Solidabis

Making the world more usable – Case Attractive.ai


Everyone has had terrible experiences with online web services: buying flight tickets, for instance, is not for the faint hearted. Indeed, there are reports claiming more than a trillion dollars is lost in e-commerce alone due to bad websites. Attractive.ai has come out with a service that constantly evaluates and fixes the user experience of websites. Their bot points out problems and areas of improvement and many, with one click, can be fixed by the bot as well. The bot is trained and kept up to date by the real life UX expertise of Solidabis.

Arto Vuori, CEO of Attractive.ai, explains the idea and philosophy behind the product: “We are on a mission to make the world more usable. While products are developed at tremendous speed, new versions being released on a weekly basis, the user experience evaluation is always lagging behind. Right now UX evaluation is either based on real user feedback and analytics tools, user testing, or heavy expert evaluation. Our key decision for Attractive was that it needs to work 100% autonomously and produce clear and easily readable user insights, essentially emulating the work of human UX experts. Only full automation makes it possible for every team to do continuous evaluation of user experience, which is exactly the impact we want, for improving the state of the UX in the world. Of course, humans are still needed for many parts of the work. For example, we need humans to say whether high level processes are logical and achieve the desired business goals. However, we can already cover a significant part of the work by robots.”

Attractive.ai’s Head of product, Kristoffer Lawson, continues: “One big theme in Attractive is accessibility. According to the World Health Organization, 15% of the world population live with some disability. Ignoring them is not only morally questionable, it’s also financially poor judgement.”

”Not only that, but designing a service to cater for blind people, or those with reduced motor skills, generally makes the experience better for everyone through better functionality and layout. An intelligent bot can notice things an able-bodied designer, with their 30” color-calibrated screens might not.”

AI spotted poor text contrast on a web page

AI can automatically recognize poor contrast and inadequate accessibility on web pages

Training the AI

Solidabis had already helped in development of previous projects and was a natural choice as a partner. It all had a humble beginning when head of Solidabis’ service design department Leena Ketonen provided  Arto the first list of important aspects in user testing: “The very first trial of creating a heuristic evaluation based on an automated testing sequence transformed rapidly to questions about the nature of expert evaluations, improving efficiency and what is the value that humans bring to tiresome checklist walkthroughs. After a few manual trials and a workshop together with Arto and Kristoffer, we set up a 5-person multi-disciplinary UX expert team to check the early AI findings for correctness and if anything had been missed. We also tried to find more examples of findings to propose for the AI to learn. Some were simple and easy to teach while others were more complex and vaguer. The more complex ones really got Kristoffer excited”

“On the other hand, when going through typical heuristic rules and especially accessibility requirements we quickly realized that a computer is far more efficient than humans in comparing multiple variables in large data sets and all possible navigation paths. For example, checking the consistency of color schemes is a surprisingly easy task for computers”, Leena continues.

Arto sees the importance of having real UX/UI professionals as a part of development team: “UX designers of Solidabis gave us professional feedback and insights. If you ask regular people on the street and they say things like “I like this” or “don’t like this” but can’t tell exactly why. Artificial Intelligence is only as good as the data it is learning from and therefore, we wanted to teach it with the best information possible.”

”As a result, using Attractive.ai differs quite a lot from regular user testing, since actually you get a group of UX professionals loaded with scientific data looking simultaneously and continuously at your site.”

Coca Cola site 404 bug spotted by Attractive.ai

404 link error bug on Coca Cola site spotted by Attractive.ai

Is AI a threat or possibility for human UX experts?

One might question why Solidabis is doing this, and is AI making the human-powered UX/UI design business obsolete? Leena does not see it this way: “As the design maturity level in the software industry rises, there is a growing demand for designers to concentrate on the service design and to verify that the experience throughout the whole customer journey meets customer expectations. As design professionals, we have to deliver more than today. The best way to accomplish this is to harness tools that allow us to turn our attention from listing problems to decision making, creative problem solving and business process improvements.”

Kristoffer agrees: “There is a symbiosis between us and Solidabis. The Attractive.ai bot tells you what to fix and the UX designers of Solidabis can come up with creative solutions on how to do it. There are already real-life examples of this. This is especially true when companies realize, sometimes as a result of our observations, that the site might need a bigger revamp.”

Arto sees Attractive.ai as a productive tool also for UX professionals:

”Talented UX people will welcome this service. It can help you concentrate on things where people excel such as creative design, bigger entities and feeling.”

Aleksi Tiira, an UX-expert from Solidabis, agrees: “If one looks at the current spectrum of predictions on how design industry and work will change in the future, a clear emphasis can be seen. This emphasis points towards technology and AI aiding designers in their work rather than diminishing and eating up designers’ work. Repetitive manual repetitive tasks can be assigned to AI and instead of having to create a layout from scratch an AI can give 5 common layout suggestions. With the help of assisting technological solutions, anybody could more easily do design and be a designer, meanwhile the core competence of professional designers would shift towards on one hand managing the design process and decision making and on the other being truly creative and disruptive. Given these kinds of predictions, Attractive.ai is truly looking forward and is strategically aiming for not just today’s markets but tomorrows as well.”

Concentrating on the right things

But what makes Attractive AI unique? Kristoffer explains: “Our solution is smarter than what our competitors are offering. It does not dump and drown users with data but knows how to concentrate on important things. It is like a team member that in meeting knows when their input is needed, gives it, but shuts up about everything else that might come to mind.”

The menus on Huawei site were inconsistent. This was spotted automatically by Attractive.ai

A question might rise on how customers are perceiving robot generated data.  Arto explains that this hasn’t been an issue: “Maybe even on the contrary. Most people actually trust algorithms, since they are usually considered to be more neutral. The data and findings provided by Attractive.ai has been utilized to sell improvement ideas. People might know of an issue but can’t get it forward in an organization. When they get the AI-generated data, it is much easier to convince the decision makers. It is not just someone’s own opinion, but a scientific fact.”

Co-operation with Solidabis

The work with Solidabis is ongoing. The AI is constantly evolving and Solidabis is helping in this process. “The robot evolves and there are checkpoints when AI is benchmarked against human UX experts. AI is also updated with the latest trends and technological advancements in UI/UX field”, explains Arto.

Kristoffer has been happy with the co-operation: “We have found the right people and developed a good process. This equals good co-operation. We have been very satisfied.”

Leena from Solidabis agrees: “Working together in creating new AI based software has naturally been inspiring. It has demonstrated what kind of assisted tools experts like us could really benefit from and pushed us to consider what an AI can spot better than a human. Our co-operation has been one of a joyful discovery and both Arto and Kristoffer as trainers have taken on our ideas and findings as a positive challenge for their inexhaustible student, the AI.”

Check out attractive.ai and test your site!

Read more about our service design services (in Finnish)!