Lenny's Podcast: Product | Career | Growth · Lenny Rachitsky

How 80,000 companies build with AI: products as organisms, the death of org charts, and why agents will outnumber employees by 2026 | Asha Sharma (CVP of AI Platform at Microsoft)

·57 min·5 clips

Asha Sharma reveals how AI is turning products into living organisms that think and learn, becoming the new IP for companies.

This episode features Asha Sharma, Chief Vice President of Product for Microsoft's AI Platform, discussing the future of AI product development and organizational change. Host Lenny Rachitsky explores Sharma's unique perspective from her roles at Microsoft, Instacart, and Meta. Sharma argues the industry is shifting from "product as artifact" to "product as organism," where AI models continuously learn and adapt. She explains that with models now capable of tool calling and taking action, products become living systems that improve with more interactions. Sharma cites a study by Nathan Lambert showing that once a model hits 30 billion parameters, the capital expenditure for data becomes extremely powerful, making post-training more economical than massive pre-training runs. She describes how Microsoft's AI product "Dragon" for physicians saw acceptance rates jump from 30-60% to 83% after fine-tuning with 600,000 expert-annotated patient-physician interactions. Sharma observes that successful companies embrace AI fluency, apply AI to existing processes for measurable impact, and then use it to inflect growth. A key insight is the rise of the "polymath" or full-stack builder, as the rapid pace of AI innovation makes traditional, multi-layered organizational processes insufficient. She predicts a move from graphical user interfaces (GUIs) to code-native interfaces, as text streams interact more naturally with LLMs and enable greater composability for agents. Sharma introduces the concept of an "agentic society," where the marginal cost of good output approaches zero, driving exponential demand for productivity scaled by agents. She forecasts that organizational charts will evolve into "work charts" focused on tasks and throughput, potentially requiring fewer management layers. Sharma reveals that over 15,000 customers are building agents on Microsoft's Azure service, with millions of agents running in the cloud. For strategic planning in this fast-moving environment, her team operates in "seasons," like the current "advent of agents," supported by loose quarterly OKRs and squads with 4-6 week goals. She believes a company's new intellectual property is the continuous "loop" of ingesting data, digesting reward models, and generating outcomes, not the static product artifact. Sharma shares that she uses AI agents to summarize live site incidents and generate creative prototypes, noting AI is now embedded in nearly every workflow. The tone is conversational and forward-looking, blending high-level strategic vision with specific examples and data points. This episode is ideal for product leaders, engineers, and strategists interested in the practical and philosophical implications of AI on business and work. Listeners seeking a debate on AI ethics or a highly technical deep dive into model architecture might find it less aligned with their interests.

As heard by us

When the org chart becomes the work chart.

Asha Sharma frames AI as moving from feature to structure, with an agentic society where the org chart starts to look more like a work chart. The broader point is that product is beginning to feel less like a finished artifact and more like an organism: once models are capable…

Read the full review in PlayNext →

Why you'd press play

If you want to hear org charts get mildly threatened by agents, start here.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Lenny's Podcast: Product | Career | Growth