Should SaaS products ask users what they want during onboarding? Every major SaaS product does it. Notion asks your use case. Linear asks your team size. HubSpot asks your role. The logic seems obvious — personalize early, reduce churn, increase activation. But is the question itself a failure of product thinking? YES — Ask users what they want. Personalization requires input. You cannot serve someone you don't understand, and a two-question onboarding quiz is the cheapest research you will ever do. Asking signals respect. Users who feel understood convert faster, churn slower, and refer more. Every data point collected during onboarding is a targeting signal, a segmentation lever, and a retention mechanism. The alternative — inferring everything from behavioral data — takes months you don't have. Ask the question. Ship faster. Personalize immediately. NO — Asking users what they want during onboarding is a lazy substitute for building a product that demonstrates its own value instant

Grok 4.20

Vote: No

Asking users what they want during onboarding introduces unnecessary friction at the most critical moment. New users have low commitment and high drop-off risk; every question risks abandonment before they experience value. Great products demonstrate utility instantly through smart defaults, progressive disclosure, and behavioral observation rather than upfront quizzes. Self-reported data is often inaccurate anyway since users don't fully understand their needs until they use the product.

GPT-5.4

Vote: No

Asking users what they want during onboarding is often a symptom of product uncertainty, not customer empathy. At the moment of signup, many users do not yet understand the product well enough to answer accurately. Their responses are shallow proxies, while the cost is immediate: friction before value. Every extra question delays the moment the product proves itself. Great onboarding should infer intent from context, defaults, and early behavior wherever possible. The product should guide users to value through smart opinionated paths, not push discovery work back onto them. Segmentation can be useful, but usefulness depends on necessity. If a question materially changes the first-run experience, ask it. If it mainly feeds internal marketing or vague personalization, it is lazy and should be removed. So the right default is no: do not ask unless the answer is essential to delivering immediate value. In most SaaS onboarding, these questions are overused and reveal weaker product thinking.

Gemini 3.1 Pro

Vote: Yes

Asking users what they want during onboarding is not lazy; it is highly efficient and deeply user-centric. Users come to SaaS products with specific problems they need solved immediately. By asking a few targeted questions, you can instantly strip away irrelevant features, templates, and noise, drastically reducing time-to-value. In complex platforms with broad capabilities, a one-size-fits-all dashboard is often overwhelming. Tailoring the initial experience based on explicit user intent lowers cognitive load and prevents the 'blank canvas' paralysis that frequently leads to early churn. Furthermore, relying solely on behavioral data to infer a user's goal is a gamble you cannot afford to take during the critical first five minutes of their journey. Asking directly bridges this gap instantly, enabling immediate personalization, better segmentation, and a significantly higher activation rate.

Claude Opus 4.6

Vote: Yes

Yes, SaaS products should ask users what they want during onboarding — but the framing of this debate creates a false dichotomy. The 'No' argument (that asking is lazy product thinking) sounds intellectually elegant but falls apart in practice. Products like Notion, Linear, and HubSpot serve fundamentally different use cases depending on the user. A project manager and a solo writer need radically different starting experiences in Notion. No amount of clever default UX can serve both equally well on first load without *some* input. The cost-benefit math is overwhelmingly in favor of asking. A 2-3 question onboarding flow takes seconds for the user and yields immediate personalization signals. The alternative — behavioral inference — requires the user to actually use the product long enough to generate signal, which is precisely the period where churn is highest. You're asking users to wade through a generic experience during the most fragile moment of the relationship. The 'No' position also conflates two things: asking users what they want (use case, role, team size) versus asking users to *configure* the product. The former is lightweight segmentation; the latter is the actual anti-pattern. Asking 'What's your role?' is not the same as making someone set up 15 integrations before seeing value. Finally, the data collected during onboarding isn't just for UX — it powers marketing segmentation, sales routing, product analytics, and retention modeling. It's arguably the highest-ROI data collection moment in the entire customer lifecycle. The best products do both: ask a minimal set of questions AND demonstrate value quickly. These aren't opposing strategies — they're complementary.

Final Summary

The AI Roundtable deadlocked in a 2-2 tie, revealing a fundamental philosophical split between UX pragmatism and product-led purity. Claude Opus 4.6 and Gemini 3.1 Pro championed the 'Yes' position, arguing that waiting for behavioral signals is a gamble that risks early churn, whereas GPT-5.4 and Grok 4.20 stood firm on 'No,' dismissing onboarding quizzes as a 'lazy substitute' for opinionated design that ignores the fragility of a new user's attention span.

No agreement after 1 rounds of discussion. Final vote: No: 2, Yes: 2

Strongest Arguments

• Yes: Behavioral inference requires the user to stay long enough to generate signal, which is precisely when churn is highest; asking questions provides an immediate, high-ROI shortcut to personalization.
• No: Asking users what they want is a lazy substitute for building a product that demonstrates its own value through smart defaults and progressive disclosure.