Original Chat Interface Designs
2026The original landing page was an unnecessary cover, which led to an incredibly simple chat interface. Status indicators for processing, pulling, and generating information were vague and had the potential to trigger more questions than provide answers (ie. for ‘Putting it all together…’ what is being put together?).
The original landing page was an unnecessary cover, which led to an incredibly simple chat interface. Status indicators for processing, pulling, and generating information were vague and had the potential to trigger more questions than provide answers (ie. for ‘Putting it all together…’ what is being put together?).
The original landing page was an unnecessary cover, which led to an incredibly simple chat interface. Status indicators for processing, pulling, and generating information were vague and had the potential to trigger more questions than provide answers (ie. for ‘Putting it all together…’ what is being put together?).
The original landing page was an unnecessary cover, which led to an incredibly simple chat interface. Status indicators for processing, pulling, and generating information were vague and had the potential to trigger more questions than provide answers (ie. for ‘Putting it all together…’ what is being put together?).
Responses were entirely text-based and extremely long, with a single line of questioning generating 7 sections of double-spaced text. Upon load, the chat would also skip to the end, forcing the user to scroll all the way back up to start reading Luca’s response. Internal reports utilize space and proper visuals to provide digestible information at a glance for their executives. Luca needed to do the same.
Responses were entirely text-based and extremely long, with a single line of questioning generating 7 sections of double-spaced text. Upon load, the chat would also skip to the end, forcing the user to scroll all the way back up to start reading Luca’s response. Internal reports utilize space and proper visuals to provide digestible information at a glance for their executives. Luca needed to do the same.
Taking inspiration from Claude and ChatGPT for the landing page, I included ‘check-in’ cards with insights of varying color-coded priority to notify a busy founder of what required attention and a ‘Go to chat’ to make getting started quick and streamlined. I also included a row of buttons above that let the user set context prior to asking a question, reducing the amount of time and effort spent writing context into each prompt.
For the response status indicators, I kept the loading to checkmark flow, but replaced the infinite loading loop with a loading bar to visualize progress. I also included sources Luca was interfacing with for each task, giving the user greater oversight regarding what data sources Luca pulls from for a given response. If there is an issue with the response, the user is given a starting point for where to start troubleshooting; whether that be themselves, a co-worker, or Luca AI providing support.
Ideally, a chat response should require little to no scrolling by the user, but text alone does not create an ideal response. Simple metric cards help the user quickly identify positive and negative trends, both numerically and within context, as well as the information source if additional confirmation outside Luca is desired. The arrow icon to the right of the chat provides the affordance to jump to the latest response, rather than forcing the user by default.
I also created some additional feature mockups, including editing sent prompts, a hover state which included exporting/pinning responses, and a text-to-speech feature for accessibility.
I also created some additional feature mockups, including editing sent prompts, a hover state which included exporting/pinning responses, and a text-to-speech feature for accessibility.
I also created some additional feature mockups, including editing sent prompts, a hover state which included exporting/pinning responses, and a text-to-speech feature for accessibility.
