.png)
Thoughts
Essays on AI, software and building companies.
Everything we have written, in one place. All of it comes from real project work in Zürich, published when we have something to say. Start anywhere.
Why AI Makes Things Up (and How We Engineer Around It)
Hallucination is not a bug to patch. It is a property to contain, in four layers.
Read the essay →A Field Guide to Problems Nobody Can Reproduce
The bug is always reproducible. You just have not found the conditions yet.
Read the essay →Context: The Real Bottleneck in Business AI
The models are interchangeable. What decides usefulness is what they know about you.
Read the essay →AI Agents in Production: What Survives?
Where demos diverge from production, and why the survivors are the supervised ones.
Read the essay →How to Measure the ROI of AI, Honestly
The J-curve, the three layers of return, and the numbers we can actually defend.
Read the essay →Build or Buy? A Decision Framework for Software and AI
The framework we use with clients, including the third option most comparisons miss.
Read the essay →Is Your Data Ready for AI?
You need less than you think, organised better than you have it.
Read the essay →How to Choose an AI Partner (and When You Don't Need One)
The red flags, the questions worth asking, and the cases where the answer is nobody.
Read the essay →AI and Data Protection in Switzerland
What the FADP actually requires, where the data may live, and the four questions that decide most cases.
Read the essay →Technology and Long-Term Thinking
Why chasing trends is expensive, and the five traits of systems that last.
Read the essay →The Tool That Learned to Think. Now What?
AI is not a faster tool. It is a mirror for how your business actually works.
Read the essay →The Fortune 5,000,000
Everyone fights for the Fortune 500. We build for the five million companies that run the economy.
Read the essay →What Are the Most Common AI Implementation Mistakes?
Seven mistakes we see most often in real projects, and the fix for each one.
Read the essay →How Do You Integrate AI Into Your Existing Software Systems?
Three integration patterns that work without rebuilding what you have.
Read the essay →Should You Modernize or Replace Your Legacy Systems?
Most companies should modernize incrementally. A framework for a defensible decision.
Read the essay →How Do You Simplify Complex Data Systems?
Complexity is accumulated decisions. A practical sequence for unwinding it.
Read the essay →When Should a Swiss Startup Outsource Software Development?
When speed to a validated product matters more than building a team.
Read the essay →What Does Secure Custom Software Cost in Switzerland in 2026?
Honest price ranges, from discovery to production systems.
Read the essay →The Uliasti Guide to Making Decisions
Clear thinking and practical frameworks for better business decisions.
Read the essay →No Big Deal or End of the World?
How small issues become crusades, and the two-token choice in customer experience.
Read the essay →Why AI Pilots Never Reach Production
Most pilots work in the demo and still never ship. The gap is ownership, not technology.
Read the essay →The Spreadsheet That Runs Your Company
Every company has one load-bearing Excel file. How to spot it and replace it without drama.
Read the essay →Technical Debt Works Like a Loan
You only pay interest on code you touch. Repay where you work, ignore the rest.
Read the essay →What to Automate First
Frequency beats drama. A simple rule for picking automations that actually pay.
Read the essay →Why Software Estimates Are Always Wrong
Estimates fail because the work is learning, not construction. How to buy software with that in mind.
Read the essay →The Software Nobody Opens
Tools do not fail in the demo. They fail in week three. Four things that predict adoption.
Read the essay →We Are an Official Claude Partner
What the partnership means and where Claude pays off first, industry by industry.
Read the essay →
