Sifted Talks: How Scaling Companies Are Implementing AI in Production

The transition from experimental use to widespread deployment of generative AI presents significant operational challenges for scaling companies. These include increasing costs, establishing governance, and building trust in autonomous AI functions.

These issues were central to a recent Sifted Talks event, where experts shared insights on how companies can successfully move beyond AI pilot phases and build resilient, AI-native operations. Omar Davison, solutions engineer at Box, emphasized that identifying where AI delivers the most value – whether for individual productivity, departmental efficiency, or organizational optimization – is crucial. For less AI-mature organizations, immediate gains are often found in enhancing individual time efficiency.

Thibault Martin, ecosystem lead at Dust, noted that technology itself rarely hinders AI scaling; rather, the lack of clear frameworks for security and budget allocation is often the bottleneck. He compared AI agents to new resources requiring defined goals and accountability for business leaders. Lucien Bredin, cofounder at Naboo, shared an example where their AI twin handles 80% of event organization daily, with human oversight for trust.

Growth investor Jannat Rajan highlighted that integrating AI into core processes can significantly reduce company margins. Successful firms are managing these costs by diversifying their AI models and repricing software. She advocated for a portfolio of AI engines, similar to the early "multi-cloud" trend, to optimize spending and avoid vendor lock-in, stressing the importance of specialized financial operations for AI costs.