Using AI to Understand Documents
RoboReader
Businesses, especially in regulated industries, circulate a lot of documents as a means of sharing information to support business decisions. For over 2 years our team worked on ways to used emerging AI technology to streamline document intensive processes. Projects included work with investment research professionals, compliance analysts and operations support teams and included phases outlined below.
Research
We embedded with our business users for ethnographic observation to learn what they do.
Personas
I developed persona and empathy maps to document our learnings and identify behavior patterns and pain points for the team that would influence our solution and designs.
Task Maps
I mapped the tasks we observed to uncover opportunities for our tech stack.
Wireframes
I created low-fidelity sketches and wireframes to get feedback on our ideas for future-state tools and workflows. In some cases, the team conducted co-design sessions with our users. In many cases, we focused on solutions to elicit the right level of trust in AI generated content.
Design System
When we got to our final designs, we implemented them in the corporate design system to create consistency for our users and the tech teams taking over maintenance of the applications.
Adding Use Cases
We also looked beyond the scope of our project build and helped our business partners envision how AI might enhance how they work in the future. Note: this was before LLMs became widely available.
Future Thinking
Envisioning an AI enabled future experience to inspire future development
Publishing
Throughout the project, we documented our work to share with our business partners. I also published learnings related to design and AI to on our external website.