On this episode of Scouting For Growth, Sabine VdL speaks with Ivan Pylypchuk, CEO of Softblues, and Olha Pylypchuk, the company’s COO, to unpack what many leaders are sensing but struggling to operationalise: Agentic AI isn’t a feature upgrade. It’s a new operating model. Together, they explore what agentic AI really is, why multi-agent systems are poised to disrupt traditional business processes, and what it takes to deploy AI agents that teams actually trust and use. Agentic AI: from “single task” to autonomous execution Ivan breaks down the difference clearly. Traditional AI systems typically handle one narrow task: image recognition, classification, or content generation. Agentic AI goes further. It can manage multi-step decision-making, coordinate actions, and operate with higher autonomy—especially when designed as controlled multi-agent systems, where each agent has a defined role and guardrails. In other words: not just intelligence… agency. The real blocker isn’t AI — it’s data fragmentation Softblues sees one challenge again and again: companies have valuable data spread across disconnected systems—CRMs, email, customer databases, business tools. And when those systems don’t talk to each other, AI can’t deliver reliable outcomes. Their message is simple: for AI to work, data must be collected neatly, accurately, and integrated across the organisation. Without that foundation, even the best model will produce weak results. The hidden issue: companies often don’t understand their own workflows One of the most striking insights from Ivan and Olha is that many organisations don’t actually know how work gets done day-to-day. When they observe real processes, they often find missing steps, informal workarounds, and operational blind spots—details leadership didn’t realise existed. That matters because implementing AI on top of an unclear process can lead to wasted investment later. Softblues addresses this by spending serious time mapping workflows from Point A to Point B, then enhancing them with AI rather than forcing automation into chaos. Adoption wins or loses everything Ivan and Olha emphasise that even a “perfect” AI solution fails if teams don’t use it. That’s why they focus on: simple interfaces recommendations explained in plain language and explainability that builds trust quickly Agentic AI only scales when humans feel confident in what it’s doing—and why. Start small. Prove value. Then scale. Their deployment philosophy is practical: start with a small part of the project where AI can show fast results. Once value is proven, scale confidently. It’s how organisations move from experimentation to transformation without breaking operations—or trust. Real outcomes: speed + quality Softblues shares measurable impact from their solutions, including: reducing time-to-hire by 60% increasing quality of hires by 40% The bigger payoff, though, is strategic: freeing teams from data wrangling so they can focus on higher-value work that moves the business forward. Why this episode matters For enterprise leaders in insurance and financial services, this episode is a blueprint for agentic AI adoption that’s grounded in reality: agentic AI thrives in controlled multi-agent systems data integration is non-negotiable workflow clarity comes before automation trust and usability drive adoption small wins create scalable momentum Because the future of work won’t be shaped by AI that’s impressive in demos. It will be shaped by AI agents that quietly make organisations faster, smarter, and more strategic—every single day.