Perplexity Chief Business Officer Dmitry Shevelenko argues that computer-using AI agents are a durable productivity category, not a hype cycle. Speaking on the Big Technology Podcast, he makes the case that agentic super apps represent a real business model, not a speculative one, and that multi-model orchestration, pulling from multiple AI systems simultaneously, is a genuine competitive moat rather than a marketing talking point.
The conversation goes deeper than product positioning. Shevelenko addresses the growth slowdown in consumer AI, the economics of AI demand at scale, enterprise adoption friction, and how Chinese open-source models are reshaping the competitive landscape on pricing. These are the structural questions the industry is avoiding, and this interview does not avoid them.
What makes this worth reading in full is the tension between Perplexity's ambitions and the hard constraints it names: trust, permissions, and the economics of running inference at scale. Shevelenko is specific about where the bottlenecks are, which is more useful than any roadmap. The Big Technology AI Summit in San Francisco on June 18 is where these threads continue.
[WATCH ON YOUTUBE →]