AI and humans: using the BDI model to check if your online UX is exceeding expectations

These days the online presence of your company is much more than a static website. Demands from the user’s perspective are always increasing, and companies go all-out to accommodate them. But how hard is it to get grips on this? And how can one know if they are doing a good job at it? The answer might be much more simple than excepted, if we look at the world of Artificial Intelligence (AI).

One of the challenges in AI is to have an easy and basic model in which intelligence and behavior can be captured and explained. Intelligent (autonomous) Agents are tiny self-contained software programs which act like humans without pre-programmed behavior. They generate their behavior on the fly, following a basic set of principles, the model. One of the simple but powerful models is the BDI model: Belief-Desire-Intention (Michael E. Bratman, 1999).

The model states that human (and therefore AI) behavior consists of the following principles:

• Beliefs: Information about the world (facts and assumptions).
• Desires: objectives, things you want to accomplish. Desires may lead to goals. Goals are desires which are actively pursued.
• Intentions: actions chosen to be executed to reach a goal (or fulfill a desire). A sequence of actions is called a plan. Plans may consist of other plans and can constantly change because any intention can change the beliefs, desires and other intentions, apart from those being changed by the outside world.

As nice as this model fits AI agents, it also fits humans. Each of us have beliefs, desires and intentions. And we also have goals and plans, which may change over time. But in principle, for any customer (user) the online presence (website, app, etc) UX should at least accommodate the BDI of that person, and if possible, exceed it.

As an example, we have Chris. Chris bought a phone from some provider, and dropped it. The phone is now cracked and bruised, and he wants to have it fixed or replaced. He does not know exactly the terms of his insurance or what the costs will be, or how long he will be without a phone. He turns to the providers website for answers. He expects he will need to do some digging around, and read a lot of fine print, but he is willing to.   

Already, we can see there are a lot of beliefs in Chris’s mind. For starters that he owns a phone and has insurance. The phone can be replaced or fixed. There might me terms and conditions which could result in costs and him being without a phone. But also, he believes the providers website will hold the answers, but he also believes it will be quite a challenge. Now, Chris has some desires. Mainly, he wants his phone to be fixed or replaced. But also not being without a phone for some time could be a lingering (subconscious) desire. His intentions are clear: he will use the providers website to get things moving and do a lot of digging. So, how will the providers website hold up to Chris’s BDI?

Let’s assume the fantastic case that the provider has set up the website to fork into a sales funnel or a customer service funnel. In this case Chris will go directly into the service part. Confronted with some frequent chosen options, he selects the “broken phone” feature page, where he can indicate if he bought the phone with or without insurance. Surely, Chris selects “with insurance”, and the website tells him “Please find the store closest to you below, and make an appointment to swap your phone for a new one. In our store we can also backup and restore your data to the new phone”.

This might sound hilarious to you, but some premium providers or insurances work exactly like this. In this case, Chris is probably overwhelmed and (maybe unknowingly) is instantly transformed from customer to ambassador! Likely, Chris will recommend the provider and insurance to his friends, or even publicly online, because he is so happy with the outcome.

But what exactly did happen in the above described scenario? Well, on all three aspects, the providers website exceeded the Beliefs, Desires and Intentions of Chris. His beliefs, that he would be going through the fine print and would be doing a lot of clicking and searching are replaced with new beliefs that the provider’s website is actually very intuitive and the hustle he was expecting was actually a false assumption. His desires are not only met (finding answers), but grossly exceeded because the UX was even accommodating desires which were no goals yet. His intentions, or plan, was cut short because the UX did not need all the steps he had planned for.

So good UX is all about exceeding Beliefs, Desires and Intentions! In every part of your UX / UI, but also in your provided feature set, you should ask yourself if you are exceeding Beliefs, Desires and/or Intentions. Are you going above and beyond existing (known) beliefs? Are you AT LEAST fulfilling expected desires (and maybe even turning other desires into goals)? Are you making sure the user’s plan (intentions) consist of the least amount of steps as possible? If the answer to one or more of these three questions is YES, in general you are doing a good job! Users tend to get frustrated if negative beliefs are added to what they already believe (bad experiences), if desires are not fulfilled and when too many intentions are need to reach the goals. The best UX exceeds all three principles at the same time.