Artificial intelligence tools that could change how we do UX
This is part two of a series of blog posts derived from our experience researching AI tools as UX Designers. The effort was co-led by Kevin Lopez and Fermín Chávez, both members of our UX Team @ Encora. The latter authors this post, but credit goes to both and the UX Team at large for their contributions. Check out part one here.
When exactly is the future happening?
Last time we promised to discuss the Artificial Intelligence (AI) tools that would soon reshape our work, but to quote The Smiths, how soon is now? Let us define what we mean by soon, which feels important when we are talking about such a fast-changing field as AI. We mean the immediate future, things we should expect within this year. No, not the killer robots, those are still far behind (fingers crossed). The efforts are just one small technological leap from taking full form, and, who knows, overtaking all others.
Sadly, we are not oracles, we cannot really predict the future. But we can anticipate it by looking at the current technology, the things already underway, and the investment that has been put into them. That is how we distinguish between empty and tangible promises. The second group is the one that interests us. Without further ado, let us jump in.
The promised land
How do we measure promise? These are tools whose performance is still lacking (if they offer a working version at all) but enough to garner the interest of practitioners and investors alike, tools whose technological grounding makes sense within the AI capabilities. Out of the 15 application areas we categorized, two met these criteria: Automatic User Interface (UI) Generation and Project planning.
Figure 1. The two application areas show promise for the near future.
Automatic UI Generation
If there is one application area capable of making designers tremble, it would be this one. Will we lose our jobs? I would argue not. Editable User Interface Designs surely capture a lot of our time, and are one, if not the most important deliverable. Yet, to reduce what we do as UX professionals to that would be a misrepresentation. Discovering user needs, gathering requirements, conceptualizing solutions, and testing them, are all things that need to happen for us to deliver the final designs. Where else did you think they came from? We will still have plenty to contribute to the value chain when their generation gets automated.
Figure 2. Image generators like Dall-E are still far from delivering editable designs.
And how will it be automated? Galileo AI makes the ambitious claim that once it is launched it will produce full UI designs from a plain text input (e.g., “An onboarding screen for Pharmacy app”). It is worth noting that so far, no tools are fully capable of delivering on that.
Text prompt image generators can provide representations of non-functional prototypes for inspiration while chatbots like ChatGPT can provide low-fidelity, interactive HTML code that represents basic ideas. Both are still far away from the original promise. It might take some time until we get tools that can fully produce quality UI automatically, but it will be interesting to see how they could complement our work as designers rather than simply replace it.
Project Planning
Oh, project planning, the balancing task we can never fully get right. Project planning involves a large scope with high contextual complexity, there are a lot of elements, variables, and constraints at play. Chatbots like ChatGPT can be a great entry point into structuring a project plan pointing to a bright future where, after describing your project in detail, it can produce an effective plan, dividing it into tasks for each day or week.
Of course, in their current state, tools are fallible and tend to over or underestimate timing and resources. Improvements in tools like this can undisputedly assist us in shortening the time required for the planning stage. Heightened accuracy and integration with our current productivity software are possible, and the differentiators we should look out for.
Should we reserve seats, wait for an RSVP?
Figure 3. Someday Galileo AI may deliver its promises or other companies will.
It is a major no-no in UX Research to ask users for hypotheticals: “Would you like a feature that does X?” We cannot really anticipate how we will react to a given situation unless confronted by it, especially not when it is something for which we have no frame of reference. These tools put us in a similar place as UX Professionals. We can play around with AI-supported project planning and assess it as it evolves, but automatic UI generation is not something we can test yet.
Does that mean we should sit and wait for these capabilities to be fully realized and then we will see about incorporating the tools? No. Must we start reshaping our processes to welcome a seamless transition? Maybe. But that does not mean working around an empty space whose chances of being occupied are not all guaranteed. What it should mean are experimentation and prospective exercises. Much as you would do in UX Research when you do not have a final product to show; promoting interactions as close to the ones you anticipate, letting the users engage with those ideas in a palpable way, then you can trust their reactions. By setting up the appropriate proofs of concept, we can be ready for the impact these tools will have on our work and identify the preparations we will need to make.
In the meantime, as discussed in Part I, there are plenty of AI tools UX professionals could, and should be working with. As for those discussed here: be on the lookout, the technological leap they need could happen any day. Maybe it already happened by the time you finished reading this.
About Encora
Fast-growing tech companies partner with Encora to outsource product development and drive growth. Contact us to learn more about our software engineering capabilities.
