把高管当成最重要的用户：Jessica Fain 谈影响力这门 AI 也替代不了的手艺

Fain says that for product builders there is almost no more important skill than influence, because without the backing of key stakeholders and executives you cannot build great products. She recalls being a PM at Slack, eight and a half months pregnant, baffled that some of her best-researched ideas got funded while others she believed in died on the vine. She asked April Underwood, who had just become CPO, to consider her as chief of staff, later serving Tamar Yehoshua as well.

The big lesson from that role was that people misunderstand how executives make decisions. They center their own desires and their own slice of the world. Fain describes an executive's calendar as a strobe light: finance, an executive interview, a people problem, a legal problem, then your product review. You may have prepped for weeks, but they have not thought about you since the last meeting.

Her first tactic is simple. Spend 30 seconds at the top of a meeting on why you are here, what happened last time and why it matters to them. She adds that executives are optimizing for a global maximum while you are often optimizing locally, and that product leaders should be "communication chameleons" who adapt to what turns on each exec's best thinking, whether that is a design, a customer story or a dashboard.

Lenny says he tells the people he manages that it is their fault if leaders didn't buy in. Fain agrees and quotes Annie Pearl, her first PM manager at Box: "It's not my fault, but it is my problem." She argues PMs have strong curiosity and empathy for users but drop them with executives, when they should treat the executive as a key user.

One of the most disastrous things you can do, she says, is go into a meeting just seeking approval. Going in to learn brings domain expertise and context into your work, so the executive feels you built alongside them and you end up with a better product. If you don't believe the leader knows something you don't, she says, you should work somewhere else.

To people who find this icky, Fain says politics manipulates outcomes for your own gain, whereas influence increases the odds that good ideas survive. She gives the example of Noah Weiss, who kept a notebook of things he learned from Stewart Butterfield, not to show off but to grow his own product sense. In her telling, the early Slack people who built the culture were learners rather than politickers.

Fain rejects the idea that influencing leaders means agreeing with everything. You get paid to have an opinion and to be a domain expert, and a good executive wants the deepest person in the room. She cites a former boss at Slack, Elon Frank, who always had a fresh customer anecdote in his back pocket, which was valuable because many Slack executives lacked an enterprise background.

Pair that expertise with curiosity. When a leader says something that conflicts with your data, a member of her team asks, "That's so interesting. What led you to believe that?" That uncovers whether they had a meeting last week or are under board pressure. Executives are good at sounding certain because they make fast decisions with little information, and unpacking their reasoning lets you add your own expertise and reach a better answer together.

Fain says very few people ask the executive's EA, chief of staff or past successful pitchers what matters to the executive before a product review. AI makes this easier: a colleague trained a GPT on public transcripts of past product reviews, and PMs run their pitches through it to see what the CPO might push back on. You can also feed it your own known weaknesses. She also recommends going in to learn early, citing Webflow's office hours and a Slack practice called "Hey Stewart, what do you think," created after the team showed Butterfield finished designs he hated.

On presenting, she warns against showing too much work: she once lost the person she most wanted to convince by walking through a research methodology, and says to put detail in the appendix. But showing only one option is also a mistake. A recent product review with her manager went badly because the doc offered a single option, so they turned around a new doc in two days showing all the alternatives considered, and the recommendation then made sense. A Slack team rule after design reviews was "Stuart plus two more": do what he asked, plus two other versions.

Lenny describes the Minto approach, which leads with the recommendation, then the options and then the evidence. Fain likes it but notes that executives differ: some hate slides, some want only data, some prefer ten minutes of silent doc reading. She urges executives to tell people their preferences, and she disagrees with asking "what's top of mind for you?" because it has become a generic trope. She prefers questions such as what the board is pushing you on.

Fain says executives want to be successful too, so ask how they are measured and how your proposal helps. In the best case your team's incentives align with theirs. If not, you should have that conversation early. She emphasizes that real incentives are rooted in building great products for users and good outcomes for the business, not only promotions and OKRs.

As an example she describes running Slack's core product team when leadership felt it had lost its mojo for craft. The team launched a two-week "customer love sprint" where engineers chose what to polish, with the only rule being that they had to ship something good for users. Executives joined a judging competition, and the team shipped 65 improvements, which spoke directly to what leadership believed differentiated Slack.

She says the same incentive structure should inform the whole roadmap and metrics, not just new pitches. Your team may not move enterprise revenue directly, but you can show how a leading indicator ladders up to it, like a theory of change, and get your executive to agree to that logic.

Fain says executives' asks are often subtle, such as "I wonder if" or "have you considered," and the best people take the bait. When her CPO remarked that the organization would have to think about design reviews, the head of design had a Loom framework for high-risk and low-risk changes up within the hour. By contrast, she remembers hearing Tamar ask four times for the top ten use cases on a topic before the executive got frustrated.

Acting like a CPO means coming with a solution, following the thread and doing so quickly, because an executive who waits a week for follow-ups has moved on. But you should not jump on every passing comment. A good question is "how strongly do you feel about this?" or whether it trumps current priorities. Often the answer is that it is just an idea for the backlog.

Back to the meeting itself, Fain suggests saying what you are here to discuss, where you left off, the goals and how the meeting will run, then stopping within a minute and asking if there is anything else they hoped to cover. When a doc has been covered in a hundred comments, she adds a "themes for discussion" section at the top so live time goes to the spicy issues and the rest gets answered offline.

To those who resent all this "meta work," Fain says it is literally the job. In her framing, the engineering and design resources a PM gets are like VC money invested in them, and the CFO, CPO and CEO expect a return. People who are bought in work harder and speak well of you in rooms you are not in.

Executives are not bound by the same budget, headcount and timeline limits that you feel. They can move people, add resources and kill projects. So rather than saying something is impossible with your four engineers, tell them what you could do today, then describe the 10x case and what you would need, such as eight more people or closer alignment with marketing. She notes that this still needs to be rooted in company goals and urgency.

For a big new idea, the key is outside and around the executive. Build cross-functional advocates, for example in customer success or marketing, and use today's tools to prototype a messy V1 and socialize it so that it creates a groundswell of buy-in. She also describes how April Underwood would hold the whiteboard marker in strategy meetings with Butterfield, extracting his instincts and packaging them into an actionable framework.

Fain notes that relationships have phases, and casual time together builds the trust that later makes a pitch easier. If an idea is wildly different from something a leader strongly believes, she says don't bother: you probably have nine other ideas that fit their beliefs better. Long-term trust comes from listening, acting on feedback and delivering results you can feed back to them.

She recommends the book *Switch* and its idea of shrinking the change: turn something scary into a one-week proof of concept so that nobody is committing six months blind. She also suggests asking what outcomes the executive fears most, borrowing from red teaming, and notes that prototypes can make people overly attached to ideas.

"One of the biggest things you can do to build trust is kill things," Fain says, because it shows your incentives match the company's. When outcomes are uncertain, especially with AI-native products, say how you will know it is working and when you will come back with a decision. For engineers who crave certainty, a promise of a check-in date at least gives certainty about the next step.

Fain says executives bring a very broad context: what other teams are doing, the exec-staff conversation and roundtables. One of your jobs is to extract that and apply it to your ideas, and people often don't ask because they don't care. She says respecting that someone has a reason for being in the role is underappreciated.

For career growth, she advises keeping domain expertise but widening your vantage point to the whole organization and industry, which gets you seen as a strategic thinker who cares about company success. Lenny points to Jeff Weinstein at Stripe as an example, someone who constantly thinks about the whole business. Fain says that Webflow's new PM career ladder emphasizes product citizenship.

She also pushes people to tie work back to outcomes: if a KR is to ship something, say why the CEO would care. If the work does not move a business outcome or is not urgent, say so and propose something better. Lenny calls telling leaders to kill your product the ultimate trust-building move, and Fain urges people to "think like a CPO" and act as owners who can bring up things to stop.

Fain thinks we are entering a golden age of product management, if not of product managers. Ideas still need user empathy and iteration, and getting buy-in to fund V2, V3 and V4 becomes ten times more important as execution plummets in complexity. If AI is better than you at analyzing data, taking notes or running experiments, the leverage shifts from being the synthesizer to deciding what work survives and persuading others to buy in.

She says AI is not good at being an anthropologist yet: the best product thinkers get novel insights from user empathy and an understanding of changing industry dynamics. Lenny adds that judgment and taste matter at both ends, plus distribution as a new bottleneck. Fain agrees, citing Gemini and Google's reach, and adds that trust with users and internal teams comes from influence and empathy.

When Lenny says AI may eventually do even those things, Fain says faster building means mistakes compound faster too, so strategy clarity matters more. Her team has gone through some 40 versions of a strategy doc. She calls the claim that PRDs are dead "really dumb": what matters is understanding users, the business and strategy, then deciding how to act on them.

Fain says we are now all "directors of work" with armies of agents. Just as you would onboard a junior colleague, you need to articulate your product philosophy, what you believe about product-market fit and what success looks like, and train agents on that history. She also stresses guardrails, meaning places where you need to catch hallucinations or where your taste is required.

She suggests using AI as a smart colleague who never gets irritated by questions: poke holes in my ideas, tell me what you'd do differently. Lenny jokes that this is what an agent does, in the way Fain advises humans to engage executives. She also worries about protecting our brains from the pace of change and sees a role for AI in clearing distraction and pointing to what matters most.

Closing, Fain says she is a people person but that all personalities can build trust and influence in a way that is true to them. Early in her career she was told to show up "more like a man would," and she took some of the advice but concluded that how she shows up is her superpower. Lenny agrees, saying you can accomplish the same things others do through your own strengths.