Gamma beat my AI slide system. I’m keeping both.
Buy the result you need today while building the option you may want tomorrow.
My wife is working on a product and needed slides, so I walked her through my short series on building an agentic slides system.
She had very little to work with: one existing PDF guide and a couple of slides she liked as inspiration.
She asked the system to study them, extract the design takeaways and use those lessons to create her new presentation.
But the first attempt did not quite get there.
My argument was that we could not expect perfection on the first try.
She had a presentation to finish, though. The inspiration slides had originally been created in Gamma.app, so she took a similar brief back there.
Gamma won, mostly on design, with more variety and flexibility.
And honestly, this is expected. Gamma is a dedicated presentation product that has been improving for several years, while our system was literally on its first day.
That does not make the experiment a failure.
The slides Gamma produced are now additional context she can bring back to the local system, and ask:
What worked better in these slides?
Which layouts should become reusable options?
What changed from the original guidelines?
What should it try differently next time?
Then she can save those lessons locally alongside the final presentation, template and design rules.
An agentic system does not improve merely because you use it. It improves because you keep giving it feedback and additional context.
As the context accumulates — and the underlying AI models become better — the gap may shrink.
You do not have to choose build or buy once and defend that decision forever.
But you can buy the result you need today while building the option you may want tomorrow.
If you are trying to decide where an off-the-shelf tool is enough, and where owning the workflow could become an advantage, this is the kind of work I do with founders.