Why Before What: The Question Startups Forgot How to Ask

A child’s tendency to ask “why?” – “Why is the sky blue?” “Why does that kid look different?” “Why do I have to go to school?” – is the most underrated skill in the product-building space. Somewhere between childhood and startup founding, most builders stop asking it.

Questions that unveil the underlying causes of current phenomena – why is this the way it is? – are being replaced with what solution can I apply? Or worse, how quickly will this generate revenue? Revenue matters. But revenue is a result of building something people actually need, and you can't get there without asking why first.

A couple of weeks ago, I asked the Director of Suffolk University Law School’s Legal Innovation and Technology Lab about his methodology for making legal tech. Before speaking at length about their academic processes and nuances, he set out an essential imperative: question why current paradigms (ex: systems, processes, documents, etc.) exist before intervening.

That stuck with me. Since our conversation, I’ve been reflecting on research in legal design, where I am using design thinking to augment immigrants' access to legal resources. I realized that I’ve never fully questioned the causes of issues in immigration law. I would jump to identifying defects and critiquing case processes. Raging on attributes I presumed to be unjust.

But taking a pause to ask why makes the problem much clearer. Understanding the root causes of a problem allows you to build a solution that attacks the problem, not just the problem’s external indicators. What if something is the way it is because of an important factor? If you know that insight, you can build something even more relevant accordingly.

I’ve been seeing this pattern everywhere since.

I recently started a human-centered design project in collaboration with the City of Anaheim and the Salvation Army, exploring unsheltered homelessness. Human-centered design is exactly what it sounds like: research first, build second, and always grounded in the experiences of the people you're designing for. My team is focused on the ACCESS Court: a network of legal and social service professionals that create case-by-case strategies to redirect homeless individuals away from criminal justice proceedings and into rehabilitation, treatment, or community service.

We recently interviewed a prosecutor within ACCESS – someone whose entire approach is to ask “why does homelessness exist?” He put it simply:

“If you think about it from a medical perspective, it’s great that we have tools to combat cancer and serious disease, but wouldn’t you rather do preventative things to never get cancer or those diseases in the first place?”

These two interventions, ACCESS and preventative medicine, are built on understanding root causes (asking why?) instead of throwing the homeless in jail or telling diabetics to simply take pills. The prosecutor knows he’s preventing crime by housing people, getting them medicated, and getting them clean. He’s familiar with adverse childhood experiences and criminogenic factors. He can empathize.

That empathy allows him to make more informed decisions – not just look for the hammer.

Question. Empathize. Understand. Make informed decisions.

Startup founders should use the same process.

The AI Wrapper Mess

Because lately, the only empathy I’m seeing in the startup space is founder’s dedication to their AI wrapper and their heartfelt Claude credits to maintain it. Because without maintenance, then their dream of a fast-generating revenue product would succumb to an ultimate failure. Because the dream isn’t to revolve their product around their target user’s needs, of course.

Take the meeting notes space. Fellow.ai is the latest addition to an already-crowded category – Fellow, Otter.ai, Fireflies, Grain, tl;dv, Fathom, Avoma, etc… Most of them are doing the same thing: transcribe meeting, summarize, done. None of them are asking why meetings are actually inefficient.

I can explain from experience. I recently tried accessing a recording and its transcript on Fellow, but couldn’t because the software only grants access within a 15-day window. Then, hit you with an “upgrade plan” button. I don’t have money for that. And the fact that I can’t intuitively remove the bot from my Zoom meeting whenever I want makes my sensitive research-based conversations awkward.

So I end up with a perfectly transcribed meeting that never should have happened. Or if it did happen, I can’t access the information that would have been the entire purpose of the product.

This is what solution-first startups look like. They see a capability (LLMS can generate text) and go looking for a problem. They're missing the value-add from real-world research and data collection from target users like me: an undergraduate student juggling meetings for internships, networking, research, and clubs.

Jasper AI is another case study. Raised $125 million at a $1.5 billion valuation by wrapping an LLM for marketing copy. Then, ChatGPT launched, and their value proposition eroded within months. Massive layoffs followed. Revenue fell from $120M to roughly half that. They never asked why marketers struggle with content, they just assumed AI-generated text was the answer.

If Jasper had actually done deep research with marketers, they would have uncovered questions like:

• Why do marketers struggle with content? It’s not because writing is slow it’s because content is disconnected from brand strategy and stuck waiting on approvals.
• Why does marketing copy feel generic? Because the real gap is context, not production speed.
• Why do teams resist AI tools? Because they don’t fit into existing workflows – Jasper operates in a silo where generated text has to be exported through other tools to get approved.

Jasper is now trying to become what they should have been from the start – a workflow and brand orchestration tool. But they burned through their first-mover advantage and $125M figuring out what user interviews or real-world data could have told them.

The pattern is the same every time. Whether it's the data collected through system analytics, human stories, or surveys, it all traces back to human behavior. And it should, because the product must revolve around a target user. What's beautiful, and currently being ignored by solution-first startups, is that meaningful building follows the human-centered design process: data gathered in the research phase naturally raises questions This leads to evaluation, which recenters you back to asking why, which leads to an insight, then ideation, then a product that ties back to a story.

Hence, a powerful product that makes a contribution to society. Instead of a techie trying to boost their ego with an AI wrapper.

What It Looks Like When You Ask Why

Let’s talk about what happens when you actually do the work – it’s so fun and satisfying.

There's a reason I use Canva instead of Google Slides to design presentations – and why all my classmates do too, and why my college pays for free accounts for every student. It's intuitive and hence less time-consuming.

That's not an accident. In 2007, a 19-year-old Melanie Perkins was teaching Adobe Photoshop to university students and watched them spend entire classes trying to find the right buttons. She didn't start with 'I should build a design tool.' She asked why is design software so hard to use? That observation was grounded in empathy, firsthand experience, and real people struggling. It became the foundation of a platform that now generates $3.5 billion in annual revenue and serves 260 million monthly users. The why didn't just shape the product. It became the product.

It may be easier said than done, but doing the hard work up front is far superior to losing millions on an AI wrapper after users either realize the product doesn’t solve their root problem, or find a product that does.

In Essence…

The ACCESS Court prosecutor asks why does homelessness exist? and builds a preventative system. Medical professionals ask why do people get sick? and builds preventative treatments. Melanie Perkins asks why is design software so hard to use? And builds a $3.5 billion platform around the answer. Fellow.ai asks what can AI transcribe? and builds a paywall.

This toddler-like curiosity determines the success of an intervention because it measures its relevancy. If one system can transfer its "ask why first" practice to another – legal to medical, medical to startups – then this approach is universal.

So yes. Let's continue asking why the sky is blue. Why homelessness exists. Why asylum applications require an approximately $800 USCIS fee. Why Melanie needs to record Zoom meetings. And why she is writing this.

Because maybe, asking why before what is the difference between building something that disappears and building something that changes the world.