There's a lot of mythology around AI right now.
I hear it in boardrooms, product brainstorms, investor meetings, and team standups. The market is flooded with "AI-powered" versions of every product. The pressure to "use AI" is real, and it's understandable. The tech is impressive, the potential is massive, and frankly, nobody wants to be left behind.
But here’s a quieter truth:
You might not need AI. What you do need is clarity.
Clarity about the problem you're solving.
Clarity about the outcome you're aiming for.
Clarity about what data you have, what your users need, and what success looks like for your business.
AI, despite the headlines, isn’t magic. It’s a tool. And like any tool, it’s only useful if it’s the right one for the job.
Let’s break this down with a few practical examples:
If your company wants to reduce customer churn.
You’ve got a few years of customer data. Purchase history, support interactions, NPS scores, maybe even user behavior patterns. This is a great candidate for AI. A prediction model could help surface warning signs early and give your team a chance to intervene.
But if all you have is anecdotal feedback and a few spreadsheets, then building a model isn’t just overkill, it’s likely to fail. You might be better off running some simple reports, interviewing customers, and improving onboarding or support workflows.
If you’re trying to build a new scheduling tool for your internal operations team.
Someone on your team suggests AI to “optimize shift rotations.” But when you dig in, it turns out the biggest problem is that people keep forgetting to log their hours. The pain point isn’t optimization—it’s adoption. You don’t need machine learning. You need a better interface, some smart reminders, and maybe an integration with your payroll system.
If you’re building an customer support system.
Sure, everyone talks about using generative AI to automate customer responses. And in some cases, that makes sense. But what if your users just want faster response times and easier access to past tickets? What if all they need is a search bar that actually works and a clean handoff to a human when needed?
This is the pattern we’re seeing again and again:
AI can be powerful, but it’s not always the right fit.
And even when it is, it’s often not the first step.
Unfortunately, most companies don’t have a good system for figuring that out. They go straight to implementation. They start writing prompts before they’ve written a problem statement. They buy access to an API without knowing what question it’s supposed to answer.
That’s where things break down.
Timelines slip. Budgets bloat. The thing you ship doesn’t match what users wanted in the first place.
So what’s the fix?
If you are headed on a vacation, you start by packing your suitcases. If you don't know what to pack, you need to do some research. What is the weather? How many days? Are we going anywhere we need to dress up? Do I need diving masks or ski parkas?
It's really the same thing when you build any software, regardless if it will utilize AI or not. Start with what you are trying to solve. Start with clarity
And if you are in a company that means finding out what the stakeholders expect to get from the trip. You need to talk to the teams, management, and business units to cut through the noise and get to the heart of what they actually need.
This isn’t about gatekeeping AI. I love the technology. I’ve seen it do amazing things. But I've also seen people waste time and money chasing it when they didn’t need to.
We don’t need more buzzwords.
We need more builders who are willing to ask: “What are we actually trying to do here?”
If we can answer that clearly, the rest gets easier—whether we end up using AI, traditional code, or something in between.
Let’s stop overcomplicating the process and start focusing on the outcomes that matter.
Good software solves real problems. Let’s build more of that.