Explainable AI begins with a simple problem. Artificial intelligence has become very good at giving us answers, but the real difficulty starts when we ask it a simple question: why? An accurate prediction may look impressive in a demonstration. But when it influences...
AI in Solution Architecture: Who Still Holds the Compass?
by Javier Elorduy | Mar 11, 2026 | Artificial Intelligence, Architecture
AI in solution architecture brings a familiar promise to the table: this time, the machine will make the difficult decisions for us. Artificial intelligence has brought that promise to the architect’s desk. It can compare patterns, generate diagrams, expose risks and...
Small Language Models: Why Smaller Can Be the Smarter Architecture
by Javier Elorduy | Feb 23, 2026 | Artificial Intelligence
In the race for artificial intelligence, many companies are buying a cargo ship to deliver an envelope. They connect their simplest tasks to the largest models available, then wonder why the bill grows faster than the business value. Small Language Models offer...
Event Driven AI Architecture: Why Your ERP Should Stop Answering Every Question
by Javier Elorduy | Feb 11, 2026 | Artificial Intelligence
The Illusion of Simplicity: Why Direct Integration Is a Trap An event driven AI architecture often begins with a simple temptation: connect the AI directly to the ERP. The REST API already exists, the data is there, and the reasoning seems sound. If the chatbot needs...
AI Cost Optimization: Five Architecture Decisions That Keep Spending Under Control
by Javier Elorduy | Jan 23, 2026 | Artificial Intelligence
An AI pilot can look inexpensive right up until it becomes useful. The demo works. People begin using it. More documents are indexed, more requests are sent, and the most powerful model is called for tasks that never needed it. Then the first production invoice...
AI Security Architecture: Five Pillars for Secure Production
by Javier Elorduy | Jan 4, 2026 | Artificial Intelligence
AI security architecture becomes a serious concern when someone asks one of the most feared questions in a digital strategy meeting: “We already have the pilot working. Can we open it to the whole company tomorrow?” The short answer is no. The longer answer begins...
Enterprise AI Architecture: What It Takes to Move from Pilot to Production
by Javier Elorduy | Dec 21, 2025 | Artificial Intelligence
An AI prototype can impress a room with a single answer. A production system has a harder job: it must survive thousands of imperfect requests, incomplete data, partial failures, security constraints and changing business conditions. This is the moment when a...
Measuring AI Impact: From Adoption to Business Value
by Javier Elorduy | Dec 1, 2025 | Artificial Intelligence
AI adoption is easy to count. Licenses activated, pilots launched and employees trained all produce visible numbers. None of them proves that the organization is making better decisions, serving customers more effectively or reducing risk. Measuring AI impact begins...
Autonomous AI Agents: From Automation to Better Business Decisions
by Javier Elorduy | Oct 14, 2025 | Artificial Intelligence
Autonomous AI agents are changing the business case for artificial intelligence. For years, that case was relatively straightforward: automate repetitive work, reduce errors, and cut costs. That logic still matters, but it no longer describes where the technology is...
FROM EXPERIENCE TO INSIGHT
Software, AI and Architecture Insights Shaped by Real Work
The Intech Heritage blog shares practical insights from building SaaS products, custom software and AI systems. We explore software architecture, product development and the technical decisions that help businesses create useful, scalable and reliable technology.








