DEEP - The first practical custom GPT I created was a tool that turned hours of research into seconds of work. I designed it to take the bare minimum in prompts (just the company name followed by just the index number of task required) and use it to scan the web to pull together a comprehensive research report to help go from 0 to 80 real fast. Of course, as is the case with any AI tool, the effort to go from 80 to 100 should always be human-made.
IDEATE - After tackling the discovery phase with DEEP, I started wondering what other parts of the design process I could automate with custom GPTs. I created IDEATE to test if I could have a GPT that assisted in ideation? Almost like having an additional participant in an ideation workshop.
User Research Assistant - Another phase I found myself wondering if I could speed up was user research. This was trickier, because nothing can replace the real thing. So I made a GPT that simulated an interview based on a given brief. The interview itself, being simulated, holds little value. The useful part is the summarized hypotheses at the end, which can serve as a strong starting point for building a questionnaire for actual user research.
German Assistant - Last but not least, a fun one from early on. Anyone who’s tried learning German knows how maddening noun genders can be. German Assistant is a GPT I created to make them easier to remember through visual output.