For more than two decades, I have worked on improving business processes.
I have mapped workflows, analyzed root causes, eliminated waste, reduced variation, and helped organizations implement new ways of working. Throughout these years, Lean Six Sigma and continuous improvement methodologies have provided a solid foundation for solving business problems.
But recently, I have started asking myself a different question:
What if artificial intelligence changes not only how we solve problems, but also which problems we are capable of solving?
Think about the improvement projects running in your organization today.
How much time do your teams spend collecting data, preparing reports, searching for information, and discussing potential solutions? And how much time is actually spent testing improvements and creating value?
Now imagine reducing some of that analytical and administrative effort significantly.
Would we simply complete more improvement projects? Or could we start addressing problems that previously seemed too complex, time-consuming, or expensive to solve?
This question has become increasingly important to me, partly because of technological developments, but also because of an unexpected personal journey in 2026.
When I Became Part of the Process
At the end of February 2026, my employment came to an end, and I began searching for a new opportunity in Denmark.
Over the following five months, I submitted more than 400 job applications and participated in over 30 interviews.
Having spent years improving business processes, I suddenly found myself navigating one over which I had very little control.
Applications disappeared into recruitment systems. Feedback was sometimes limited or nonexistent. Decisions took weeks, requirements changed, and the next step was often unclear.
I understand that recruitment is complex, and I am certainly not suggesting that every unsuccessful application represents a process failure. Nevertheless, experiencing recruitment from the candidate’s perspective gave me a different appreciation of transparency, communication, and uncertainty.
There was another challenge.
As a Turkish citizen living in Denmark, my employment situation was closely connected to my residence and work permit. Finding a new job was not simply a professional decision; it also had implications for my family’s future.
Suddenly, process efficiency was no longer just a professional interest. It was personal.
I frequently caught myself thinking like a process improvement professional: Where are the bottlenecks? Why is information not flowing? What creates unnecessary uncertainty?
Of course, understanding a process and having the authority to improve it are two different things.
But this period also gave me something valuable: time and motivation to explore artificial intelligence more seriously.
From Process Improvement to an AI Startup
My initial focus was pharmaceutical quality management, particularly deviation investigations.
Anyone familiar with pharmaceutical manufacturing knows how demanding these investigations can be. They involve reviewing historical records, identifying similar events, evaluating root causes, assessing potential impact, and preparing documentation that meets regulatory requirements.
These activities are essential, but investigators often spend considerable time gathering, organizing, and comparing information before they can focus on the actual problem.
I started asking myself:
What if AI could help quality professionals spend less time searching for information and more time making informed decisions?
That question gradually developed into Causix, an AI-based business concept designed to support pharmaceutical deviation investigations.
The objective was not to replace professional judgment or automate regulatory decisions. It was to support investigators with relevant information, historical cases, and more structured analysis.
I also began developing an initial software prototype using AI-assisted development tools.
This was perhaps one of the most surprising parts of the experience.
Previously, translating a business improvement idea into a software application usually meant preparing requirements, involving developers, discussing technical possibilities, and waiting for an initial solution.
Now, I could explore ideas, create prototypes, test assumptions, and refine solutions much more directly.
Of course, building a prototype is not the same as delivering secure, validated, production-ready software. But the speed at which I could move from an idea to something tangible was remarkable.
I eventually applied to the Startup Denmark programme.
My first application was rejected.
I reviewed the feedback, challenged some of my assumptions, strengthened the proposal, and tried again.
In August 2026, the Startup Denmark expert panel approved my revised business plan.
It was an important milestone and encouraging external validation of the concept.
But around the same time, another opportunity emerged.
A Change in Direction, Not in Ambition
I returned to corporate employment, joining Coloplast in a project management role.
After months of uncertainty, returning to work was a welcome development. It also gave me the opportunity to learn a new organization, build relationships, and explore different business challenges.
However, my employment commitments and immigration circumstances made pursuing an independent startup more complicated than I had originally anticipated.
I needed to reconsider my priorities.
Initially, I saw two different paths: building my own AI company or continuing my corporate career.
Then I realized that perhaps I was asking the wrong question.
My startup concept focused on one specific process: pharmaceutical deviation handling.
But the principles behind it — understanding a business problem, combining domain knowledge with data, and using AI to support better decisions — were not limited to pharmaceutical quality management.
They could potentially be applied to many other business processes.
And one area particularly caught my attention.
Supply chain management.
The Supply Chain Opportunity: Optimizing the Whole, Not the Parts
Global supply chains are fascinating systems.
Demand forecasting, inventory, production capacity, batch sizes, changeover times, supplier constraints, transportation, and customer service levels continuously influence one another.
One of their biggest challenges is that improving one performance indicator does not necessarily improve the overall business.
Consider a hypothetical manufacturing company supplying multiple markets from several production facilities.
The production team wants to maximize equipment utilization and minimize changeovers. Larger production batches often help achieve these objectives.
The supply chain team wants to reduce inventory, improve responsiveness, and release working capital.
Smaller batches and more frequent production may help achieve those objectives.
Both teams have reasonable goals.
But optimizing them separately can produce a suboptimal result for the organization.
Imagine a product manufactured every eight weeks because changing the production line requires significant downtime.
Producing more frequently could reduce inventory but increase changeover losses. Producing less frequently could improve equipment efficiency but increase inventory, working capital requirements, and exposure to demand fluctuations.
Now add several manufacturing sites, different market requirements, varying lead times, and hundreds of products.
The number of possible alternatives grows rapidly.
How do we determine the best overall solution?
Traditional process improvement methodologies, analytical models, and specialized planning systems remain essential.
But I see an opportunity to combine them with AI, making sophisticated analysis more accessible and allowing teams to explore alternatives more efficiently.
For example, imagine a supply chain professional asking:
- What are the main drivers of our inventory levels?
- How much inventory could we potentially reduce by improving changeover times?
- What would happen if we redistributed production between manufacturing sites?
- How would a different batch-size policy affect capacity, inventory, and customer service?
- Which improvement opportunities would create the greatest overall business value?
These questions require more than a large language model.
They require reliable data, mathematical models, operational constraints, and people who understand the business.
But AI could help connect these elements and make the interaction between decision-makers and analytical systems significantly easier.
This is where I believe the opportunity becomes particularly interesting.
Diagnose, Simulate, Optimize: A Practical Approach
Rather than introducing AI simply to automate individual tasks, I believe organizations should explore how it can strengthen three connected capabilities.
1. Diagnose — Understand What Is Really Happening
Most organizations already have enormous amounts of operational data.
The challenge is turning that data into meaningful insights.
Why is inventory increasing? Which products or markets are driving the problem? Are the main contributors forecast accuracy, manufacturing constraints, batch sizes, or planning policies?
AI-assisted analysis could help identify patterns, connect information from different sources, and guide deeper investigation.
The objective is not to produce more dashboards.
It is to understand the underlying causes of business performance.
2. Simulate — Explore What Could Happen
Once we understand the problem, we need to evaluate alternatives.
What happens if we reduce changeover time by 30%? What if we change production frequency? What if demand increases in one market while declining in another?
Simulation allows us to explore these scenarios before making expensive operational changes.
AI could make this capability more accessible by helping users formulate scenarios, interact with simulation tools, and interpret results.
Instead of debating assumptions for weeks, teams could test them more systematically.
3. Optimize — Identify the Best Overall Decision
Understanding a problem and evaluating scenarios are important steps.
But ultimately, organizations must decide what to do.
Optimization techniques can help compare alternatives while considering multiple business objectives and constraints.
For instance, reducing inventory should not come at the expense of unacceptable service levels, quality risks, or operational instability.
The goal should be to optimize overall business performance rather than individual departmental KPIs.
AI could help decision-makers interact with these optimization models, understand trade-offs, and explain the implications of different choices.
Diagnose. Simulate. Optimize.
This is not a replacement for Lean Six Sigma, operations research, or supply chain expertise.
It is a way of bringing these disciplines together with new technological capabilities.
And perhaps most importantly, it provides a practical starting point for organizations wondering where AI could create meaningful operational value.
Does AI Change the Role of Process Improvement Professionals?
I believe it does.
But not necessarily in the way some people expect.
Many activities that previously required substantial manual effort can now be supported by AI: reviewing documentation, organizing information, preparing initial analyses, generating code, and creating prototypes.
If some of these activities become faster and more accessible, where will experienced process improvement professionals create the most value?
I believe several capabilities will become even more important.
First, identifying the right problem.
A sophisticated AI application that automates unnecessary work is still waste.
Second, understanding the whole system.
An algorithm that maximizes equipment utilization while increasing overall supply chain costs may simply optimize the wrong objective.
Third, challenging assumptions and validating outcomes.
AI can generate convincing answers, including incorrect ones. Domain knowledge, analytical thinking, and critical judgment remain essential.
And finally, leading change.
A technically excellent solution creates little value if people do not understand, trust, or use it.
These are not new principles. They have always been fundamental to operational excellence.
But AI may make them more important than ever.
I also believe the boundary between business improvement and software development is beginning to change.
Experienced process professionals can increasingly experiment with digital solutions themselves, using AI-assisted development tools to translate business ideas into working prototypes.
That does not eliminate the need for professional software engineering, security, governance, or validation.
But it can significantly shorten the distance between identifying an improvement opportunity and demonstrating a possible solution.
And that changes what continuous improvement teams may be capable of achieving.
Learning, Unlearning and Adapting
Looking back, the past several months have been an unexpected journey.
I started the year facing career uncertainty.
I navigated hundreds of job applications and a complicated immigration situation.
During that period, I developed an AI-based startup concept, experimented with software development, experienced rejection, improved my business proposal, and eventually received approval from the Startup Denmark expert panel.
Then I returned to corporate employment and had to reconsider how to pursue my entrepreneurial ambitions.
Now, I find myself exploring how the same technologies could create value across broader business processes, particularly supply chain management.
None of this followed a plan I had prepared at the beginning of the year.
But perhaps that is the point.
For years, I have encouraged organizations to challenge established practices, learn from failures, and adapt to changing circumstances.
The same philosophy should apply to us as professionals.
More than 20 years of experience provides me with a strong foundation. But it also comes with familiar methods, established habits, and assumptions that occasionally need to be challenged.
AI is evolving quickly, and I certainly do not have all the answers.
I am experimenting, making mistakes, learning new tools, and discovering opportunities that would have been difficult to imagine only a few years ago.
My startup ambitions remain alive, even if their timing and direction evolve. At the same time, I see opportunities to apply what I am learning within established organizations.
And I believe this is relevant far beyond my personal experience.
Whether you are an operations leader, supply chain professional, process improvement specialist, or business manager, it may be worth looking at your current improvement activities from a different perspective.
How much effort goes into preparing analyses rather than acting on them?
Which business problems remain unresolved because the analysis is too complex or time-consuming?
Could AI, combined with your existing expertise and analytical tools, help you approach those problems differently?
I do not believe organizations need to abandon proven improvement methodologies to benefit from AI.
But we should be willing to challenge how we apply them.
The future of operational excellence will not be defined by technology alone. It will depend on how effectively we combine human expertise, systems thinking, analytical methods, and artificial intelligence.
Continuous improvement has always been about learning and adapting. Perhaps the most important process we need to improve is our own way of thinking.
Öner Tank — Industrial Engineer | Lean Six Sigma Master Black Belt | Operational Excellence & AI