What AI means for your product strategy | Paul Adams (CPO of Intercom)

Episode Summary

Executive Summary: Paul Adams, Intercom’s chief product officer, shares lessons from career-defining failures and explains why AI is a once-in-a-generation shift that product teams must treat strategically. He argues companies should map core user jobs against AI capabilities, distinguish replacement from augmentation, and avoid bolting AI on as an afterthought. He also offers practical product frameworks on pricing, differentiation, and job-to-be-done thinking.

Main Topics: Career failure and resilience (Priority: 5/5): Paul recounts major public failures, including freezing on stage at Cannes and leaving a Google social project midstream. He frames these as formative experiences that taught him adaptability, humility, and the ability to move forward after embarrassment or setback. How to think about AI in product strategy (Priority: 5/5): He argues AI is a major societal and product shift, comparable to the internet or mobile. Teams should start from the product’s core job, then assess whether AI can replace, augment, or partially help that job, rather than beginning with the technology itself. Intercom’s AI pivot and Fin (Priority: 5/5): Intercom changed strategy after ChatGPT’s launch, focusing on customer support as a prime AI use case. Its AI chatbot Fin now handles customer questions directly and also assists human agents, illustrating both replacement and augmentation modes. Organizational change, team structure, and capabilities (Priority: 4/5): Paul says AI should not be isolated in a siloed team. Instead, companies need strong ML specialists plus broad product-team AI literacy, with embedded collaboration across PM, design, engineering, and support workflows. Product frameworks for decision-making (Priority: 4/5): He shares several simple frameworks: before/after moments, swinging the pendulum, product-market-story fit, differentiation vs. table stakes, and jobs to be done. All emphasize clarity, practicality, and avoiding overcomplication. Pricing and roadmap discipline (Priority: 3/5): Paul explains how Intercom struggled with pricing complexity and overcorrected in various directions. His key lesson is to keep pricing simple, align it to value carefully, and avoid layering too many tiers and add-ons. Personal operating principles (Priority: 3/5): He closes with advice on focus, skepticism, and kindness: only work on what matters most, don’t worry about what you can’t control, read opposing views, and be nice to people because you rarely know what they’re dealing with.

Key Arguments: Failure is essential for growth; public or strategic failures can become durable learning experiences if teams adapt quickly. AI should be evaluated from the bottom up: identify the product’s core job, then ask whether AI can do it fully, partially, or not yet. Many products in B2B SaaS and media-adjacent workflows are exposed to AI because AI can write, summarize, search, reason, act, and process multimodal inputs. Intercom’s decisive move into AI was justified by the customer support use case, where AI can already resolve a meaningful share of requests and reshape the org. AI adoption requires broad organizational literacy, not just a specialist AI team; otherwise AI becomes a bolt-on instead of a product foundation. Simple frameworks often work better than academic complexity because they keep teams focused on customer problems and decision quality. Pricing should be simple and understandable; too many tiers, add-ons, and exceptions make bills confusing and undermine value communication. Jobs to be done and the four forces are useful because they center teams on customer motivation, habits, anxieties, and switching behavior. Strong leadership buy-in helps teams navigate ambiguity, but conviction should be balanced by skepticism and alternative viewpoints. Being nice and focusing on controllable priorities are practical life and management heuristics that reduce unnecessary conflict and distraction.

Data Points: Years as Intercom CPO: Over 10 years - Paul has held the chief product officer role at Intercom for more than a decade. Google tenure: 4 years - He worked at Google for four years before moving to Facebook. Facebook tenure: About 2.5 years - He spent roughly two and a half years at Facebook. Cannes keynote freeze duration: 3–4 minutes into the talk - Paul froze onstage early in a keynote at Cannes before walking off and returning. Intercom company size at recruitment pitch: 10-person company - Owen and Des pitched Paul to join Intercom when it was very small. ChatGPT launch date referenced: November 29 (year implied: 2022) - Paul repeatedly marks ChatGPT’s launch as the moment that changed his AI thinking. Team workflow ratio: 1 PM to 5 engineers - Paul mentions Intercom’s typical product team staffing ratio when discussing AI’s impact on engineering roles. Support deflection/automation rate: 50%–70% of inbound questions - Some Intercom customers reportedly resolve half to nearly three-quarters of support queries with Fin. Work hours estimate: About 50 hours/week - Paul says he avoids working crazy hours and believes beyond this decision quality declines. Radiologist training: 7 years - He cites the length of training to be a radiologist as a comparison for AI’s capability in x-ray analysis. Product framework: Horizon 1 / 2 / 3 - He references the Horizons framework in a discussion about AI roadmap uncertainty. Support market size: Decades of table stakes - He describes customer support as a mature category with many established incumbents.

Pivotal Quotes: "This is a like meteor coming towards you. This is going to radically transform society." — Paul Adams: His framing of AI as an unprecedented technology shift that product leaders must take seriously. "At Facebook, you can design the product, but at Intercom, you can design the company." — Owen (as quoted by Paul Adams): The line that convinced Paul to join a small Intercom and helped shape his tolerance for experimentation and failure. "Go back to basics and then ask, can AI do that?" — Paul Adams: His central recommendation for evaluating how AI should affect a product strategy.

Implications: Product teams should rethink core workflows through an AI lens now, not later. Winners will combine clear use-case mapping, simple pricing and positioning, and fast learning loops while avoiding AI theater and organizational silos.