Key Moments
The last roadmap | Claire Vo
Lenny's PodcastKey Moments
The rapid advancement of AI has made building software effortless, but product leaders now face a crisis of conviction, struggling to identify truly valuable ideas amidst a sea of possibilities.
Key Insights
Product leaders now have more execution capability due to AI tools but are running out of good ideas, shifting the bottleneck from engineering capacity to conviction about what's worth building.
The traditional roadmap, effective when engineering was scarce, now creates three traps: task accumulation, feature parity, and abandonment, leading to perceived progress without real customer value.
Claire Vo argues for a shift from 'building' to 'proving,' emphasizing the need for strong convictions about desired outcomes, defined by evidence, rather than focusing on feature lists and arbitrary dates.
A 'zero roadmap' state, where every task is executable, renders traditional roadmaps and prioritization obsolete, necessitating a focus on ambitious goals and genuine customer validation.
The future of roadmapping requires embracing a 'race for ambition' over a 'race for speed,' encouraging large investments in ambitious projects and bold experiments, even if most fail.
Instead of promising specific features and dates, teams should communicate their core convictions and the evidence supporting them, distinguishing between 'bets' and 'promises' to manage customer and internal expectations.
The crisis of conviction in a hyper-executable world
Claire Vo opens by noting the dramatic shift in product management since the last summit. While the field isn't dead, its landscape has fundamentally changed. Previously, engineering capacity was the scarce resource, forcing product managers to act as gatekeepers, saying 'no' to countless ideas. Now, with AI-powered coding agents and advanced developer tools, the ability to build almost anything is unprecedented. Vo confesses her own predicament: she has run out of *good* ideas, not due to a lack of potential projects, but a lack of conviction that any of them are truly valuable or marketable. This presents a new, more profound challenge: moving from a scarcity of execution to a scarcity of belief in what should be built.
The obsolescence of traditional roadmaps
The traditional roadmap, designed for an era of limited engineering resources, has become detrimental. In such times, scarcity naturally filtered out bad ideas, as they simply couldn't be built. Now, with AI, all ideas can be executed, leading to three major pitfalls. First, the 'task accumulation trap': AI rapidly populates backlogs, creating the illusion of progress as tasks are completed, but not necessarily solving customer problems. Second, the 'parity trap': with competitors having similar access to information and AI, teams tend to build similar products, leading to a 'parody' of innovation where everyone is doing the same thing. Third, the 'abandonment trap': teams release features, encounter minor issues, and abandon them because they can always build something new, preventing learning and iteration. These traps accelerate teams towards mediocrity, focusing on output (shipping features) rather than outcomes (delivering value).
The 'zero roadmap' and the demise of prioritization
Vo introduces the concept of a 'zero roadmap,' not meaning a lack of a roadmap, but a state where every item on the to-do list is executable. When everything is feasible and effort is no longer a significant differentiator, traditional prioritization loses its meaning. The bottleneck shifts from 'can we build it?' to 'should we build it?'. This renders the roadmap, as a list of features with estimated impacts and dates, not just outdated but dangerous, especially when combined with AI's execution power. AI amplifies the consequences of poor evaluation, turning bad ideas into problems much faster.
Shifting focus from building to proving
The proposed solution is a transition from 'building' to 'proving.' This involves articulating core convictions about the desired future state, not just for the next quarter, but for the long term. Crucially, these convictions must be accompanied by clearly defined evidence that will validate them. Teams need to establish what evidence is required to confirm they are on the right track and, just as importantly, what evidence would signal the need to stop or pivot. This requires a robust 'factory' capable of rapid iteration and learning, allowing for swift interaction with reality and redirecting investments based on evidence.
Ambition over speed: the race for impact
Vo contrasts the recent 'gold rush' of rapid feature shipping with the upcoming 'race for ambition.' While speed was critical in the last 1-2 years to leverage AI and gather feedback, the next phase requires focusing on radical, ambitious changes and large-scale experiments. This means posing questions like: 'What audacious bets can we make?' and 'What experiments can we run in weeks that used to take years?'. The roadmap should reflect these large investments in ambitious projects, rather than just incremental feature velocity. The goal is to create systems that can surprise us, leading to genuinely better outcomes.
Embracing conviction with mutable features
The new approach demands strong convictions but mutable features. Teams need to be steadfast in their core beliefs (good conviction) while remaining flexible on the specific solutions (features). This requires distinguishing between 'bets' – hypotheses tested with evidence, which may be proven wrong – and 'promises' – features customers can rely on and build upon. Transparency about the level of conviction and commitment for each initiative is paramount. The ultimate aim is to build trust with customers, recognizing that while code is abundant, customer trust is rare. A successful roadmap will show clear progress toward convictions, without attachment to specific feature implementations.
The future roadmap: ambition, evidence, and honesty
Vo's final call to action is to create a 'last roadmap' – not a final plan, but a reimagining of what a roadmap should be. This means moving away from spreadsheets of features with guessed impacts and fixed dates. Instead, teams should focus on nurturing their core convictions and ambitions, defining what success looks like in 1-2 years. This involves making big decisions, accepting that most bold experiments will fail, and setting high standards for AI to generate superior ideas. The new roadmap will embrace ambiguity in details, focus on ambitious goals, and be honest about the level of commitment to each initiative, recognizing that customer trust is the most valuable currency.
Mentioned in This Episode
●Software & Apps
●Concepts
●People Referenced
Common Questions
The speaker argues that traditional roadmaps are becoming obsolete due to increased execution capabilities fueled by AI. They suggest shifting focus from feature lists and dates to core convictions and ambitious goals.
Topics
Mentioned in this video
More from Lenny's Podcast
View all 51 summaries
97 min90 minutes of unfiltered product advice from Snap and Discord’s product chief | Peter Sellis
83 minHow we built Grok Bot in a month | Roman Ugarte (SpaceXAI)
80 minWhy companies are becoming a series of loops | Anish Acharya (a16z)
82 minAI’s third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI’s product lead)
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free