AI Chat Product Design

Delivering design solutions that make AI assistance approachable and accessible for all types of users.

About the project

A client hired my company to create a closed-AI platform for small to mid-sized businesses that would not only include an AI assistant, but also store useful company files and documents. That said, the client was worried about the user experience of the product, stating that artificial intelligence can often feel intimidating to the average user. To their point, many people interact with AI through interfaces that give little to no explanation of how decisions are made, which can create a sense of mystery or even distrust. In order to create a solution for this client, I utilized UX research and design practices to create an AI product that not only felt visually welcoming and secure, but provided the user with valuable features that they could use for their day-to-day tasks. I created user and admin screens in order to showcase features available for both views.

Client:

pipIQ

Services:

UI/UX design and testing

My Approach

I began my process with thorough client and user interviews as well as competitor research. I identified common sources of confusion for users, such as unclear AI capabilities, lack of transparency in responses, and a perceived lack of control with the information being shared/addressed within a company. Based on these insights, I then began to align ideas for creating a better product for users with long-term business goals for stakeholders.

Next, I set out to use the initial research to further define what the product would need to accomplish in order to be successful. I created personas based on types of users that would be expected to interact with the platform, as well as diagrams to illustrate the ideal user flow for specific tasks. (Examples shown below)

I then used all the data and research I had gathered thus far to begin wireframing a more intuitive interface. I created wires of features such as onboarding tutorials, inspirational focus areas, correlating auto-populating prompts, accurate tooltips, and other options to clarify the AI chat assistant’s role and limitations. Once initial concepts were validated and discussed with the client and developers, I moved into high-fidelity mockups and prototyping with Figma. The prototypes were tested in usability sessions, where I obtained interaction feedback. Changes based off of the feedback I received were implemented and final mockups were prepared, presented to the stakeholders, and seamlessly handed off to the development team.

Conclusion

The successes of the project are as follows:

The design successfully created a more calming and helpful interface for new AI users to utilize the product to its full capabilities
The product itself was white labeled, so additional features were added so companies who purchase could upload their own brand logo and colors; this gave even more comfort to employees logging on for the first time, as they could see familiar visuals that signaled they were in the right place
I not only gained a deeper understanding of what UI for AI products should include, but I also learned valuable information about what is necessary on the development side during collaborative efforts with engineers