Personalised print campaigns used to take days. Now Genscom ships them in 10 minutes. With its AI-powered Curator, Genscom freed up 340 hours of manual work a week, tripled revenue and doubled its headcount. See how we made it real → hubs.la/Q04ypKHS0 #AgenticAI #GenerativeAI

We've had agents since 2016. Long before it was cool. Our agents don't run on tokens. They run on coffee, curiosity and a drive to turn AI hype into lasting business impact. 10 years at the AI frontier. Now we're scaling our agents—by hiring them. Become one 👉 hubs.la/Q04yjdWS0 #Hiring #LifeAtML6

The thicker the policy binder, the safer we feel. But what if AI governance is compensating for missing design? Francesco Cilurzo argues for codifying guardrails, friction, overrides, explainability and accountability into AI systems by default—leaving governance for the hard questions design can’t solve. Read more: hubs.la/Q04yjd1t0 #AIGovernance #ResponsibleAI

Autonomous cyberattacks aren’t a future problem. They’re already here. As AI models handle longer, more complex tasks, attackers can target hundreds of systems in parallel. Is your security ready for attacks at AI scale? See how we help teams prepare → hubs.la/Q04yd3qR0 #Cybersecurity #CyberRisk

AI is changing decisions across the entire energy asset lifecycle. ⚡ From smarter design and development to real-time operations, optimisation and trading. Over the next 3 weeks, we’ll explore where AI can create tangible value, starting with Asset Design & Development. #EnergyTech #EnergyTransition

Good AI ideas shouldn’t stay ideas because time, budget or resources are missing. ML6 for Good helps turn meaningful AI challenges into action. Got an AI challenge with real-world impact? Let’s build it. 🔗 hubs.la/Q04xw9KD0 #AIforGood #ML6forGood #TechForGood

From pilot to scale: HeyImmo, ImmoScout24’s AI assistant, cut response times from 8s to under 5s while improving satisfaction from 68% to 75%. See how ML6 and Scout24 reworked the architecture to scale speed, quality and accuracy → hubs.la/Q04xlygj0 #AI #AgenticAI

The cheapest AI model isn’t always the cheapest option. GPT-6 Astra costs 2.5× more per token than its predecessor. But on a 300+ file migration, Astra got it right first time for $10.80. The cheaper model? $18.25 to get a working result. Don’t compare token price. Compare the cost of a finished job. 🔗 hubs.la/Q04xc-Gs0 #GPT6Astra #GPT6

We don't follow the hype: we build the future. UNUM is one of the ways we do it. We're now exploring platform engineering alongside our custom project work, and that means two kinds of engineers we didn't need to hire two years ago. 👉 Check our current openings to be part of the journey: hubs.la/Q04w_-_g0

In life sciences, a better model rarely fixes the real problem. AI gains often vanish once messy data gets cleaned up. It wasn't the model. It was the data. We saw it firsthand building an AI tumor board with Princess Máxima Center — specialists lost hours to scattered data before diagnosis even started. 🔗 hubs.la/Q04wQDg80 #LifeSciences #HealthTech

#VoiceAI in support isn't a switch between "fully automated" and "not used." It's a design question: what should happen while the customer waits? In this clip, Jens Crokaert walks through how it plays out based on queue time alone. 🎥 Watch the full recording: hubs.la/Q04wv1hT0 #CustomerExperience #ConversationalAI

More applications. Fewer strong matches. AI made applying nearly free — but ML6 saw applications rise 45% while offers rose only 12%. We're not screening with AI. We're adding steps that are harder to fake. 🔗 hubs.la/Q04w2p9n0 #FutureOfWork #HRTech

What actually gives an #AIagent the power to act? In July, an AI agent reportedly breached part of Hugging Face's prod environment while chasing an OpenAI security benchmark. Not the model's intelligence alone — the harness around it. Full breakdown 🔗 hubs.la/Q04w26yY0 #Cybersecurity #SecureAI #AgenticAI

📍 ML6 is growing, and yes, also in Munich. We opened the doors to our Munich office in 2026. Now, we’re ready to bring more brilliant people through them. We’re scaling the team across the board; from Sales and Project Management to Senior Engineering roles and our UNUM team. Want to build next level AI and help shape ML6’s next chapter in Munich alongside Jasper, Luca and Konstantin? 👉 Apply: hubs.la/Q04vl8Nd0

UI testing tools have existed for years. The bottleneck? Writing and maintaining the tests. We built cua-uat, putting Computer Use Agents to work on it: mapping user flows, finding failures and turning natural-language flows into Playwright tests. On a real test app, it caught 27/28 planted bugs and generated 14 Playwright tests that made it into a PR. Biggest lesson? Input quality mattered more than the model. 🔗 hubs.la/Q04vl5_R0 #AgenticAI #UITesting #AIEngineering

𝗣𝗥𝗘𝗦𝗦 | Anthropic is the first AI company to watermark all its output — every Claude text now carries an invisible marker. Our own Ryan Ott (ML6) told @kristofvds why: it's not just the EU AI Act. AI companies are running low on fresh training data, and watermarks help them filter their own text back out. 🇳🇱 hubs.la/Q04v9SGr0

As part of ML6 for Good, we partnered with the AMES Foundation to build AI-enabled drone surveillance on Google Cloud, turning live video into real-time alerts for anti-poaching teams. Responsible AI. Real impact. 👉 Read the full case study. hubs.la/Q04tNVYH0

Voice AI can catch a name in seconds. Numbers are harder. Take Belgium’s 11-digit national registration number. Before hitting the database, the system can first check: is this even a valid 11-digit input? Small design choice, real impact: lower latency and fail fast instead of fail expensive. One detail from our webinar on taking Voice AI from pilot to production 👇hubs.la/Q04tJVqn0 #VoiceAI #ConversationalAI #CX

Most companies don’t have an AI usage problem. They have a context evaporation problem. Engineers learn how to make AI work on a codebase - the right context, workflows, guardrails - but that knowledge disappears when the chat ends or the project changes. The fix: context as code. In our latest blog, Georges Lorré explains how to capture that knowledge and make it scale across an engineering org. 🔗 hubs.la/Q04tw4YB0 #AIEngineering #AIAdoption #AICoding

Where does your team’s AI knowledge go when the chat closes? Too often, the next engineer starts from zero. Treat AI context like code: put it in the repo. An hubs.la/Q04sVdGb0, reusable skills, reviewed via PR. One project. One AGENTS.md. One skill. Read the breakdown: hubs.la/Q04sV79s0 #AIEngineering