Product strategy questions can sound smart and still waste a room. Shreyas Doshi’s X article argues that “should we go wide or deep?” is often the wrong opening move, especially for an AI startup suddenly facing larger incumbents. The better question is smaller and harder: which customer, which pain, which feature, and which reason to buy?
Table of Contents
The short version
- Doshi describes an AI startup founder whose team started debating whether to widen the product or deepen the current workflow after two large incumbents entered the space.
- His advice is to reject the binary because it pulls teams into abstract language before they have named the customer bet.
- The useful product strategy questions sit one level lower: what feature will resonate, who will buy because of it, and why will they stay?
- For founders and PMs, the article is a reminder that frameworks do not rescue weak customer understanding.
What happened
Doshi published an X article titled “Get to the Core of the Thing” after advising a founder running an AI startup. The founder’s team was anxious because two established companies had moved into the same market. Their proposed frame was familiar: should the product expand its surface area, or should the team sharpen what it already had?
Doshi’s answer was blunt. Drop the frame. In his view, a wide-versus-deep debate lets smart people sound strategic while avoiding the work that actually matters: naming the specific bet on a specific feature for a specific customer.
That distinction matters because many product meetings drift upward. Teams start with a real market threat, then jump into platform versus point solution, CAC versus LTV, horizontal versus vertical, or whatever analogy sounds good that week. Those phrases can be useful later. They are dangerous when they arrive before the team has done the customer work.
Why this is worth watching
The article lands because AI product teams are living through exactly this kind of pressure. When a bigger company enters a category, a smaller team can feel pushed to look broader, more platform-like, or more defensible on a slide. That instinct is understandable. It can also blur the only question a customer cares about: does this product solve my problem better than the thing I already use?
The piece is also useful for non-AI teams. “Wide or deep” is only one version of the trap. Founders can swap in “enterprise or SMB,” “workflow or infrastructure,” “self-serve or sales-led,” and still avoid the harder work. The language changes. The escape hatch is the same.
A better meeting starts with product strategy questions that make the team prove what it knows. Which buyer felt the pain last week? What did they try before? Which feature would change the buying conversation? What can the team ship quickly enough to learn from real use?
For more technology and AI briefs, the IT & AI archive tracks similar product and builder signals without turning every link into a trend forecast.
What the discussion is missing
There does not appear to be a Hacker News thread tied to this article. That is probably fine. Doshi’s post is less a news event than a product operating note, and the missing debate is the practical one inside teams: when is a framework helpful, and when is it camouflage?
The useful objection is that teams still need high-level strategy. A startup cannot interview its way out of every positioning decision. The point is not to ban strategy language. It is to use it after the team can state the customer bet in plain language.
The other open question is speed. Doshi says the team needs real differentiation and needs to build it quickly. That is the part many teams will agree with and still struggle to do. The test is whether the next roadmap meeting produces a feature bet someone can validate, or another hour of vocabulary.
The practical read
If your team is stuck in a wide-versus-deep debate, pause the labels and rewrite the agenda around product strategy questions.
Ask who the customer is in a way that points to a real person or account, not a segment name. Ask what that customer is doing today instead of using your product. Ask which feature would change the purchase or retention decision. Ask whether your team can build enough of that feature to learn before the market moves again.
If you cannot answer those questions, choosing “wide” or “deep” will not fix the product. It will only make the uncertainty sound organized. If you can answer them, the shape of the product usually becomes less mysterious. You go wider where the customer bet requires reach, and deeper where the buying reason requires depth.
Product strategy questions to ask first
Use these product strategy questions before the roadmap turns into a framing contest:
- Which customer call, support ticket, renewal risk, or lost deal are we using as evidence?
- Which feature would make that customer buy, stay, expand, or switch?
- What do we believe competitors cannot copy quickly enough to erase the advantage?
- What can we ship in the next cycle that will make the answer clearer?
That is less glamorous than a strategy offsite. It is also harder to fake.