I started working on Bright Ape just before Christmas 2024. I had a completely empty calendar for 2 weeks except for “Get an Oil Change” so was looking forward to some much needed relaxation. Alas, ChatGPT (ChattyG) had other ideas. I’d been receiving emails reminding me Invision was closing down permanently at the end of the year and struck up a conversation with ChattyG about what this meant for designers and product people and (of course) what startup opportunities there might be to replace some of what InVision provided (“possibly using AI”).

After we traded a few ideas back and forth, one stuck out as being particularly interesting, “predictive design insights”. This was super interesting to me because I’d been complaining for months about a product I’d been working on not having enough traffic/users to make confident “data-informed decisions”. When you don’t have data, anyone’s opinion is just as good as anyone else’s since you just can’t prove one way or another which is better, which of course leads to feature bloat, missing the mark, and ultimately not hitting growth targets.

So far I’ve built out some really great persona generation (if I do say so myself), behavior simulation of those personas based on the uploaded design, accessibility auditing, as well as comparing the uploaded design to other designs (competitors, other designs, etc.) and have started to train a custom computer vision model to be used for more behavior and heat map simulation.

The results so far have been incredible and even unfinished the insights I've received for some of my work have been quite useful.