From AI pilots to P&L: What leaders must do to scale AI value

Six takeaways for CEOs and investors 

1. Start with the value chain, not the tool. AI should solve business problems with a clear economic case. Vanessa Stützle, CEO of LUQOM Group, described how the lighting retailer uses AI across product design, content creation, marketing and operations. According to the figures presented, AI-enhanced lifestyle imagery delivered a 43% uplift in click-through rate in testing, while the company reported more than 40% EBITDA growth in 2026. These outcomes illustrate the importance of embedding AI in strategic, revenue-generating processes rather than treating it as a standalone experiment. 
2. Make proprietary data usable and reliable. Rajesh Shanmugasundaram, CTO & Chief AI Officer at Breitling Group, stressed that companies need control of their own data before they can create meaningful AI applications. He shared examples spanning AI-assisted product discovery, conversational commerce, store merchandising, demand forecasting, and accounts payable. Dr. Andreas Fruth, Co-Founder & Managing Director of Atolls, showed why access alone is insufficient: without clear definitions and data lineage, even capable models can produce misleading business analysis. 
3. Put AI-building capabilities in the hands of employees. Atolls has deliberately enabled non-technical teams to develop and test their own workflows. At a company-wide hackathon, employees submitted 43 ideas and built 30 prototypes. Fruth described how domain experts can now create useful software themselves, reducing the handoffs between business teams, product managers, and developers. Governance, security and ownership then become critical as those experiments move toward production. 
4. Design end-to-end processes, with adoption built in. Oliver Herzig, Chief Product Officer at eurodata, presented an AI automation layer for tax and accounting practices. Its “virtual employee” achieved 88% accuracy from the outset and up to 97% after learning in the pilots he discussed, while reducing manual processing time. Yet Herzig emphasized that the harder challenge was earning the trust of practitioners and scaling adoption through training, local ambassadors, and aligned incentives. The objective is not automation for its own sake, but freeing capacity for higher-value client advisory work. 
5. Combine automation with trust, accountability and measurable economics. Elias Apel, CEO of Lucanet, highlighted the importance of bringing deterministic calculations together with AI reasoning in finance, where traceability and auditability are essential. He reported 98% daily AI usage within Lucanet, supported by a company-wide transformation of workflows and ways of working. Roy Tuvey, Co-Founder & CEO of Vertice, framed a related challenge: enterprises must rethink how they purchase, govern, and measure software and AI spend as technology economics change. 
6. Rethink the operating model for an AI-native enterprise. Matias Undurraga Breitling, Principal Enterprise Technologist at AWS, argued that AI-native organizations require more than a technology strategy. Leaders need to redesign decision-making, budgets, team structures, and jobs around cooperation between people and AI agents. Raffael Fruscio, CEO of RAAN Corporate Group, brought the discussion into the circular economy: connecting material flows and data can help move an established industry from waste management toward more intelligent, closed-loop resource systems. Both perspectives point to a transformation led by leadership and adopted across the business.