top of page

From data to decision

How progressive disclosure reduced cognitive load during live training.

Enode, Berlin, Germany

image.png

View mobile app

Context

Enode ONE is a B2C mobile app for athletes to track and optimize their training performance. The training tracking screen is the most critical touchpoint in the app. It is where athletes make real-time decisions during live sessions.

Problems

  • High cognitive load

  • Hard to understand features on first use

  • Hard to make decisions during training

  • Users didn't trust AI Recommendations

Deliverable

Redesigned training tracking screen, including revised information architecture, progressive disclosure system, and AI Recommendation transparency layer. The AI trust journey, from onboarding through in-session recommendations, was designed in collaboration with the sports scientist.

Role & Duration

Solo UX/UI Designer, end-to-end redesign of the training tracking flow. Duration: 2 weeks · Mobile-first (iOS)

Grounded in strategy

This work is part of a broader UX strategy I independently defined for Enode across both products, Enode ONE and Enode PRO, aligned with the company's product vision. The strategy was reviewed and approved by the Product Manager. After the first iteration, I stepped back to define a clear strategic direction before going deeper.

UX Strategy. by Neda Mokarami

image.png

0. First iteration

Design thinking before user feedback

When I joined Enode, the training tracking screen was hard to understand and created friction during live sessions. Without formal user feedback yet, I made an initial round of improvements based on my design thinking, reviewed and approved by the team. This first iteration became the foundation for a deeper UX strategy I later defined, which shaped the next round of research and redesign.

Before: App screenshot

After: First iteration

image.png
image.png
image.png
image.png

What I improved in the first iteration:

  1. Visual consistency: color system realigned

  2. Navigation:  bottom nav removed

  3. Progress clarity: IN PROGRESS and COMPLETED badges added

  4. Load control: target indicator on slider

  5. Key metric focus: metric chart inside set cell

  6. AI Recommendation trust: reframed as system suggestion

  7. Workout timer: moved to previous screen

1. Problem discovery

App Store review

In this app: why is the weight target not shown for current set? After setting up exercises it's not clear what weight target is. It just says 80%. It's hard to input the weight you want because you have to click on a bunch of extra stuff to get the keyboard to come up. The slider is useless except for micro adjustments."

1 star · App Store · Jan 17 · Jkcr5y

Three specific pain points from this review:

  1. Weight target not visible for the current set

  2. "80%" means nothing without context, what is 80% of?

  3. Too many taps to enter weight, the slider is useless for anything but micro adjustments

Team feedback

Based on direct conversations with developers who were in contact with users:

  1. Users hesitated during live sessions, didn't know what to do next

  2. AI Recommendations were ignored or dismissed, users didn't understand where they came from

  3. First-time users found the screen overwhelming, too much information competing for attention

Patterns found

Three consistent themes emerged:

image.png

1. Seamlessly integrated new functions into main and sub-menus
2. Consolidated similar content into clear and cohesive categories
3. Standardized button text for consistent user experience

2. Problem framing

From user pain to design direction

The feedback was clear, but feedback alone doesn't create better design. I translated what users experienced into specific UX problems.

  • Where clarity broke down

2. Visual hierarchy, decision first:  Multiple metrics competed equally for attention with no clear priority. For in-session UX, each means something different:

KG = action,

RIR = effort and feeling,

% = system logic.

Only KG and RIR should be immediately visible,% needs context or should be hidden.

image.png

1. Advanced tools, progressive disclosure: Features like focused metric, tracking and distance tools don't belong in the default view. Surface them only when the user needs them, keep the default state calm and focused.

3. Load control, friction at the critical moment:

Adjusting weight required too many taps to reach the keyboard. The slider, designed for precision, was useless for entering a completely different load, forcing users to work around the most important input on the screen.

  • AI trust problem

1. Ambiguous timing, next set or current set?

"Next set" was unclear, users didn't know if the recommendation applied to what they were about to do or something in the future.

image.png

2. Conflicting numbers, what should I actually do?

The recommended load in the card didn't match the target load shown on the slider below, two numbers on the same screen contradicting each other destroyed confidence.

3. Misleading CTA, accept only load?

"Accept load" implied only the KG would be applied, but RIR and reps were also shown. Users didn't know what they were actually accepting.

From insight to action

With the problems clearly framed, I defined a direction before opening Figma. Every design decision that followed was tied back to one of these four stages.

What user said

What I heard

What I decided

What I did

image.png

3. Design decisions

Every change was tied to a specific user problem identified in the research. Nothing was added for its own sake, the goal was a screen that guides athletes through a live session with clarity and confidence.

Calm by default

1. Decision-first hierarchy:

KG is now the dominant metric. RIR provides effort context. % is de-emphasised, visible but not competing for attention.

2. Future sets, hidden by default:

Upcoming sets are collapsed and accessible via a chevron, reducing cognitive load during a live session. Athletes focus on the current set, not what's coming next.

image.png

3. Advanced tools, hidden until needed:

Camera and session insights moved to a floating button at the bottom right, reachable with the thumb, where they're needed most.

Rarely used features are tucked into the settings icon at the top. The default state stays focused on what matters: the next set.

Designing AI that earns trust

Once I identified that users couldn't trust the AI recommendations during training, I collaborated with the sports scientist on the team to extend the trust journey back to onboarding.

Rather than trying to explain the recommendation at the moment it appeared, I designed a foundation earlier, so that by the time users reached the training screen, the logic behind every suggestion was already familiar.

Trust doesn't start at the recommendation.

It starts at step one.

Onboarding, building the contract:

  • Before the first recommendation ever appears, users tell the system their goal, experience level, session duration, and maximal loads.

  • Each screen explains exactly why this data matters. Users understand what the system knows about them, and feel in control of it.

  • "Can be adjusted at any time", this one line does more for trust than any recommendation card ever could.

image.png
image.png
AI recommendation worth trusting

When a recommendation appears during training, users already know where it comes from, because the foundation was built in onboarding. The source is explicit. The action is singular. No conflicting numbers, no ambiguity. AI recommendations are only trusted when users understand the logic behind them, not just the output.

2. Load and reps together

KG and reps shown as a pair, matching how athletes actually think about a set.

image.png

3. Clear CTA:

"Apply" replaces "Accept load", one action, one outcome, no ambiguity.

1. Source made explicit 

"Based on 85% of your current max" replaces the anonymous suggestion. Users now know exactly why this load is recommended.

Guidance on demand

1. Coach context, when needed:

Set instructions are hidden by default and revealed on tap, keeping the screen calm during normal flow.

image.png
Input without friction

1. Keyboard first:

Full numeric keyboard appears immediately, eliminating the extra taps users complained about in the App Store review.

image.png

2. Target always visible:

TARGET ~8KG shown above the slider gives athletes a clear reference point while adjusting.

3. Slider for micro-adjustment :

Slider remains for fine-tuning, but keyboard handles the primary input, each tool doing what it does best.

What I learned

This project pushed me beyond visual problem-solving into behavioural and trust design, understanding not just what users see, but what they feel confident acting on.

What worked

  • Starting from user feedback grounded every decision, nothing was added without a reason

  • Treating AI trust as a journey, not a feature, led to a more coherent product experience

  • Collaborating with the sports scientist early saved significant rework later

What I'd do differently

  • I'd run usability testing on the final design with real athletes, especially the recommendation card and load control interaction

  • I'd explore edge cases for AI recommendations, what happens when the system has no data yet, or when the suggestion is significantly different from what the user expects?

  • I'd design explicitly for moments where AI trust is most fragile — when the athlete consistently overrides suggestions, or when the system conflicts with their own instinct. These patterns reveal where human-AI collaboration needs the most design attention.

Next steps

  • Validate the progressive disclosure system with real session data

  • Extend the AI trust framework to Enode PRO, coaches face similar trust challenges with team-level recommendations

© 2024 by Neda Mokarami

bottom of page