How to build AI-native products with Des Traynor, Nicolas Sharp, and Vas Zhovner

Last week, Marc Hillander - Managing Director of CUBE Nordics - sat down with Robert Pop, our Head of AI at We As Web. Robi’s been building real-impact software for 10+ years—and agent-based systems for 3, before they were even a thing. Marc’s first question: “When should you use agents?” Robi’s answer: “When tools aren’t enough. But never just because it sounds cool.” That mindset drives our work: start tool-first, add agents only when necessary, and optimize every step. It’s how we’ve turned complex flows from 40s → 2s in production. Link to Marc’s original post (in Swedish) in the comments. #AI #AgenticAI #MLOps #Engineering #WeAsWeb #CUBE

Agent-based systems are powerful, but only when they serve a clear purpose. Impressive insights from Robi and Marc on how disciplined engineering choices lead to production results that truly scale.

Revolutionize Your Business with Low-Code AI Tools — In an era where technology continues to evolve at a breakneck pace, businesses are constantly seeking innovative solutions to streamline processes, enhance efficiency, and gain a competitive edge. Low-code AI tools have emerged as a powerful means to achieve these goals. By minimizing the complexity associated with software development, these platforms empower… https://lnkd.in/d-SMKtXH #2025Innovations #AiProductivityAssistant #AiTools #BusinessTransformation #LowCode #TechnologyTrends

Revolutionize Your Business with Low-Code AI Tools — In an era where technology continues to evolve at a breakneck pace, businesses are constantly seeking innovative solutions to streamline processes, enhance efficiency, and gain a competitive edge. Low-code AI tools have emerged as a powerful means to achieve these goals. By minimizing the complexity associated with software development, these platforms empower… https://lnkd.in/dGuXVFan #2025Innovations #AiProductivityAssistant #AiTools #BusinessTransformation #LowCode #TechnologyTrends

A few years ago, our development sprints looked very different. Whiteboards filled with logic flows, long nights of debugging, and endless lines of code — all written manually, piece by piece. Then came AI. At first, it was just another buzzword. But slowly, we began to see the shift — smarter tools, predictive analytics, AI copilots, and automation that didn’t just assist developers, but amplified their creativity. Today, our engineers at Thrymr don’t just build software — they train intelligence. They collaborate with AI to design better architectures, predict issues before they occur, and build systems that actually learn from user behavior. The heart of development hasn’t changed — it’s still about solving real problems. But the way we do it? It’s faster, smarter, and far more human than ever before. The fusion of human ingenuity and artificial intelligence isn’t the future anymore — it’s our everyday reality. 🚀 At Thrymr, we’re excited to be shaping this new era of intelligent software. #AI #SoftwareDevelopment #Innovation #DigitalTransformation #ThrymrSoftware #Technology #FutureOfWork

Every software team feels it — the pressure to build faster, smarter, and better. But here’s the truth: AI isn’t replacing developers. It’s empowering them. Our new carousel dives into 3 AI trends reshaping software in 2026 — trends that are already changing how teams code, test, and launch. If you want to stay ahead of the curve, start here. Because in 2026, those who understand AI won’t chase innovation, they’ll create it.

🚀 Bringing Agentic AI into Application Reliability Engineering at PickMe We’ve taken a step forward by introducing Rovo AI Agents from Atlassian, bringing Agentic AI into our workflow to: 🤖 Instantly analyse incoming incident tickets 💬 Request missing details from reporters in real-time 📚 Recommend investigation paths based on past incidents & knowledge base ✅ Evaluate investigation quality and flag tickets needing deeper analysis This has transformed how we handle incidents — reducing turnaround time, increasing consistency, and helping our engineers focus on what truly matters: delivering reliability with intelligence. We’re also exploring deeper AI integrations — from real-time observability using Prometheus + OpsGenie to automated knowledge extraction from Confluence and Jira. Read the full article on Medium for more insights. #AI #AgenticAI #ApplicationReliabilityEngineering #PickMe #Innovation #AtlassianRovo #EngineeringExcellence #Automation

Production-grade GenAI isn’t just plug-and-play. What separates a demo from a scalable, secure deployment? It’s the invisible engineering: vector DBs tuned for latency, RAG pipelines that don’t hallucinate, and cost controls that actually work at scale. Most teams stall when PoCs break under real-world load—or when integration with cloud-native stacks exposes hidden complexity. Our team bridges that gap: deep backend roots, relentless AI R&D, and a business-first mindset. We build GenAI systems where reliability, observability, and cost efficiency are engineered from day one—not retrofitted after launch. Ready to move from prototype to production? Let’s talk about real-world GenAI that scales with your business. 💡

🚩 Claiming to be a tech company but not owning your code. ✅ Building on a platform to accelerate your product delivery How does this apply to Voice AI? Tech companies are making absolutely wild decisions that will come back to haunt them right in a rush to implement voice AI. Real tech companies are relying on point solutions for their core product offering or they are trying to scale enterprise grade products using no-code tools. Absolutely wild decisions that will have repercussions down the line. Given the urgency, how does a tech leader seize the massive voice opportunity quickly without making irreversible decisions? 1) Understand that no-code tools are for prototyping and ultra-simple use cases. Serious tech enterprises don't build products on no code tools. 2) Use a platform to handle voice infrastructure & tooling (don't stitch together multiple. Find a trusted partner that can do it all for you) 3) Make sure your product is written in code that you own. Even if someone helps you build V1, you aren't a tech company if you don't own the code. Happy Friday and happy building!

Generative AI is rewriting the rules for open-source software—what was once built on collaboration may be turning into fractured silos. ⚠️ 🔗 https://bit.ly/4oIy8zK #OpenSource #GenAI