Skip to main content
← Lab

Learning technology · Concept and paper prototypes · 2024 to now

Focus

In 2024 I designed Focus, an AI-integrated knowledge management system meant to increase memory retention and facilitate human cognition. It starts as a reader. Everything you read and add stays in it, it builds its own model of that material, and AI helps you find the connections between ideas.

I designed and tested Focus in two UX courses at Kent State in fall 2024. Claude drafted this page from those assignments, and I reviewed it.

2024first sketches
22paper panels
2test sessions
5stages mapped
Hand-drawn storyboard: the learner, distracted by the inside and outside world, engages Focus mode, notifications and noise are eliminated, and the user retains more information
October 2024. The first storyboard: a distracted learner turns on Focus and keeps more of what they read.

01 · The problem

Reading competes with everything else on the phone.

Focus began as an assignment to design a better e-reader. The first thing I drew was the problem. A page of text sits on a phone that never stops asking for attention: banners, sounds, and then the room itself.

I split distraction into two kinds. Focus 1 handles what comes from inside the device and silences other apps. Focus 2 handles what comes from outside, with music, binaural tones, and noise cancellation.

Sketch titled Focus Mode Level Two: a reader with Focus buttons beside the text, and a drawing of a person reading while speech bubbles shout Honk and Blah around them
Focus Mode Level Two: distraction also comes from the outside world.
Six-panel storyboard: a reading page, the same page buried in dings, beeps, and chatter, a swipe up to reveal controls, the Focus levels appearing, a tap, and a quiet page with a smiling reader
The interaction in six frames: interruptions arrive, the reader swipes up, taps a level, and the page goes quiet.

02 · First version

One tap for help, one long press for control.

Sketch of single-click Focus buttons: Focus 1 silences all phone notifications, Focus 2 starts binaural beats and noise cancelling, and a press and hold opens finer adjustments
Single-tap buttons block distractions; press and hold opens the finer settings.

The first prototype gave every reader the same simple entry point. A beginner taps Focus 1 or Focus 2 and keeps reading. Someone who wants more can press and hold to choose the music, the binaural tones, and how much noise cancellation to apply.

That pairing has stayed with me. A tool should give a newcomer a clear way in and give an experienced person exact control, without making either one wade through the other's interface.

Eight rough phone screens sketched in marker: reading pages, a stats sidebar, a progress view, and an options screen listing Focus 1, Focus 2, Focus 3, and a fourth Focus option
Screen sketches from the first round, including an options screen with three Focus levels.

03 · Paper prototype, November 2024

22 panels, two people, a 50/50 split.

Paper prototype laid out on a table: a quiet phone frame, a loud phone frame surrounded by drawn speech bubbles, sticky notes, and six sheets of hand-drawn reader screens
The full prototype: a quiet frame, a loud frame, sticky-note menus, and the screens I swapped in by hand.

I built an iPhone reader out of paper and tested it with two people. The first scrolled down through pages and never tried to swipe. The second swiped and never tried to scroll. On the most basic action in a reader, my two participants split down the middle.

Press and hold to open the Focus options confused the first participant, so I changed it to a single tap before the second session. The second participant learned the app by tapping everything, which showed me the prototype gave too little feedback when a function started or stopped. I drew a menu panel, a stats panel, and a navigation panel between sessions because I expected her to explore, and she did.

  • Paper phone screen with three notification banners stacked over the reading page
    Notifications from other apps pile up over the page.
  • Paper phone screen with a small stack of Focus levels rising from the Focus button
    One tap on Focus raises the levels of help.
  • The same Focus stack drawn inside a frame of speech bubbles reading Honk, Woof, and Blah
    The loud frame stood in for noise in the room.
  • Paper phone screen with a Focus menu offering type of music, binaural tones, and noise cancelling
    The Focus menu: music, binaural tones, noise cancellation.
  • Paper phone screen with a stats stack listing time, pages, words, percent complete, and all time
    A stats panel showing the reader their own progress.

04 · Mapping years of use

From a reader to a PhD.

In the second course I mapped how one person would use Focus over several years. The persona was Mikaela, a bilingual graduate student who works full time in research. She starts with the reader app, begins linking her notes, learns she can ask questions of everything she has read, and eventually decides she has built enough knowledge to pursue a PhD.

The map was written for investors whose original brief was a device that could measurably increase reading comprehension. I wanted to show them a system that could do that and keep growing with the person using it.

Experience map with five stages: initial discovery, learning the Focus Reader app, expanding into the Focus ecosystem, buying the Focus Reader device, and fully integrated. A curve of speech bubbles rises through moments such as I can ask it questions about my own materials I have read
The experience map. The peak comes when Mikaela sees that everything she has read is linked together. Select the map to open it full size.

05 · What happened next

I built the memory first.

I did not build Focus. Over the next two years I built Clarence, the AI system I work with every day. It keeps what I give it, links it, and answers questions about my own material. In practice it is a version of Focus for one user.

What Clarence does not do is the part I care about most now. It remembers what I have read. It does not know what I have learned, and it does not decide what I should learn next.

06 · Where it can go

A system that keeps teaching the person using it.

In fall 2026 I am taking a graduate course on AI in education and human performance. The research there names the parts Focus still needs.

  1. A learner model

    The system keeps a running estimate of what the person understands, separate from what they have read. Intelligent tutoring systems work this way, and two meta-analyses, together covering more than 150 comparisons, found they outperform large-group teaching.

  2. Support that fades

    Help a beginner needs gets in an expert's way. The one-tap and long-press controls from 2024 were a first attempt at both modes. The system should move between them as the person improves.

  3. Pacing by mastery

    The person moves on when they have learned something, at the fastest pace that still works for them.

  4. A model the learner can see

    The person should be able to see what the system believes about them and correct it. The stats panel in my paper prototype was an early version of this.

  5. Proof

    The investors in my map wanted measured gains in comprehension. Research on personalized learning makes the same demand: it is the benefit most often claimed and least often shown.

07 · Privacy

The vault belongs to the person.

Marketing, ethics, and design often want different things, because people's goals depend on who signs their checks. If no one can get into a person's vault and the technology never harvests information from them, the privacy problem is solved.