Notes on building, validating, and operating with AI
Short pieces on what we learn building our own SaaS products with artificial intelligence at the center of the operation.
Inside Cliniotica: the problem behind an ERP for ophthalmology clinics and optical stores
Cliniotica is still a prototype, not an available product. This is how a real problem becomes a candidate for the DeltaEdge AI catalog before it reaches subscription.
November 16, 2026
How the DeltaEdge AI subscription catalog works
Subscribing to a product in the DeltaEdge AI catalog is not commissioning development. It is using a SaaS that already exists, operated by us, with the cost of evolution shared across every subscriber.
November 9, 2026
Build, validate, operate: the lifecycle of an AI-native SaaS product
Treating building, validating, and operating as three separate phases, rather than a linear sequence where the next one only starts once the last one ends.
November 2, 2026
Why human validation still matters in AI-driven workflows
The more a process relies on agents, the more critical it is to define precisely where a person needs to confirm the result before it moves forward.
October 26, 2026
AI agents in internal workflows: where they actually save time
The real payoff from AI agents is not in rare, complex tasks. It is in frequent, well-defined tasks with low risk per individual run.
October 19, 2026
Validate your own SaaS idea with AI before you write code
AI does not replace talking to real customers, but it can compress weeks of early research into days, without skipping the step that actually decides whether the idea survives.
October 12, 2026
What it means to be an AI-native company (and why that's different from "using AI")
Using AI tools and being an AI-native company are different things. The difference is where AI enters the decision process, not how many tools sit in the stack.
October 5, 2026