Perplexity is increasingly positioning itself as more than an AI-powered answer engine. In a recent conversation on The Deep View Conversations, chief business officer Dmitry Shevelenko said the company now sees a future built around AI agents, digital coworkers and highly leveraged teams rather than just search.
Shevelenko described Perplexity’s evolution from a relatively small startup into a larger company focused on building tools that can do more of the work people once handled themselves. He said the company has grown from about 20 employees to roughly 400, reflecting both demand for its products and the broader expansion of AI across the business world.
A central theme of the interview was accuracy. Shevelenko said Perplexity has treated reliability as a core principle, rather than an optional feature, as it tries to stand out in a crowded AI market. That focus, he argued, is especially important as AI systems take on more complex tasks and users depend on them for decisions that carry real consequences.
The discussion also touched on how Perplexity has changed its business strategy. According to Shevelenko, the company moved away from advertising as a main path to growth. Instead, it has been leaning into products that support AI-driven workflows and new forms of compute, including Perplexity Computer, which can coordinate several AI models at once.
Shevelenko said the rise of AI agents is likely to alter how knowledge work is done. Rather than serving only as chatbots or search tools, these systems are expected to act more like digital coworkers that can help complete multi-step tasks. That shift, he suggested, could reshape everything from company structure to hiring.
One implication is that smaller teams may be able to do much more with less headcount. Shevelenko said the future may favor compact organizations with outsized output, especially when supported by AI systems that can automate or accelerate routine work.
He also pointed to pressure on entry-level roles. As AI takes over more basic knowledge tasks, the traditional path into the workforce could change. In that environment, Shevelenko said entrepreneurship may become a more common route for people starting their careers.
The conversation explored the idea of “tokenmaxxing,” a term used to describe pushing AI systems to generate as much output as possible. Shevelenko suggested that approach may not hold up over time, implying that the most effective use of AI will depend on better design and deployment, not just higher usage.
The interview also addressed a question many workers are now asking: what human abilities are least likely to be replaced? While the source material does not list the three skills in detail, it frames the discussion around traits that remain valuable even as automation spreads.
That includes the ability to lead, to think strategically and to use AI for leverage rather than treating it as a full substitute for human judgment. Shevelenko’s comments suggest that the workers and leaders most likely to thrive in the AI era will be those who can combine technical tools with discernment, adaptability and initiative.
Perplexity’s broader message is clear. As AI systems become more capable, the competitive advantage may shift from producing more work to orchestrating it better. For companies and workers alike, the challenge is no longer just adapting to AI. It is learning how to work with it in ways that preserve human value.