Undoubtedly, this is the age of artificial intelligence; this is also the age of experience, and the two cannot be separated.
While the power that drives artificial intelligence is the availability of large tranches of data and explosive computing ability, what will really enable businesses and individuals to make sense of Artificial Intelligence and its applications, is the deployment of AI capabilities where they really matter.
So, as designers, engineers, and data scientists, how do we unlock the true potential of the capabilities that artificial intelligence presents to all?
Finding the humanity in AI
The classical human-centred design approach, whether it’s architecture, web design, industrial design, engineering, etc., relies heavily on the stories that our users tell us. These stories contain insights not only about the needs our users have but also enable us to see the problem itself from the user’s point of view. What this does is it automatically inhibits our impulse to start building the solution right away without truly understanding if we are solving the correct problem.
To illustrate this, let’s take an example of an image-generating model, think Dall-E or Midjourney. Right now, their powerful image models allow you to create images using text prompts that control everything about the final output. You can have highly contextual prompts, which can include people, places, time periods, and styles, or it could be something as random as the lyrics of Eminem’s “Lose Yourself”.
Although they are powerful, they lack specificity in what problems they are trying to solve. The next step for these products would be to create specific workflows for use cases like branding, conceptual arts, social media content, etc. Users will require systems to help them control and create something meaningful and valuable out of randomness. It’s the product’s job to get better prompts from users and translate those to the model in a meaningful way. User stories help a product root itself in real-world use cases, something any business needs to do.
User Stories = Articulation of the correct problem
User stories will not only allow you, the designer, to see the problem clearly, but they will also allow the users themselves to realize and define the problem better. Often what happens is that the users themselves are not aware of the problem. They have ‘an idea’, but it’s usually very vague and difficult to design.
Asking for stories, and probing for themes, helps the user articulate their problems better and therefore set the designers on the right path. Having a feedback cycle, for example, can be a great data point to monitor in order to grow that model in a meaningful way.
This will result not only in a solution that is appropriate — desirable to the users, but will also save a lot of time and cost of unnecessary and futile iterations.
User Stories = Intuitive AI
Coming back to artificial intelligence, today, AI is in almost every field of work. For any application of artificial intelligence that is being developed, it is important to unlock the stories of each of the users, the humans who are a part of the solution ecosystem. The technology you use should be guided by the user experience you want to achieve. Instead of diving headfirst into algorithms, think about how people do the task today.
Most solutions that fail do not fail because of technology; they fail because of the lack of adoption — because of the lack of desirability.
When solutions are built without taking user stories into account, they often face a lot of resistance, and why this happens is because the first criterion by which any solution is judged is its desirability. To make a solution desirable to the user, you first need to understand the user, their worldview and the problem itself from the user’s point of view and start building a solution. This Universal rule applies to any solution that is being developed using artificial intelligence or any other technology.
A lot of the applications that use artificial intelligence are at an algorithmic level and do not necessarily expose controls to users. With the image generators, there was a way to enter prompts, but processes like automated customer service calls, image processing, etc., work under the hood. Right now, users are not really in control of all the variables that could help you dial in your AI to specific clients, use cases and workflows. It’s a designer’s job to make the experience flow in a way that helps users, and a huge part of that is to make the decision of using or at least trying to make your product as easy as possible to use, by targeting specific use cases.
User stories not only have the power to help create meaningful interactions between Artificial Intelligence and humans, but they also have the power to affect the very algorithms on which your AI is built.
User Stories = Better algorithms
Allow me to give you an example of how empathy generated from user stories can affect your algorithms:
Let’s say, for example, you are building an AI-based solution that detects anomalies within given data sets, and let’s also assume that this solution is in the form of an algorithmic suite that helps managers and decision-makers identify an act upon any threat or opportunities that may come up in a seemingly blue sky market scenario.
You have an algorithm in place which can detect anomalies — upcoming trends within the data sets that you provide to the AI. Now, there might be a lot of anomalies that your algorithm detects within the data given to it. You cannot feed all of that information to your user; it would simply result in a cognitive overload.
For any analytics platform, the key stories to map out are the ‘data-to-decision journies’ of the users. To do this, what you need to do, is before you start designing the algorithm talk to your users, assimilate their stories and analyze them to understand which data points, which insights, are important to your users and then use that understanding to build stories that are easy for your end users to consume.
In this case, the user stories have affected the same algorithm based on which your artificial intelligence decides which stories need to be told and how to tell them.
Artificial Intelligence is everywhere. Every organization is talking about building a capability in AI. The difference between those who succeed and those who fail will be their ability to understand, decode, and utilize the stories of their users.
Consciously controlling the biases going forward
Crafting these magnificent AI models is akin to the art of Bonsai. The data engineers act as Bonsai masters guiding these models, rebuilding and automating workflows that would take a lot longer. These models not only reflect some real-world biases but also present an opportunity to introduce better ones. Thinking about users and their use cases can help concentrate that effort and even act as a starting point for conversations like the ethics in designing AI, design inclusivity and design regulations.
Canvs Editorial regularly brings you insightful reads on design and anything related. Check out the work we do at Canvs Club.
Follow Canvs on Instagram for more design-related content.