Insights

The AI agent, one year later: a 2026 architecture - and where it actually plugs into a banking platform

Ivan Kun — Fri, 24 Jul 2026 13:03:58 GMT

A year is a geological era in agentic AI. Twelve months ago, a typical “AI agent platform” diagram (we're looking at one from mid-2025 as we write this) put a vector database at the center, treated agents as a sub-box inside an “AI layer,” wired every tool integration by hand, ran monitoring as four dashboards (cost, usage, data, LLM), and handled deployment as classic CI/CD with prompts as an afterthought.

From Regulation to Architecture: Designing Human Oversight for AI Systems

Valentin Mazareanu — Fri, 24 Jul 2026 12:38:36 GMT

What regulators expect

One of the most common misconceptions I encounter is that “human oversight” simply means placing a person at the end of an AI workflow to approve its output. In reality, that interpretation falls short of both the spirit and the intent of modern AI regulation. Frameworks such as the EU AI Act, ISO/IEC 42001, and the NIST AI Risk Management Framework all recognise that humans must remain capable of understanding, supervising and, when necessary, overriding AI-driven decisions.

DORA and Your SDLC: The 5 Things Auditors Really Check

Valentin Mazareanu — Fri, 24 Jul 2026 12:31:26 GMT

DORA (the Digital Operational Resilience Act) has changed the way software development is assessed. During audits, the discussion rarely focuses on the wording of the regulation itself. Instead, auditors want to understand whether the Software Development Life Cycle (SDLC) consistently produces software that is governed, secure, traceable, and resilient. The difference between a mature organization and one that simply claims compliance is almost always found in the evidence generated throughout the development process.

AI in Regulated Software Delivery: DORA, EU AI Act, ISO & FINMA - The Complete Guide

Valentin Mazareanu — Fri, 24 Jul 2026 12:29:56 GMT

Artificial intelligence has fundamentally changed how software is designed, developed, tested, and maintained. Development teams now rely on AI to generate code, review pull requests, create test cases, produce documentation, and even suggest architectural improvements. For regulated organizations, however, the discussion has never been about whether AI improves productivity. The real challenge is demonstrating that every AI-assisted change remains controlled, attributable, and compliant with the expectations of regulators and auditors.

Rethinking Software Modernization: Replacing Months of Discovery with AI Modernization Framework

Andrei Postaru — Fri, 24 Jul 2026 12:25:06 GMT

The Uncomfortable Truth About Software Modernization: Why Projects Stall in Discovery

The dual-run bill: why time is the real cost of modernization - and how AI shrinks it

Daria Muehlethaler — Fri, 24 Jul 2026 12:23:33 GMT

Here is the sentence most modernization business cases are afraid to put on slide one: for the entire duration of the migration, you will pay for two systems.

Is AI-generated code allowed in a bank? A traceability answer.

Valentin Mazareanu — Fri, 24 Jul 2026 12:23:02 GMT

Somewhere in Switzerland this week, an engineering lead will be asked the question in a risk committee: “Are we actually allowed to put AI-generated code into production?” It sounds like a yes/no question. It isn't — and treating it as one is where most internal AI policies go wrong.

AI ambitions, legacy reality: what "not modernizing" actually means

Daria Muehlethaler — Fri, 24 Jul 2026 12:17:23 GMT

There is a pattern playing out in boardrooms right now, and the numbers describing it are brutal enough that they need no dramatization.

Agentic System Design: Patterns That Work When the Demo Becomes a Product

Andrei Postaru — Thu, 25 Jun 2026 11:03:40 GMT

Every agentic system has a demo that works. The agent takes a goal, calls a handful of tools, reasons its way through the task, and produces an outcome that makes the room lean in. It's compelling, fast, and convincing enough that everyone starts asking the same question: When can we ship it?

How We Review AI-Generated Code Without Slowing Down the Team

Andrei Postaru — Tue, 12 May 2026 09:09:24 GMT

A year ago, AI-generated code in pull requests was a curiosity. Today, in many of the codebases we work in, it's the default.

What hasn't kept pace is how teams review it. We see two failure modes, both broken.

The first is review theatre: a reviewer scrolls through a 400-line diff, sees that it compiles and tests pass, and clicks “LGTM” without fully understanding what changed. The assumption is that if the model wrote it and nothing failed, it must be fine.

The second is the opposite extreme: over-review. Teams react to the risks of AI-generated code by scrutinizing every line. Naming, formatting, structure — everything gets debated. Reviews take longer, throughput drops, and the team slows to a crawl.

Neither approach scales. Neither reflects what code review is actually for.

The model that works — the one we've settled on across our embedded engagements — treats review as a layered system.

Machines handle form. Humans guard intent.