Christian Byza

Applied Intelligence & AI

The core message of the provided transcript excerpt is that applied AI creates value when it moves past demos and into specific operational environments: the software backlog, the phone line, and the shop floor. The moderator frames the panel around three companies applying AI to concrete workplace bottlenecks: SolveAI for enterprise software creation, Synthflow for phone-based customer interactions, and Dryft for manufacturing operations. The setup emphasizes practical deployment rather than abstract AI discussion, with the moderator explicitly encouraging disagreement and deeper questioning rather than polite consensus. Anna-Julia “AJ” Storch introduces Dryft as a company built around manufacturing expertise. She describes her and her co-founder’s background across automotive, HVAC, traffic signs, medical devices, turbines, and other industrial settings, then explains that Dryft captures operator knowledge across supply chain, production planning, and engineering. Its system simulates the outcomes and consequences of operational decisions, with the goal of helping manufacturers move faster from new product development to customer shipment. Dryft is headquartered in San Francisco, recently opened a Munich office, operates across countries including India, Germany, Austria, and the US, and is hiring for engineering, forward deployed engineering, product, and sales roles. Storch also notes the company’s current high-intensity work culture, including six-day office weeks, while acknowledging that such a pace needs a time limit. Steve Basher then pitches SolveAI from the perspective of his Palantir background, where he saw how relatively straightforward bespoke software could create major enterprise impact. He argues that LLMs have become capable enough to build end-to-end solutions, but that generic AI-generated software is often unsuitable for enterprise production because IT teams need systems built in approved languages, connected to existing infrastructure, and compliant with internal specifications. SolveAI’s value proposition is letting enterprise employees build tools and applications while respecting IT requirements, approval flows, infrastructure, and the highly specific ways companies connect to systems such as SAP. Basher highlights one example where a law firm had been quoted £250,000 for a solution that SolveAI built in two hours, using it to illustrate the gap between demo ware and production-grade AI-enabled software. Hakob from Synthflow closes the visible excerpt by introducing the company’s work in AI voice agents for call centers. He describes a non-technical, business-consulting background and says Synthflow started before the current AI wave, when working with earlier models was a “terrible experience.” Despite the competitiveness of the market, he argues that the space is becoming more interesting as models improve, and he contrasts conversational AI with older IVR systems that force callers to press numbered options. The excerpt ends as he begins answering why many call centers still use “press 1, press 2” systems, so the remainder of the session is not available in the supplied transcript.