In late 2017 I had the opportunity to work as a UI/UX designer for Oracle's new Artificial Intelligence Applications. Wherewith machine learning and data-based engineering, we set out to provide users with a new set of resources that not only strive for better results and decision making but keep the user at the center of the process, providing better AI adoption and increasing trust.
However, this was not an easy task; the following is a summary of my journey through this experience.
To bring the best of our capabilities, my team and I needed to have a better understanding of AI and how the user would have the best possible experience and adoption of these technologies.
We took this daunting task as another project, where we started with a lot of research and then come up with a system that would meet our needs and those of the company.
Learning the basics
We start with the basics, the most essential and simple concepts of artificial intelligence. Our task was to investigate different publications and scientific papers.
And in fact, AI was everywhere, from airplanes to translation tools. Knowing this made it easier to detect patterns and uses in other applications, finding references to use cases and potential solutions to problems.
We also encounter the different classifications of artificial intelligence, being categorized by techniques and applications. Soon we start to differentiate and understand concepts, participating more proactively in the early stages of the planning of our applications. Then we realized that it was crucial to automate processes, shorten tasks and allow the user to work side by side with the AI instead of relying only on AI-driven recommendations.
Setting the principles
All this led us to follow a series of principles that we gathered throughout our research, applying them, and adjusting them to the diverse needs of the product. We started with trust issues in order to generate the right level of interaction between the user and the AI and then addressing ethics and human rights topics, as well as possible cases of AI misuse.
We also began collecting examples of AI usage patterns in other applications, so that we could have a visual and practical reference, to be able to link it to its respective principle and easily illustrate its use.
This helped considerably to illustrate our decisions in a better way as well as to provide us with a good foundation of good AI practices and to be able to link this in a design perspective.
Framework and governance
Finally, we needed to share this knowledge across the company and help other teams who were recently interacting with this new AI approach to their applications. We teamed up with Oracle's Research team in order to take on this crucial task.
Our first steps were a series of interviews with the different team leaders to understand their products and their struggles better. Then, we moved on to a design review and audit where we detect the parts of the product where AI was present or could have potential automation, as well as the likely misuses. Finally, customized recommendations would be provided for each product, referring to the general principle from which they originated.
However, our task was far from over. Helping other design teams and being a reference point for AI design was already a full-time task; we also needed to take charge of our products and continue providing solutions.
So we decided to optimize our process as much as possible, starting with the mockup process and the maintenance of our design library, being these two the most time-consuming areas. Therefore by providing the development team with a design theme within their development framework, we were able to eliminate these two tasks.
Our deliverables would focus on having accurate flows and solid wireframes; these would serve as a direct reference to the developers, using our design theme to assemble the layouts accurately.
This way, we could focus on continuing our research to deliver more effective solutions for our products and the rest of the company.
It was an excellent opportunity to work in a company as big as Oracle, especially on a new suite of products that would shape the future for all their upcoming applications.
AI is not rocket science - In general terms, working with artificial intelligence products is not much different from any other design project. However, having at least basic concepts and understanding how AI works helped me get a broader perspective of possible areas of improvement and constraints alike.
Research - It is essential to do research work, and many times it is taken for granted, I admit, I have made this mistake in the past. This experience, however, helped me to understand its importance and how it plays a critical role in making us better designers, expanding our vision to better solutions.
Experiment - Trying out different approaches to improve our work process was very helpful in taking away the belief that each project must follow a rigorous checklist. Not every methodology or process will work the same for every project, so adapting them is a useful way to find practical solutions