Case Study: Best Generative Image AI (Part 1)

Generative image AI is a hot topic (as well as anything related to Artificial Intelligence), and many people are eager to take advantage of these creative tools. In this case study, we conducted a thorough research process from a UX standpoint to evaluate the top 5 generative image apps and reach a verdict.

Initially, I thought that a two-part article would be sufficient. I could have condensed the insights as we typically do for our clients, but I realized that most readers would not have access to the underlying data. As a result, they would not be able to understand the testing methodology or the objectives and goals for this research. Therefore, I decided to include additional information to provide context.

As I progressed through the report, I realized that it was becoming excessively lengthy. To address this, I divided it into four parts, transforming it from a mere article into a comprehensive case study. It is important to note that while this study contains abundant information, it still represents only a fraction of what our clients receive. However, I understand more readers will be interested in the results rather than the methodology aspects.

Therefore, if you wish to directly access the results, feel free to navigate to the specific pages. In this case, Part 3 includes the apps we tested and the search results, while Part 4 encompasses the analysis and insights for the review, along with generative image AI experts’ opinions. Apart from that, you can visit the sections of this case study that interest you the most, especially if certain parts are less relevant to you.

Evaluating generative image AI Tools

First things first.

Since everything related to creative apps is subjective, we used a methodological approach to evaluate these apps. After all, we’re a UX research company, so what would you expect? 😉

Before anything, we needed to define which ones are the most important right now. The problem is that there are a lot of apps for AI-based image generation, so we needed to set some rules.

For this purpose, we used the following criteria:

• PR and Media mention
• Social media mention (in dedicated groups)
• Amount of users (when publicly available)
• App overview (consumer-facing)

We joined different AI image generation groups on Facebook and Reddit, followed all companies on their social media accounts, and also followed influential people in the generative image field. This process took three months.

You may ask: why three months if this could have been answered in a week or less?

Well, the thing is that everything related to Artificial Intelligence is moving too fast, and whatever is the “cool thing” today might change in a couple of weeks.

As a matter of fact, if we expand the timeframe to six months, the players would be completely different, and the top option wouldn’t be the one that you’ll see at the end of this report!

Initial Insights on AI-based Image Generation Apps

Our first discovery was important: similar to most gaming startups, generative AI startups have a weak presence in social media, with one notable exception: Discord. Most AI apps thrive on Discord and use it as their primary channel to connect with users. Furthermore, Discord can be used to provide prompts and obtain the generated output, as we will explore further below.

This insight holds great significance from a UX perspective.

• First, it marks a significant paradigm shift in user behavior and expectations, moving from an “app first” approach to a more conversational one.
• Second, it creates a clear division from a marketing standpoint. They are explicitly conveying the message, “We’re a demographic with our own codes; we don’t cater to the old schoolers.” This insight is intriguing and bears several consequences for the future.
• Third (but not finally), it allows for the quick creation of a community base with multiple touchpoints, ranging from galleries featuring users’ creations to direct contact with developers—something that most traditional apps don’t offer.

These initial insights were highly valuable when designing the research methodology for this case study. They helped us identify where to locate the necessary information and determine the specific information we needed.

However, they also posed a challenge for us as UX researchers. The emergence of new paradigms meant that our interviewees required varying degrees of assistance. While they possessed a strong command of the English language, it was not their native tongue, and unfamiliar terminologies could potentially discourage their use of generative image AI applications.

Nevertheless, we are accustomed to facing such challenges. In fact, they aided us in assessing the accessibility of these tools for different user groups and determining if certain skill levels were required. We considered various factors, such as technology accessibility, design proficiency, age, and familiarity with similar apps, to obtain more insightful results.

With all that being said, I believe this case study on AI-based image generation will be of interest to a wide range of individuals. It will appeal to AI novices seeking unbiased assistance, as well as UX students searching for guidance on how to construct UX case studies and reports using third-party resources that lack internal information. It is important to note that we do not hold any bias towards any particular app in this study; it is entirely objective and impartial. If any of the mentioned companies would like us to review specific aspects due to recent changes, we are open to addressing those modifications and adjusting this case study accordingly.

In summary, the results of this case study are not definitive, and it would be beneficial to bookmark this page for future reference, as it will undergo significant updates in the coming months!

Acknowlegments

This case study was conducted by our UX team and external collaborators, which included the Experts Panel. However, for methodological reasons, we are unable to disclose the identities of the Expert Panel members.

The present case study would not have been possible without the invaluable insights and support provided by Fabio Devin, MsC; Mariana Arismendi, PhD; Alice Durham, PhD; and Santiago Figuera, PhD.

Generative Image Tools: Continuation

The Generative Image AI article isn’t finished! This wasn’t too long, but the results and complete insight from research as well as UX analysis will come in the next parts, and they’re juicy enough to satisfy any taste!

This is the first part of a four-part case study on tools for AI generated images. If you’re interested in the research methodology we used for this article, please continue to Choosing the best AI image generation apps. Alternatively, if you arrived from another page on this site or an external site, make sure to scroll to the top of the page to access all sections of this study.

Ariel Gaster

Ariel is a Bachelor in Computer Sciences and writer for technology related sites. He lived in Nicaragua, Mexico, Spain, France, Ireland and Argentina. Ariel is a big fan of sports, specially football. And he swears PHP is not going anywhere!