Innovation is usually framed as invention, something new, shiny, and never-before-seen. But often, the most powerful breakthroughs come from something much simpler: using what already exists in an unexpected way.
Here are three lenses I’ve found useful when trying to build something different.
1. Innovation as Misuse
Some of the best innovations started as accidents:
- Bubble Wrap was designed to be textured wallpaper.
- Microwaves were discovered when a radar engineer noticed a chocolate bar melting in his pocket.
- Viagra was a failed heart medication.
In AI today, misuse might be the whole point.
We think of it as a chat tool. But what if that’s not its true form?
Maybe AI is best used as an operating system for decisions, one that observes how we work, identifies friction, and reconfigures systems in real time.
2. Innovation as Refusal
Another way to innovate is by saying no:
- Netflix refused the physical store.
- Canva refused complexity in design.
- Tesla refused dealerships.
What’s accepted as normal in your industry that you could simply reject?
In retirement savings, is it the assumption that advice must be expensive? That customer service must be reactive? That scaling requires headcount?
Refusing the norm might be your most original act.
3. Innovation as Inversion
Flip the goal:
- Instead of using AI to recommend, use it to filter or eliminate.
- Instead of automating a bad process, ask if the process should exist at all.
- Instead of asking for more data, ask how you’d build something valuable with less.
Constraints force clarity. Inversion forces questions others don’t think to ask.
Innovation at Sevaka
At Sevaka, Inversion is our starting point for innovation.
We don’t begin by asking how to make advice cheaper or faster. We ask why the system needs to work the way it currently does.
- Why does financial advice need to start with a meeting?
- Why does customer service rely on staff searching through Word documents?
- Why do compliance teams spend hours reviewing things the system already knows?
By flipping the question, we’ve built an AI that supports humans by removing the friction they never signed up for, the searching, the form-filling, the rework.
That’s how we’re approaching innovation: not by stacking on more features, but by quietly taking away what doesn’t need to be there.
If you’re building something similar, or thinking this way, I’d love to swap notes.