Irfan Mir's Portfolio

Ellsi (research & design)

User Research

After conducting surveys with a diverse group of students exploring how they used technology, with their consent, I accompanied them in certain situations observing their interactions with notifications on their phones and computers all aiming to learn how notifications, and technology in general, manipulate users.

We (a team I had recruited) interviewed students, the campus counseling and psychiatric services, and off-campus services further seeking to learn more about technology and notifications' impact on millennials’ mental health. In addition, we interviewed older "Gen Z" (late middle-school, high-school) individuals with their own and their parents’ consent to explore those that the grew up in an age surrounded by the internet and technology. We learned, with anonymity, about the effect notifications have on human wellbeing including anxiety developed through anticipation of notifications over time to instances of ringtone hallucinations; for example, phantom vibrations. However while the impact of technology and notifications remained diverse as the groups we were exploring this impact with, an emerging sentiment was revealed where notifications and technology resembled stress, work, and labor whereas the desired or expected sentiment from technology was relief and entertainment.

To conclude our research, although what determined to be an equally important phase, we conducted contextual inquiries with our participants on breaks at their place of work, when they used their phones and had an opportunity to read received notifications (they were and were not anticipating), as well as in other locations in their life when given permission. When given consent, we measured their heart rate during these inquiries to determine a physiological effect when anticipating or receiving notifications.

We reached the conclusion that much of technology including that which may be built with user experience and empathy in mind, or so appears to be, actually has a pervasive, intrusive, addictive effect on the user in their everyday life. And while that gains more users or a cult following, it is detrimental to the wellbeing of the humans the technology should serve or at least be useful for in their lives.

Classical Research

While conducting the user studies summarized under User Research, I independently and individually conducted research in Artificial Intelligence from a philosophical and cognitive psychological perspective under supervision and later review of experts in the field. In doing so, reflecting on my readings and times studying consciousness from theistic, atheistic, and non-theistic perspectives, I wrote three papers (introduced in the page on my papers and presentations). My aim in doing so was to create a strategy reflecting on consciousness’ evolution and from that create a methodology and techniques for technology to be a more human-centered and empathetic experience unlike that which was shown in the User Research. To promote and benefit the user and gain a genuine, non-cult following. I did this by exploring behavior, goal creation, and drive towards goal achievement and reflected on the pathways manipulated in the user—perhaps even a weaponizing of empathy. Whereas my aim was empathy and compassion from the technology itself, I then sought to explore levels of consciousness and their evolution from sentiment onwards across species and through evolution. Through both of these, I developed a system and formula to afford a natural drive in technology, to be more empathetic, through accomplishing its goals of serving the user.


After concluding these two phases of research, I began prototyping formulas of my research in software as well as interface concepts and experience elements for an assistant. I believed other than a set of techniques by which to build more human-centered technology, which was also developed, a system to check notifications in a more relevant, contextual manner, to the users' wants and ambitions not manipulating instinctive drives, was needed, as per the research. Between 2013 and 2016 there were a few phases of iterative prototyping, design, and testing; but, among them were gaps as I was restrained by intellectual property and no-compete agreements. However, when I was intellectually free and legally able, I built a high-fidelity prototype featuring more interactions and a complete experience designed to serve relevant notifications with empathy so as to empower and promote the user, the human, in their life.

A key element of these prototypes were a formula and implementation of artificial emotions to recreate drive and purpose for the user and their goals as explored in my papers. For more on this and how the papers relate to Ellsi, I refer you to Ellsi's introductory homepage.

I am able to demonstrate live instances and recordings of these prototypes in action. I can not however make it live for unlimited, readily use due to API limits and other costs that I can not maintain on my own, I hope you can understand.

Visual and Interaction Design

Now in the process of a true visual design, we researched further to construct human-centered interface elements:

A feature I designed through user and scientific research studies is the placement of the call-to-action button. While most humans are right handed, one can assume through efficiency to place a call-to-action button in the lower right would be optimal; however, this isn't inclusive. While the lowest area of a screen is most reachable, the thumb is at its resting point when extended and not flexed at the joint, its natural position, meaning the call-to-action in the bottom right would not be the optimal place for a right-handed individual, but perhaps a button on the left. For a more inclusive design, we use an elongated ellipse call-to-action button reachable by both thumbs when relaxed, which is; the button, retractable to follow the thumb, in the lower area of their choice as visible in the designs below.

In addition to a conversational interface, I built more direct interface controls to optimize completion of quick tasks or to prompt direct action in an unobtrusive, clear and concise, fashion.

Preview of moment view in Ellsi
This is first done in the moment view where we summarize what, through the user’s indication and our research, is most relevant to the user at this moment and reveal three prompts to promote use.
Preview of conversational view in Ellsi
The main form of user interaction is the conversational interface where we use natural language understanding and standards for conversational design, cards, and visual design to best deliver information and prompt action across the app and the device.
Two previews of mail view in Ellsi
From conversation mode, one can view their relevant notifications within the app, for example, email, showed by two different methods in these two views. I’ve also used chips to easily find emails of certain types; for example, payments or packages. Using badges, we allow for ease of recognition through verification from official domains and sources to hinder phishing attempts and indicate which conversations one is following, grouping flagging and favorites.
Preview of Today's weather view in Ellsi
Another view is one to display the weather as shown here in the Today’s weather view reachable from expansion of the introductory weather card in the conversation interface. In this view, the user has more control as per the ability to shift through the day, confirm or deny weather preference, or do more with the app based on this information.
Preview of weekly weather view in Ellsi
In this forecast view, we designed a different way of visually delivering the weekly weather. In what our research, including usability tests, were able to observe what was most quickly interpretable and recognizable. The process of reading numbers involves sight and recognition and then pairing that recognition with intelligence. While this happens quickly mentally, there seems to be less of a cognitive load in recognizing colors and distance that does not require math and relies on visually recognizable information/memories. Similar to seeing and recognizing something is hot v.s. seeing, reading, understanding numbers and their meaning/correlation to weather. Therefore, we used shades of warmer colors to represent heat and cooler colors to represent the opposite as well as fluctuations in length to represent temperature and precipitation.

Read more on Ellsi's website.