Product management is not chiefly the production of specifications. It is the work of helping a team understand who a product is for, what it helps them do, and why that matters in their lives. In his September 14, 2026 essay for Andreessen Horowitz, partner Josh Elman argues that this shared story matters even more now that AI makes prototypes faster and cheaper to create.
Why a product story matters more than a specification
Asked what artifact a product manager produces, Elman recalls Reid Hoffman posing the question and describes a familiar answer: a specification. Specs can be useful, but they explain a system and its requirements. They do not necessarily give a team a common understanding of the people who will use it or why it should matter to them.
As an Amazon Associate I earn from qualifying purchases.
Elman’s alternative is a clear, repeatable account of the product’s users and purpose. As he puts it, “Product management is about telling the story of people who are going to use the product, and why it will matter in their lives.” A specification can support that work; it is not a substitute for it. A lengthy document—even the 120-page spec Elman recalls—cannot by itself ensure that a team is building something coherent or valuable.
This is not an argument to abandon requirements or planning. It is a way to keep them in proportion: use documentation to coordinate execution, while using the product story to guide decisions about what belongs in the product.
#1 Best Overall
How AI changes the path from idea to product
Elman describes a shift from moving directly from an idea into specification, scoping, and construction toward using an AI-assisted prototype to explore the idea first. A team can interact with a rough version, learn what feels useful or confusing, and then design and build the real product.
| Approach | What it emphasizes | What it can miss |
|---|---|---|
| Specification-first | Defining requirements and scope before building | How the idea actually feels to use, until later in the process |
| Prototype-first learning | Making an idea tangible early, then playing with it and learning | A demo’s appeal can be mistaken for product readiness |
Prototyping changes the cost and timing of learning; it does not remove the need for judgment. A prototype is not automatically a finished product. The team still has to decide whether the idea fits, design it for real use, implement it reliably, and learn from what happens after shipping.
Rank #2
- Physical Condition: No Defects
- Great one for reading
- It's a great choice for a book person
That distinction reframes a common product debate. “Can this fit in the schedule?” is a resourcing question. “Does this fit in the product?” is a question about coherence and impact. When ideas are cheap to demonstrate, the scarce work may be deciding which ones deserve to become part of the experience. More features do not necessarily make a better product.
Define the product through purpose, actions, and cycle
Elman’s practical framework for product vision has three parts. Together, they connect the promise a team makes to the behavior it hopes users will adopt.
Rank #3
- Purpose: Why would someone choose to bring this product into their life?
- Core actions: What does that person actually do with it?
- Cycle: How often should each core action happen?
The cycle matters because products serve different rhythms. A product intended for daily use should be judged differently from one whose useful action naturally happens less often. Teams need to know what meaningful use looks like for their product before choosing measures of success.
Ask whether people really use the product
Elman’s reader-facing question is direct: “Are people really using your product?” Counts such as signups, waitlists, revenue, app-store rank, or raw traffic can describe attention or business outcomes, but they do not, on their own, show whether users return voluntarily and perform the actions the product exists to support.
Rank #4
For an AI product, Elman recommends reading user transcripts. Look for what people expect, where they struggle, and when they rephrase a request. AI can help surface recurring patterns, but a product manager still needs to interpret them and decide what they mean for the product. The goal is not to treat every interaction as a feature request; it is to understand whether the product’s purpose, interface, and actual use line up.
Onboarding is where the product explains itself
Onboarding is a crucial chance to tell the product’s story because new users are curious and paying attention. Some arrive already eager and informed; others are casual visitors, with many people in between. Designing only for people who already understand the product leaves that larger curious middle without a clear route to value.
Best Value
Elman’s approach is to make onboarding teach, not merely collect information or move users through screens:
- Repeat the core message. Explain why the product exists and what it helps the user accomplish.
- Explain requests for information. Make clear why the product asks for particular details rather than treating data entry as self-evident.
- Break the product into concepts. Introduce the important ideas in manageable pieces.
- Give each concept a clear action. Let the user do something that demonstrates how the idea works.
For an AI product, a blank prompt box may leave newcomers guessing what the system can do. Show useful capabilities and help people reach a relevant case, ideally using their own data. Then judge onboarding by what users do afterward: whether they return and perform core actions, not simply whether they completed the introductory flow.
What Elman says Twitter’s Learn Flow changed
Elman recalls that early Twitter attracted curious people who did not yet understand what Twitter was or what to do there. He says the team rebuilt onboarding as a Learn Flow that introduced tweets, following, and timelines step by step. In his account, that work moved retention more than anything else the team shipped that year. This is Elman’s recollection of the experience, not an independently measured study presented in the essay.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallWhat product teams can take from the argument
- Write down the user, purpose, core actions, and expected cycle before treating a feature idea as a product decision.
- Use prototypes to learn earlier, while keeping a separate decision about whether the idea fits the product and is ready for production.
- Look beyond attention and acquisition counts to voluntary return and completion of the actions that deliver the product’s value.
- Make onboarding teach users how to get value, then assess it by subsequent use rather than flow completion alone.
- For AI experiences, use transcripts to identify expectations and friction, but leave interpretation and product judgment with the team.
Elman’s essay does not offer an external benchmark or study establishing a universal measure of product success. Its case is a product-management argument: a shared story helps teams decide what to build, how to explain it, and what evidence of real use to look for.
Read Josh Elman’s essay, “Product Management is Still All About Telling Stories,” at Andreessen Horowitz. Elman is an Andreessen Horowitz partner focused on consumer technology and AI, according to his official biography.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
