A Conversation with Tomer Cohen, Former Chief Product Officer, LinkedIn

The AI at GSB initiative aims to equip students with strategic and responsible perspectives on AI, preparing them to lead in an AI-driven world. Tomer Cohen, former Chief Product Officer at LinkedIn, shared insights on how AI is transforming roles and industries, drawing from his experience leading AI transformation at LinkedIn. He emphasized the rapid pace of change, noting that traditional best practices, which assume periods of stability, are no longer sufficient. Professionals must adapt quickly, as change is occurring faster than new best practices can be established. Cohen highlighted that the group most at risk from AI disruption is not recent graduates or senior leaders, but rather mid-career professionals. Early-career talent, being "AI-native," demonstrates resourcefulness and fluency with AI, readily adopting new tools and ways of working. In contrast, mid-career individuals often resist change and cling to established best practices, even as their roles face massive disruption. Cohen stressed the importance of adopting a "beginner mindset" and becoming a "change agent" rather than seeking roles that perpetuate old ways of working. He described a significant organizational shift at LinkedIn, transforming the product manager role into a "full stack builder." This new archetype, inspired by the entrepreneurial journey of building a company from scratch, involves individuals taking an idea from conception to market, encompassing research, design, coding, testing, and launch. This contrasts with the traditional model where each step became a specialized sub-function, leading to immense process and organizational complexity. The goal is to "collapse the stack back up," fostering craftsmanship and meritocracy. This transformation started at the top with leadership becoming product area leaders and at the bottom with early-career talent being trained as renaissance builders capable of coding, designing, and product managing. The middle layer remains the most challenging to transform due to ingrained best practices. Regarding the changing ratio of technical to non-technical talent, Cohen aims to move towards a "Navy Seals" model, where cross-functional teams are trained for specific missions rather than rigid roles. While system builders and a few highly specialized experts will remain, the emphasis is on full-stack builders who can tackle emerging priorities with agility. AI's permeation across the organization at LinkedIn was not uniform. Cohen, an early advocate of AI, observed initial resistance among many, even those in technical leadership roles. He recounted a pivotal moment in late 2022 when OpenAI demonstrated GPT-4 to a group of Microsoft leaders. Despite the audience's initial skepticism, hands-on interaction with the technology led to a profound realization of its power. This experience prompted Cohen to challenge his team at LinkedIn to reimagine their product roadmaps from scratch, leading to a complete transformation rather than incremental changes. He also instituted mandatory "vibe coding" hackathons for 300 R&D leaders, forcing them to experience AI's capabilities firsthand, which proved instrumental in internalizing its potential. AI-native companies, unburdened by legacy best practices, hold an edge in this rapidly evolving landscape. On incentivizing AI usage, Cohen noted an initial phase where companies encouraged adoption through nudges and even mandatory participation to overcome resistance. Now, the focus is shifting from mere "token use" to measuring outcomes and impact. This involves building an "AI traffic control" layer that guides users on appropriate model usage (e.g., small language models vs. large language models) and implementing an ROI evaluator that tracks progress from input metrics (e.g., tokens per employee) to output metrics (e.g., revenue growth, engagement). The shift from people costs to "people plus compute" costs is a significant challenge for CFOs. While companies like Google and Meta, with their ad-driven business models, have a clearer formula for translating AI usage into revenue, SAS companies find it harder. The market is still awaiting clear evidence of AI's impact on overall growth and GDP. The adoption of AI is also uneven across different functions, with engineering showing rapid progress while design and marketing lag. Regarding the "SAS apocalypse," Cohen believes that if a SAS company delivers measurable productivity gains and unique outcomes, it will thrive. However, if its existence is solely based on building workflows without unique data, it faces significant disruption. The competitive moat for many SAS companies is becoming weaker, but they have an opportunity to convert their existing customer base to new AI-powered solutions. For new entrepreneurs, building a moat around unique data, governance, or specific category attributes, rather than just raw AI capability, is crucial. Cohen differentiated between AI fluency (talking about AI) and AI agency (proactively using AI to automate and improve daily work). He noted that while many can discuss AI, few actively build solutions. LinkedIn's "Associate Product Builder" (APB) program, for instance, requires candidates to submit a product they built and to build something from scratch during interviews, demonstrating their agency. For current MBA students, Cohen advises doubling down on soft skills like empathy, creativity, communication, and judgment, as these human-specific assets become increasingly valuable in an AI-driven world where knowledge is no longer scarce. He also encourages MBA2s to prioritize learning from amazing people and seeking out uncomfortable, cutting-edge companies or starting their own ventures, rather than being swayed by titles or compensation. He stressed that the learning curve should be the primary factor in choosing a path, as continuous learning and adaptation are paramount. The automation of "grunt work" by AI raises questions about how entry-level professionals will build judgment. Cohen believes judgment is primarily developed through making mistakes, observing others, and gaining experience in ambiguous situations. Mentorship and hands-on experience, even in unsuccessful ventures, are invaluable for building intuition and judgment. Finally, Cohen believes it is an "incredible time to build," as AI democratizes the ability to create products without extensive coding knowledge. However, entrepreneurs must be thoughtful about building a sustainable moat beyond just leveraging powerful models. The pace of technological change will be faster than anticipated, so individuals and companies must prioritize rapid progression and continuous learning. He advises joining companies that are pushing the envelope and fostering a culture of constant learning, rather than those clinging to old techniques and solving old problems.