase. Not to automate the old process, but to help architect the perfect one.
AI can:
- Analyze thousands of process variations to find optimal paths
- Identify potential failure points before deployment
- Simulate different designs under various conditions
- Suggest connections between systems you hadn't considered
- Model data flows to prevent the Data Hedges we discussed in Edition #2
This is about using intelligence to design for intelligence amplification.
Second: AI Embedded Within the Perfect Process
But here's the part most people miss: AI shouldn't just help design the perfect process. It should be embedded within that process as an active participant—continuously working to keep the process running as close to perfection as possible.
Think about what we discussed regarding Data Hedges. Systems that should talk to each other often don't. Data that should flow freely gets trapped. Connections that should exist remain broken. And here's the reality: every process, no matter how perfectly designed, will statistically deliver an error rate. There will always be edge cases, unexpected variations, system hiccups, and exceptions that no design can eliminate completely.
AI embedded within your process can take multiple forms:
Generative AI can create content, documentation, or communications as the process runs, adapting to context and maintaining quality standards. It can synthesize information across systems, generate insights in real-time, or produce outputs that would otherwise require human intervention.
Predictive AI can anticipate bottlenecks, forecast exceptions, and flag potential issues before they occur, allowing proactive intervention rather than reactive firefighting.
But perhaps most powerfully, Agentic AI can serve as the connective tissue that makes the perfect process actually perfect in practice.
Agentic AI lives in the gaps between systems, running continuously within your process as an autonomous operator. It becomes the connector that:
- Translates between systems that speak different languages
- Enriches data as it flows through the process
- Handles exceptions intelligently without breaking the flow
- Adapts to variations while maintaining process integrity
- Bridges the Data Hedges we can't eliminate entirely
This is fundamentally different from traditional automation. Traditional automation says: "Execute this defined sequence of steps faster."
Agentic AI embedded within the process says: "Maintain this process outcome while intelligently adapting to reality."
And here's the critical insight: Agentic AI amplifies human oversight by handling the routine variations autonomously, allowing human intelligence to focus precisely where it creates the most value—on the imperfections, the anomalies, the edge cases that require judgment, creativity, and strategic thinking.
This is the Intelligence Triad in action: AI manages the predictable complexity, freeing humans to focus on the meaningful exceptions.
The Control Tower for Perfect Processes
Now here's where it all comes together.
You've used AI to design what you believe is the perfect process. You've embedded AI within that process to connect and enable. You've deployed it to production.
But perfection is theoretical until it meets reality.
This is where the control tower becomes non-negotiable. Specifically, this is where process mining becomes your ultimate monitoring tool.
The control tower monitors three critical dimensions:
1. Design vs. Execution
Is the process running as designed? Where is reality deviating from intention? Not to enforce rigid compliance, but to understand whether the design assumptions were correct.
2. Theory vs. Practice
Are the outcomes matching expectations? Is the embedded AI performing as predicted? Where are the gaps between what should happen and what does happen?
3. Present vs. Past
How is the process evolving over time? Where is drift occurring? Is it drift toward better optimization (good) or drift toward workarounds and inefficiency (bad)?
Process mining gives you the data. The control tower gives you the visibility. Together, they reveal the uncomfortable dots that tell you where your perfect process isn't so perfect after all.
And here's the crucial part: those uncomfortable dots aren't failures. They're learning opportunities.
Embracing the Uncomfortable Dots
Remember the Schiphol control tower? It doesn't exist to prove the flight schedule is perfect. It exists to manage the constant gap between schedule and reality.
Weather changes. Mechanical issues emerge. Passenger connections run late. The control tower doesn't ignore these uncomfortable realities—it responds to them.
The same must be true for your business processes.
Your control tower will show you uncomfortable dots:
- The step in your process that takes twice as long as designed
- The integration that fails 15% of the time
- The exception handling that's become the rule
- The workaround that 60% of users have independently invented
- The AI connection that works in theory but hallucinates in practice
These dots are uncomfortable because they challenge our narrative of successful delivery. But they're essential because they show us where process debt is accumulating.
The question isn't whether uncomfortable dots will appear. The question is whether you have the courage to make them visible and the discipline to act on what you see.
Connecting Back to the Frameworks
The control tower isn't a standalone concept—it's what makes all our previous frameworks operational and sustainable.
Think of the Intelligence Triad from Edition #1. The control tower is where Human Intelligence, Process Intelligence, and Artificial Intelligence are continuously orchestrated. It's not a one-time design activity—it's ongoing intelligence synchronization that detects when any of the three elements drift out of harmony.
Remember Data Hedges from Edition #2? Control towers, especially those powered by process mining, detect when natural boundaries are growing into silos. They show you where data should flow but doesn't, where master data breaks down, where the cipher key we discussed in Edition #4 stops working in practice.
And when you embrace the White Canvas Moment from Edition #3—when you design from scratch with AI—you need the control tower to validate that your bold new design actually works. Theory is beautiful. Reality is messy. The tower bridges that gap, ensuring your perfect process stays perfect as it meets the real world.
The Path Forward: Building Your Control Tower
So where does this leave you? Three things to do right now:
1. Map Your Process Debt
Choose one critical process. Ask yourself: How closely does execution match design? Don't rely on assumptions. Actually observe. Talk to the people doing the work. Look at the data. Map the gap between theory and reality.
2. Define Your Uncomfortable Dots
What would you rather not know about this process? What metrics would reveal inconvenient truths? What variations would challenge your design assumptions?
3. Build Monitoring Before Optimization
If you're implementing AI, resist the temptation to deploy first and monitor later. Build the control tower in parallel with the process. Use process mining if you can. Make the uncomfortable dots visible from day one.
Because you can't achieve the perfect process without continuously measuring the gap between design and reality.
The Bottom Line
The Schiphol control tower doesn't make flights run on schedule. Weather, mechanics, and human factors make that impossible.
What it does is make deviations visible immediately, so they can be managed before they cascade into system-wide failures.
Your business process control tower serves the same function.
The perfect process is an ideal we strive toward, not a destination we reach and then walk away from. AI helps us design closer to that ideal and live within it more intelligently. But the control tower—powered by process mining—ensures we continuously monitor our distance from perfection.
Most organizations celebrate delivery and avoid monitoring because they fear what they'll discover.
The organizations that will lead the transformation are those brave enough to make the uncomfortable dots visible—and disciplined enough to act on what they see.
The control tower awaits. The question is: are you ready to see what it reveals?