How to get your whole team excited about AI (and actually using it) | Brian Greenbaum (product designer at Pendo)

Brian describes his breakthrough moment using Cursor to build a music player app while on paternity leave, which led him to send a bold message to leadership proposing an AI upskilling initiative for Pendo's entire product organization. This inception moment demonstrates how personal experimentation can catalyze organizational change.

• •Built a working QR code-based music player prototype in hours using Cursor despite limited coding experience
• •Immediately recognized the potential for creating interactive, data-driven prototypes for product design work
• •Sent unsolicited message to CPO and leadership during paternity leave proposing cross-functional AI training initiative
• •Positioned the initiative with two goals: improve team productivity and establish Pendo as thought leader in AI-powered product development
• •Leadership responded enthusiastically, requesting immediate all-hands presentation

Brian outlines his structured approach to AI adoption using biweekly synchronous sessions combined with a public Slack channel for asynchronous learning. The strategy addresses the common excuse of 'not having time' by creating dedicated space in calendars while fostering a culture of radical peer-to-peer sharing.

• •Launched biweekly 'Product AI' sessions to create mandatory time for AI learning and experimentation
• •Created public Slack channel with 200+ members (beyond just product org) for continuous knowledge sharing
• •Emphasized 'radical many-to-many sharing' to combat information hoarding and secret AI usage
• •Made sessions interactive with hands-on exercises, not just presentations
• •Focused on building public experimentation culture to accelerate team-wide learning

The first Product AI session featured a hands-on exercise where all participants built to-do list apps using Bolt.new, then experimented with creative modifications. This approach helped teams experience AI's capabilities firsthand, understand its variability, and reignite their ability to think beyond MVP constraints.

• •Had entire team simultaneously build same to-do app with identical prompts, revealing AI's non-deterministic nature
• •Encouraged 'going wild' with modifications (retro themes, dark mode, Myspace styling) to make learning fun and expand creative thinking
• •Demonstrated that AI enables designers/PMs to escape 'MVP mode' and build more ambitious, polished products
• •Showed real-time problem-solving when some builds failed, normalizing iteration and debugging
• •Created immediate shared experience that sparked ongoing Slack channel discussions

The Slack channel became a hub for spontaneous innovation, with team members sharing experiments like Midjourney-generated animated characters for product interfaces. This demonstrates how async sharing enables discovery of use cases that wouldn't emerge from top-down planning.

• •Designer created animated waving characters using Midjourney for intro screens in just a few iterations
• •Channel enabled discovery of creative applications (animations, illustrations) that would previously require professional resources
• •Peer sharing created 'wow moments' that inspired others to experiment
• •Async format allowed people to contribute at their own pace without meeting fatigue
• •Public channel attracted interest from other departments beyond product organization

Brian joined a cross-functional team responsible for a company-wide AI OKR, implementing sentiment surveys to measure five key dimensions: sentiment, awareness, proficiency, policy understanding, and tool availability. The biggest improvements came from clarifying usage policies and available tools.

• •Conducted baseline and end-of-quarter surveys measuring sentiment, awareness, proficiency, policy knowledge, and tool availability
• •Largest improvements were in policy awareness and tool availability (not just sentiment)
• •Survey revealed widespread confusion about what AI tools were approved and what data could be shared
• •Positive sentiment about AI's impact on employees increased significantly over one quarter
• •Measurement approach provided concrete evidence of transformation success to leadership

Pendo created a comprehensive Confluence document listing all approved AI tools, their security status, permitted data types, and request processes. This 'golden path' eliminated shadow IT, accelerated experimentation, and required close collaboration with security, legal, and IT teams.

• •Built alphabetized table of approved AI tools with security status, data permissions, and access instructions
• •Worked closely with security, IT, legal, and finance to enable rapid tool approval (typically one week)
• •Eliminated confusion about shadow IT and personal account usage for work
• •Created fast path for requesting new tools to experiment with emerging technologies
• •Documentation became single source of truth, reducing repeated questions and enabling peer-to-peer support

Brian built a Model Context Protocol (MCP) server for Pendo's analytics platform during evenings with his newborn, enabling AI to query product data and generate dashboards. This prototype, built without deep technical knowledge using Cursor, directly influenced Pendo's product roadmap and agent strategy.

• •Built functional MCP server using Cursor and public APIs despite not fully understanding the underlying technology
• •Created demo showing Claude querying Pendo data and generating interactive dashboards from natural language
• •Prototype solved real problem: automating time-consuming data investigation when metrics dip
• •Demo changed CTO's perspective, leading to productionization efforts and roadmap acceleration
• •Demonstrates power of prototyping to influence product direction - code quality less important than demonstrating value

Leading AI transformation initiatives creates unique career opportunities for individual contributors to demonstrate cross-functional impact and thought leadership. Brian emphasizes this is 'promo-making work' that opens doors internally and externally, regardless of current level.

• •AI transformation leadership is rare opportunity for ICs to show broad organizational impact
• •Initiative opened doors to speaking opportunities, cross-functional projects, and executive visibility
• •Being first to raise hand for AI leadership positions you as internal thought leader
• •Requires some nights/weekends investment but aligns with personal interest in technology
• •Creates platform for external thought leadership (podcasts, conferences, LinkedIn presence)

Brian argues that product managers and designers must develop technical literacy (understanding LLMs, agents, basic HTML/CSS) to effectively apply AI to customer problems. One Product AI session specifically taught HTML/CSS to designers, unlocking their ability to use AI coding tools.

• •PMs and designers need to understand how AI technology works (LLMs, agents, APIs) without being ML engineers
• •Ran intro to HTML/CSS session for designers to unlock AI coding tool capabilities
• •Technical understanding enables creative problem-solving by connecting technology capabilities to customer needs
• •Analogous to architects understanding plumbing and electrical systems to design functional buildings
• •Engineers should similarly develop business and customer empathy to maximize AI application