Uber engineers have developed an internal chatbot that imitates CEO Dara Khosrowshahi’s feedback style, allowing teams to rehearse presentations and pressure-test proposals before taking them to the real executive. The tool, informally known as “Dara AI,” highlights how deeply AI has become embedded in the company’s internal workflows.
Inside Uber’s CEO simulation tool
Khosrowshahi revealed the existence of Dara AI during an appearance on the podcast The Diary of a CEO. According to him, some teams now run their pitch decks and strategy proposals through the bot as a practice round before executive reviews.
The goal is straightforward. By the time a proposal reaches the CEO’s desk, it has already been refined through multiple iterations. The chatbot provides a lower-stakes environment where teams can sharpen their narrative, anticipate tough questions, and tighten their slides.
While Uber has not publicly detailed the training pipeline, reports suggest the system likely draws on Khosrowshahi’s public interviews, past communications, and meeting patterns to approximate his questioning style.
AI now deeply embedded across Uber engineering
The Dara AI experiment reflects a broader shift inside Uber’s engineering organization. Khosrowshahi said roughly 90 percent of the company’s software engineers now use AI tools in their daily work, with about 30 percent classified as heavy users who are actively rethinking system architecture around AI capabilities.
He described Uber less as a traditional ride-hailing company and more as an enormous software platform, where engineers continuously build and refine the underlying infrastructure. In that context, AI has become part of the core development fabric rather than an optional productivity layer.
Uber already operates an internal AI ecosystem that includes tools such as Genie, which answers policy and engineering questions, and Picasso, a workflow platform that embeds conversational agents into developer and support environments.
Bottom-up AI adoption takes center stage
Khosrowshahi framed Dara AI as an example of organic, bottom-up innovation rather than a top-down mandate. Teams built the tool themselves to solve a practical communication problem: preparing more effectively for high-stakes executive reviews.
The approach signals a broader trend in large tech companies, where employees are increasingly creating bespoke internal agents tailored to specific workflows, personalities, or decision processes. Instead of relying solely on general-purpose assistants, organizations are moving toward highly contextual AI tools.
What this signals for the future of corporate workflows
The emergence of leader-simulating agents like Dara AI underscores how quickly human feedback loops inside large companies are being augmented by AI. Key meetings, architecture proposals, and strategic discussions are increasingly filtered through machine-generated critique before reaching senior leadership.
Khosrowshahi described the productivity impact of AI at Uber as unlike anything he has previously witnessed. At the same time, he joked that building a digital version of the CEO might raise uncomfortable questions about job security, even at the top.
For now, Dara AI appears to function less as a replacement and more as a rehearsal partner. But its existence offers a glimpse into how AI is beginning to reshape not just how software is written, but how corporate decision-making itself is prepared and refined.
Post Comment
Be the first to post comment!
