As any student in design school can tell you, the first design principle you come across is “KISS” (Keep It Simple, Stupid). This acronym, attributed to the late Lockheed Skunk Works lead engineer Kelly Johnson, is best understood if you remember that Lockheed’s products were often designed to be used in the theater of war. Kelly’s acronym would remind the designers at Lockheed that whatever they designed and built had to be simple enough so that it could be maintained and repaired in the field using basic training and simple tools. As Lockheed products, their use case would not allow for more complexity than that. In other words, if your products were not simple and easy to understand, they would quickly become obsolete and essentially worthless in combat conditions.
Decades later, this axiom applies, whether it’s conceptual physics, elaborate engineering, or consumer products. The end user doesn’t care how clever the creator is, they care about being able to use the output of this creativity, to make it useful to their own application. The simpler the product or execution, the more likely it is that this output will be useful to the user.
What’s true for fighter planes and mobile applications is especially true for Artificial Intelligence (AI-) and Machine Learning (ML-) powered features. When you think about it, AI and ML algorithms are an extreme example of the importance of the KISS principle. Highly complex in nature, AI and ML are perceived as a complete and untouchable black box by most users. In order to use them properly, the widest possible audience must be able to understand their output and effect on the user’s task. And even the most complex intelligent systems must still feel simple.
As Informatica products are built with the CLAIRE engine at their heart, it becomes a top priority for us to simplify the design of features that are expanding in back-end complexity. We simplify by following a few Simple (J) guidelines:
- Favor text over visual explanations
A picture is worth a thousand words, and that’s exactly the reason you want to avoid using images to explain a complex task. We use simple, one-sentence explanations to help our users quickly understand and decide on any actions related to AI-based features.
- Use the right vocabulary
When explaining AI decisions and rationale, we try to avoid using highly technical or scientific terms related to Artificial Intelligence or Machine Learning those terms would require prior knowledge and education. Instead, we present explanations in simple language that everyone can quickly understand.
- Break down the complexity
Any complex idea can become a lot simpler when it’s broken down into smaller steps. We apply this stepped approach—using clear, understandable language—whenever we’re trying to explain tasks that cannot be briefly summarized.
We apply these principles consistently throughout our design. Here’s an example of how we handled algorithm-driven recommendations for tags or stakeholder assignment in input fields:
Recommended stakeholders or tags are shown inside the input field, and visually branded to indicate that they are driven by the CLAIRE engine. The reasoning behind the recommendation is provided in a short tooltip.
This is a perfect example of KISS in action: The user has an option to apply or dismiss the recommendation, and a simple UI control that does not require prior knowledge of the intricate details of the algorithm’s technical details is working behind the scenes.
Branding these “smart” inputs with a distinct CLAIRE color and appearance in various places further reinforces user’s awareness of these unique items.
We simplify the user experience whenever we can to allow the user to carry out their tasks with our AI-based products. By maintaining our other design principles of Trust, Clarify, Control, and Humanize, our users are able to leverage the intelligent engine for more areas of their work.