Problem description:
When a user starts typing into a Custom GPT, the predictive autocomplete suggestions are currently generated in a generic ChatGPT style (e.g., “Do you want to talk about your favorite TV shows?” or “Have you tried a new food recently?”).
These suggestions are irrelevant and break the experience when the GPT has been carefully designed for a specific business purpose.
In our case, we use BIZBIA, an assistant built with the IDES methodology (Innovation, Differentiation, Efficiency, Satisfaction) to support business growth with AI orchestration. Users expect suggestions related to strategy, OKRs, growth diagnostics, or design systems. Instead, they see generic social chit-chat prompts, which undermines user trust.
Proposed solution:
-
Allow predictive autocomplete to be fed by the same context as the configurable conversation starters in the GPT settings.
-
Alternatively, give creators the option to disable generic autocomplete entirely in Custom GPTs.
Benefits:
-
Stronger alignment between GPT purpose and user experience.
-
Avoid user confusion and distraction in professional settings.
-
Builds on the already existing improvement of customizable conversation starters.
Example:
Instead of showing:
“Hi there, would you like to talk about your favorite series?”
The GPT should suggest (if configured):
“Shall we review your OKRs aligned with the Differentiation vector (IDES)?”
“Do you want to run a growth diagnostic for your Innovation area?”
Conclusion:
This change would make Custom GPTs feel like true purpose-built assistants rather than “ChatGPT in disguise”, which is critical for adoption in business environments.