Q agency

15/10/2026

This time, the lights stayed on. 💡

After a power outage put the first event on hold, our Al Meetup happened this Tuesday, with Zlatko Matokanović taking the room through Human-Agent Pods. 🤖

If you haven’t picked up on it yet, we’re talking about a spec-tirst approach where agents move tast, humans hold the gates, and every shipped feature can be traced back to what was actually asked for.

Because ‘Al made it’ is cool. But ‘here’s exactly how and why it works’ is way better.

Big thanks to Zlatko and to everyone who joined us, asked questions, and made round two well worth the wait. 😎

05/10/2026

Manual QA works. Until it doesn’t scale anymore. ⚖️

When manual website checks stopped scaling, Q’s QA team rebuilt the suite with , and , with one goal: make it fast to build and hard to break.

The result covers dozens of pages 🔎, multiple viewports, navigation, content and SEO without turning into a pile of scripts nobody wants to touch.

In our latest blog post, we break down the workflow, the rules and what we learned along the way.

How Q built 110+ Playwright tests for q.agency using Cursor and Playwright MCP: smoke, regression, content, and SEO coverage with CI gates.

01/10/2026

[AI Meetup, take two 🔌]
Our first attempt got taken down by a power outage that hit the whole neighbourhood.⚡

But we’re not letting one blackout have the final say.

On 13 October, Zlatko Matokanović, our Director of AI R&D, is taking the AI Meetup through Human-Agent Pods, a spec-first delivery model where agents work across the SDLC, humans own the gates, and every requirement can be traced all the way to shipped functionality. 🎯

Because speed is no longer the bottleneck as implementation is getting easier, but specs and verification aren’t.

Expect a hands-on look at the actual workflow, from spec review and pod setup to guardrails, verification and the metrics behind provable delivery.

📍 Q agency, Reljkovićeva 4 | 🗓️ Tuesday, 13 Oct | 🕠 17:30

Save your spot here by clicking the link in our bio. 🔗

Come for the talk, stay for the questions. And snacks.

See you there.

30/09/2026

Well, this should be a good one! 😎
Q is joining , bringing our experience in digital product development, AI and turning demanding business challenges into solutions with proper business value.

And our CEO Hrvoje Gorajščan will also be there and share his views on what makes the difference when access to technology is no longer the main advantage. 🗣️

If you’re at MBF, come find us and talk tech, business or whatever’s on your mind.

See you in Rovinj, 8–11 October. 📍

29/09/2026

What happens when one Java API has to make sense of the same profile data coming from several different systems? 🧩

This blog explores that problem through a real implementation of the , from the contract itself to the role of each visitor and how the service executes them.

It shows how enrichment and validation logic can stay separate, while also explaining when that extra structure is worth it, and when a simple mapper is the better choice.

🔎 Read the full story for a practical look at how the Visitor pattern works in a real Java API. ⤵️

04/09/2026

Flutter makes cross-platform development easy. Push notification actions are one of the exceptions. 📲

Firebase Messaging doesn’t currently provide an out-of-the-box way to implement notification actions across iOS and Android.

So this guide breaks down the native setup for both platforms, how to bridge those actions back into , and how to test the whole flow. 🛠️

From quick replies to contextual actions, here’s how to implement it properly. 👇🏼

20/08/2026

What happens when speed stops being the differentiator in software delivery❓
The harder question is whether you can reconstruct how a piece of work was produced, what it passed, and who put their name behind it. 🛠️

That’s the thinking behind Human-Agent Pods: a delivery model built around traceable artefacts, explicit gates and human accountability at every stage.

The result is a process you can inspect instead of simply trust.

Read the full story to see how it works from spec to release. ⬇️

IDE copilots govern access, not work. Human-Agent Pods run AI in a governed harness where a named human signs every artefact. Part 1 of 2 from Q.

20/07/2026

Your data may be real-time, but your architecture might not be. ⚡

When thousands of devices send information continuously, every weak point in the pipeline becomes visible.

Our latest blog explores how , Azure Data Explorer and .NET can turn complex telemetry flows into reliable, low-latency analytics, from covering the architecture behind data collection, multi-protocol ingestion to faster decision-making.

Read the full post below!

Learn how Azure, Azure Data Explorer and .NET support real-time data ingestion, telemetry pipelines and low-latency analytics.

02/07/2026

Before you build, choose the risk you’re actually trying to reduce. 🧭

A POC, prototype, and MVP are not three names for the same thing. One tests feasibility, one tests usability, and one tests market demand.

Pick the wrong one, and you can spend weeks proving something that was never the real problem.

Read this piece for a practical breakdown of when to choose a POC, prototype, or MVP, and how to reduce project risk faster.

📚🔗 https://q.agency/blog/poc-vs-prototype-vs-mvp-which-reduces-risk-fastest/?utm_source=facebook&utm_medium=social&utm_campaign=june_26_POC_prototype_MVP

25/06/2026

Most AI projects fail before the technology becomes a problem. ⚠️

Because the real issues are usually:

• the wrong use case
• unclear ROI
• undefined production path
• weak validation
• no operational ownership

So instead of starting with a massive AI transformation initiative, we built something smaller, faster, and easier to pressure-test: the . 🏃🏼

📊 The outcome? In just 10 days, clients get:

• a working prototype
• validated metrics
• an assessment of business value
• a clear go/no-go direction for next steps

This carousel walks through you the idea in a few slides, and the full blog goes deeper into how it works. ⚙️

Head to the link to read the full story: https://q.agency/blog/agentic-sprint-proving-ai-value/

Have a workflow you’d like to pressure-test? Reach out to our team and let’s explore what an AI agent could realistically do with it.