There’s a common assumption around AI right now: if you don’t know how to build it, you probably don’t have much to contribute. We think that gets things backwards.
The people with the most useful AI ideas are often not the ones who know the most about models, agents or APIs. They’re the ones who know a particular industry, workflow or customer problem inside out.
They know where time gets wasted, what people keep checking manually, which tasks always fall through the cracks and which “small annoyances” are actually costing a team hours every week.
That knowledge is not the easy part. It’s often the valuable part.
Good AI products start with a real problem
It’s easy to start with the technology.
“What could we automate with AI?”
“Could we build an agent for this?”
“What can the latest model do?”
But starting there often produces something technically impressive and commercially uninteresting. A better question is much simpler: what is unnecessarily difficult about the way this work happens today?
Take a sales operations team preparing for a weekly pipeline review. Someone might spend an hour checking which deals moved, which didn’t, whether close dates changed and whether there’s enough context in the CRM to understand why.
A generic AI idea might be: “Let’s build an AI pipeline summary.”
Someone who actually does the job will usually describe the problem much more precisely: “I don’t need another summary. I need to know which deals changed in a meaningful way, where the information is missing and who I need to chase before the meeting.”
That distinction matters and now you have something worth building around.
Industry expertise tells you what matters
Technical capability can tell you what AI can do. Industry experience tells you what it should do.
A Customer Success leader knows which signals are genuinely worrying before a renewal. A recruiter knows which candidate information actually changes a decision. An operations manager knows which exceptions matter and which ones can safely be ignored.
That judgement is difficult to recreate from the outside because so much of it is learned through experience.
People who have spent years in an industry often carry around a huge amount of knowledge they barely notice anymore. They know the shortcuts, the edge cases, the recurring frustrations and the difference between a process that looks sensible on paper and one that actually works.
That’s exactly the kind of knowledge you need if you want to build useful AI.
You don’t need a finished AI product idea
This is where people often overcomplicate things. You don’t need to arrive with a technical architecture, a feature list or a detailed explanation of how an AI agent should work.
A much better starting point might be: “Our team spends three hours every week doing this manually.”
Or: “Our customers keep asking for this, but there still isn’t a good solution.”
Or: “If I had another really good person on my team, this is what I’d ask them to take care of.”
That last question is particularly useful because it moves the conversation away from AI features and towards actual work. What would you want that teammate to notice? What information would they need? What could they handle themselves? When would they need to bring a human in? Once you can answer those questions, the technical conversation becomes much more concrete.
The best setup is usually a combination of both
None of this means the technical side doesn’t matter. Of course it does, after all someone still needs to work out how the AI connects to the right systems, what data it can access, which actions require approval and how to make the whole thing reliable enough to use in a real business.
But the person who understands how to build the AI does not also need to be the world expert in your industry.
And the industry expert does not need to become an AI engineer. That’s where the interesting opportunity sits: bringing those two sides together.
That’s the idea behind our design partner model
At Lobu, we’re looking to work with people who deeply understand a specific industry, customer or workflow and can see a problem that is worth solving. You bring the industry knowledge, how the work happens today, where it breaks down and what a genuinely useful solution would need to do.
We bring the AI infrastructure and product-building side, and together the idea is to turn that knowledge into something real, test it with customers and see whether there is a product worth taking to market.
You don’t need to know how to build AI. But if you know exactly what it should do, you may already have the part that matters most.
If there’s something in your industry you keep thinking “surely there’s a better way to do this”, we’d like to hear about it.
Learn more about becoming a Lobu design partner →
Hi, I’m Emre 👋
I’m the founder of Lobu. We’re building AI teammates that understand their role, have the context they need and help move work forward without waiting for another prompt.
I started Lobu because I think AI should feel less like another tool you have to manage and more like a teammate you can actually delegate to.
See what we’re building at Lobu →Want to follow along with what we’re building? Connect with me