Lean Smith: AI, Workflows, and Building Businesses

Summary

Lean Smith, director at Google Cloud, emphasizes practical application of AI for developer experience and agentic AI. He highlights his journey from early entrepreneurial ventures like selling candy to roles at Microsoft, Amazon, and Google, focusing on computer vision and AI perception. Smith stresses the importance of sales and demand generation, even in technical fields, and advocates for a shift towards more natural communication methods like voice agents over typing. He shares his experience with tools like Whisper for transcribing speech to text, making LLM interaction more fluid. Smith oversees Google Cloud services and product growth, focusing on developer tools and growth engineering, which he defines as selling products without a dedicated sales team, by understanding behavioral economics. He details his entrepreneurial philosophy, valuing the zero-to-one journey and supporting smaller players through technology. Smith views Google as a collaborative 'college campus' of smart individuals. He discusses the 'barbell effect' in AI, where the middle is squeezed out, leaving advantages for small and large entities. He elaborates on his approach to business, prioritizing recurring revenue, B2B models, and local operations, using a pre-construction business as an example of streamlining inefficiencies. Smith advocates for workflow design and systems thinking as crucial for AI implementation, using a 'parcel grader' agent as a case study for building specialized AI skills by first documenting the process in chat. He notes that seventy-five percent of code at Google is now AI-assisted, a significant increase from fifty percent the previous year. Smith stresses that successful AI implementation requires treating agents like efficient workers with clear Standard Operating Procedures, not high-paid knowledge workers, and warns that large companies may struggle with the necessary organizational changes. He advises against viewing AI tools as mere assistants, advocating instead for system orchestration and management, similar to managing people. Smith outlines a 'growth, delivery, and support' framework for businesses, transcending traditional departmental silos. He also addresses security concerns, recommending isolating agents, providing them with dedicated email and phone numbers, and emphasizing a culture of trust and transparency for effective AI adoption, especially within enterprises. He concludes by describing the future of work as 'always changing' and requiring continuous upskilling and resilience, urging individuals to invest in systems thinking and process understanding over simply learning new tools.