Researched how AI assisted navigation can improve night time safety for vulnerable riders by conducting interviews and field observation.

Researched how AI assisted navigation can improve night time safety for vulnerable riders by conducting interviews and field observation.

Researched how AI assisted navigation can improve night time safety for vulnerable riders by conducting interviews and field observation.

The Late Night Journey Home

The Late Night Journey Home

The Late Night Journey Home

Every weekend, millions of people face the same dilemma: how to get home safely after a

night out. Whether tired, tipsy, or simply disoriented in unfamiliar areas, they make critical

navigation decisions when their judgment is compromised. This vulnerable state leads

to poor choices, unsafe situations, and preventable accidents.

Every weekend, millions of people face the same dilemma: how to get home safely after a

night out. Whether tired, tipsy, or simply disoriented in unfamiliar areas, they make critical

navigation decisions when their judgment is compromised. This vulnerable state leads

to poor choices, unsafe situations, and preventable accidents.

Every weekend, millions of people face the same dilemma: how to get home safely after a

night out. Whether tired, tipsy, or simply disoriented in unfamiliar areas, they make critical navigation decisions when their judgment is compromised. This vulnerable state leads to poor choices, unsafe situations, and preventable accidents.

37%

37%

37%

of pedestrian accidents occur

between 9pm-3am


of pedestrian

accidents occur

between 9pm-3am


of pedestrian

accidents occur

between 9pm-3am


2.3M

2.3M

2.3M

ride-share requests cancelled

due to confusion

ride-share

requests cancelled

due to confusion

ride-share

requests cancelled

due to confusion

68%

68%

68%

of users report anxiety about

late-night navigation

of users report

anxiety about

late-night navigation

of users report

anxiety about

late-night navigation

Impaired Vision Focus

Impaired Vision Focus

Impaired Vision Focus

Users struggle to read small text and process complex navigation instructions when tired or affected by alcohol, leading to wrong turns and missed stops.


Users struggle to read small text and process complex navigation instructions when tired or affected by alcohol, leading to wrong turns and missed stops.


Users struggle to read small text and process complex navigation instructions when tired or affected by alcohol, leading to wrong turns and missed stops.


Decision Paralysis

Decision Paralysis

Decision Paralysis

Standing at intersections unsure which way to go, users second-guess themselves and waste time, increasing exposure to unsafe situations and cold weather.


Standing at intersections unsure which way to go, users second-guess themselves and waste time, increasing exposure to unsafe situations and cold weather.


Standing at intersections unsure which way to go, users second-guess themselves and waste time, increasing exposure to unsafe situations and cold weather.


Unsafe Route Selection

Unsafe Route Selection

Unsafe Route Selection

Without proper awareness, users choose poorly lit or isolated routes that put them at risk, especially in unfamiliar neighbourhoods late at night.


Without proper awareness, users choose poorly lit or isolated routes that put them at risk, especially in unfamiliar neighbourhoods late at night.


Without proper awareness, users choose poorly lit or isolated routes that put them at risk, especially in unfamiliar neighbourhoods late at night.


Interface Fumbling

Interface Fumbling

Interface Fumbling

Loss of balance and coordination makes it difficult to interact with complex interfaces, leading to cancelled rides, wrong addresses, and frustration.


Loss of balance and coordination makes it difficult to interact with complex interfaces, leading to cancelled rides, wrong addresses, and frustration.


Loss of balance and coordination makes it difficult to interact with complex interfaces, leading to cancelled rides, wrong addresses, and frustration.


When good judgment is compromised,

small UI friction becomes dangerous.

When good judgment is compromised,

small UI friction becomes dangerous.

When good judgment is compromised,

small UI friction becomes dangerous.

Understanding the Vulnerable User

Understanding the Vulnerable User

Understanding the Vulnerable User

Typical Late-Night Journey: Confidence Deterioration

Typical Late-Night Journey: Confidence Deterioration

Cognitive Impairment Factors

Cognitive Impairment Factors

Visual Processing

Visual Processing

Visual Processing

Blurred vision and depth perception issues.

Difficulty reading small text on screens.

Reduced contrast sensitivity in low light.

Eye strain from screen brightness.

Blurred vision and depth perception issues.

Difficulty reading small text on screens.

Reduced contrast sensitivity in low light.

Eye strain from screen brightness.

Blurred vision and depth perception issues.

Difficulty reading small text on screens.

Reduced contrast sensitivity in low light.

Eye strain from screen brightness.

Decision Making

Decision Making

Decision Making

Analysis paralysis at route choices.

Risk assessment failure in dark areas.

Impulsive choices without full consideration.

Delayed emergency response thinking.

Analysis paralysis at route choices.

Risk assessment failure in dark areas.

Impulsive choices without full consideration.

Delayed emergency response thinking.

Analysis paralysis at route choices.

Risk assessment failure in dark areas.

Impulsive choices without full consideration.

Delayed emergency response thinking.

Motor Skills

Motor Skills

Motor Skills

Typing errors and autocorrect dependency.

Missed taps on small UI elements.

Balance issues affecting phone use.

Difficulty with multi-step interactions.

Typing errors and autocorrect dependency.

Missed taps on small UI elements.

Balance issues affecting phone use.

Difficulty with multi-step interactions.

Typing errors and autocorrect dependency.

Missed taps on small UI elements.

Balance issues affecting phone use.

Difficulty with multi-step interactions.

Critical Failure Moments

Critical Failure Moments

The Frozen Decision

The Frozen Decision

The Frozen Decision

Users stand at intersections paralyzed by multiple route options, unable to process trade-offs between time, safety, and distance. Decision fatigue compounds with cognitive impairment, leading to prolonged exposure to elements and increased vulnerability.


Users stand at intersections paralyzed by multiple route options, unable to process trade-offs between time, safety, and distance. Decision fatigue compounds with cognitive impairment, leading to prolonged exposure to elements and increased vulnerability.


Users stand at intersections paralyzed by multiple route options, unable to process trade-offs between time, safety, and distance. Decision fatigue compounds with cognitive impairment, leading to prolonged exposure to elements and increased vulnerability.


The Safety Gamble

The Safety Gamble

The Safety Gamble

Tired users consistently underestimate danger and overvalue convenience. The temptation to save 10 minutes by cutting through poorly lit areas overrides safety instincts. Impaired risk assessment leads to choosing isolated routes that seem "fine" at the moment.


Tired users consistently underestimate danger and overvalue convenience. The temptation to save 10 minutes by cutting through poorly lit areas overrides safety instincts. Impaired risk assessment leads to choosing isolated routes that seem "fine" at the moment.


Tired users consistently underestimate danger and overvalue convenience. The temptation to save 10 minutes by cutting through poorly lit areas overrides safety instincts. Impaired risk assessment leads to choosing isolated routes that seem "fine" at the moment.


The Help Hesitation

The Help Hesitation

The Help Hesitation

Users delay calling friends or emergency contacts, not wanting to be a burden or cause worry. This hesitation means help arrives later when situations deteriorate. Social anxiety about appearing "too drunk" or "overreacting" prevents timely safety interventions.


Users delay calling friends or emergency contacts, not wanting to be a burden or cause worry. This hesitation means help arrives later when situations deteriorate. Social anxiety about appearing "too drunk" or "overreacting" prevents timely safety interventions.


Users delay calling friends or emergency contacts, not wanting to be a burden or cause worry. This hesitation means help arrives later when situations deteriorate. Social anxiety about appearing "too drunk" or "overreacting" prevents timely safety interventions.


Where Existing Solutions Fail

Google Maps

Google Maps

Google Maps

Requires users to compare multiple route options and make complex trade-off decisions. Small text and detailed information overwhelm impaired users. No proactive safety awareness or assistance.


Requires users to compare multiple route options and make complex trade-off decisions. Small text and detailed information overwhelm impaired users. No proactive safety awareness or assistance.


Requires users to compare multiple route options and make complex trade-off decisions. Small text and detailed information overwhelm impaired users. No proactive safety awareness or assistance.


Uber/Ride-Share Apps

Uber/Ride-Share Apps

Uber/Ride-Share Apps

Requires manual address entry and typing, which is difficult with impaired motor skills. No AI assistance to detect confusion or automatically help. Users must actively request a ride rather than being guided to the safest option.

Requires manual address entry and typing, which is difficult with impaired motor skills. No AI assistance to detect confusion or automatically help. Users must actively request a ride rather than being guided to the safest option.

Requires manual address entry and typing, which is difficult with impaired motor skills. No AI assistance to detect confusion or automatically help. Users must actively request a ride rather than being guided to the safest option.

Generic Safety Apps

Generic Safety Apps

Generic Safety Apps

Requires multiple deliberate taps to activate emergency features. "Hold for 3 seconds" is too long when judgment is impaired. No contextual awareness of user's state or automatic intervention entirely reactive rather than proactive.



Requires multiple deliberate taps to activate emergency features. "Hold for 3 seconds" is too long when judgment is impaired. No contextual awareness of user's state or automatic intervention entirely reactive rather than proactive.



Requires multiple deliberate taps to activate emergency features. "Hold for 3 seconds" is too long when judgment is impaired. No contextual awareness of user's state or automatic intervention entirely reactive rather than proactive.



Understanding trust, safety, and autonomy

in vulnerable states of individuals

Understanding trust, safety, and autonomy

in vulnerable states of individuals

Understanding trust, safety, and autonomy

in vulnerable states of individuals

User Interviews

User Interviews

User Interviews

24 participants • Ages 21-35

24 participants • Ages 21-35

24 participants • Ages 21-35

In-depth conversations with regular nightlife goers and occasional users to understand their navigation challenges, safety concerns, and decision-making patterns when tired or tipsy.


In-depth conversations with regular nightlife goers and occasional users to understand their navigation challenges, safety concerns, and decision-making patterns when tired or tipsy.


In-depth conversations with regular nightlife goers and occasional users to understand their navigation challenges, safety concerns, and decision-making patterns when tired or tipsy.


Behavioural Studies

Behavioural Studies

Behavioural Studies

Simulated environments: Real-time tracking.

Simulated environments: Real-time tracking.

Simulated environments: Real-time tracking.

Observation of navigation behaviour in controlled scenarios simulating tired or tipsy states. Tracked decision delays, route confusion, and risk assessment capabilities.


Observation of navigation behaviour in controlled scenarios simulating tired or tipsy states. Tracked decision delays, route confusion, and risk assessment capabilities.


Observation of navigation behaviour in controlled scenarios simulating tired or tipsy states. Tracked decision delays, route confusion, and risk assessment capabilities.


Safety Analysis

Safety Analysis

Safety Analysis

9pm-4am data: Urban areas

9pm-4am data: Urban areas

9pm-4am data: Urban areas

Analysis of crime statistics, accident reports, and ride-share patterns during late-night hours to identify high-risk areas and understand existing safety infrastructure.


Analysis of crime statistics, accident reports, and ride-share patterns during late-night hours to identify high-risk areas and understand existing safety infrastructure.


Analysis of crime statistics, accident reports, and ride-share patterns during late-night hours to identify high-risk areas and understand existing safety infrastructure.


Research questions for participants

Research questions for participants

Research questions for participants

How do impaired users currently navigate?


How much control vs automation do they want?


How do they feel about location tracking?

How do impaired users currently navigate?


How much control vs automation do they want?


How do they feel about location tracking?

How do impaired users currently navigate?


How much control vs automation do they want?


How do they feel about location tracking?

What safety decisions do they struggle with?


What builds trust in AI decision-making?


What constitutes a 'safe' route?

What safety decisions do they struggle with?


What builds trust in AI decision-making?


What constitutes a 'safe' route?

What safety decisions do they struggle with?


What builds trust in AI decision-making?


What constitutes a 'safe' route?

Characteristic values of our interviewees/participants

Characteristic values of our interviewees/participants

Samuel Okonkwo

26 • Marketing Manager • Urban

I always second-guess myself at night. I want help but don't want to give up control completely.

James Ebong

23 • Graduate Student • Suburban

After a few drinks, I can't tell if I'm making the right call about walking vs. getting a ride.

Stella Maris

29 • Software Engineer • Urban

I trust AI for recommendations, but I need to see WHY it's suggesting something before I commit.

Juliet Okon

31 • Teacher • Urban

My biggest fear is my phone dying when I need it most. I need something reliable and predictable.


Mary Ukpi

24 • Graphic Designer • Suburban

I'd love an app that just handles everything when I'm too tired to think, but only if I can override it.

Obinna Ezenandu

28 • Nurse • Urban

Working night shifts, I'm often exhausted. Safety isn't just about crime it's about making smart choices when you're not at your best.

Samuel Okonkwo

26 • Marketing Manager • Urban

I always second-guess myself at night. I want help but don't want to give up control completely.

Stella Maris

29 • Software Engineer • Urban

I trust AI for recommendations, but I need to see WHY it's suggesting something before I commit.

Mary Ukpi

24 • Graphic Designer • Suburban

I'd love an app that just handles everything when I'm too tired to think, but only if I can override it.

James Ebong

23 • Graduate Student • Suburban

After a few drinks, I can't tell if I'm making the right call about walking vs. getting a ride.

Juliet Okon

31 • Teacher • Urban

My biggest fear is my phone dying when I need it most. I need something reliable and predictable.


Obinna Ezenandu

28 • Nurse • Urban

Working night shifts, I'm often exhausted. Safety isn't just about crime it's about making smart choices when you're not at your best.

Studying 47 late-night riders revealed how

fatigue shapes safety choices and autonomy.

Studying 47 late-night riders revealed how

fatigue shapes safety choices and autonomy.

Studying 47 late-night riders revealed how

fatigue shapes safety choices and autonomy.

1

Users want safety assistance

but fear losing autonomy.


While 78% wanted AI help during vulnerable moments,

82% wanted ability to override decisions. Trust requires

transparency and control.

While 78% wanted AI help during vulnerable moments, 82% wanted ability to override decisions. Trust requires

transparency and control.

While 78% wanted AI help during vulnerable moments, 82% wanted ability to override decisions. Trust requires

transparency and control.

Data Privacy

95%

Transparency in AI Reasoning

92%

Human Backup Option

87%

Clear AI Decisions Logic

84%

Emergency Contacts Notified

71%

2

Cognitive load increases dramatically in compromised states


Cognitive load increases dramatically in compromised states


Cognitive load increases dramatically in compromised states


Decision time for simple navigation choices increased by 675% when tipsy compared to baseline, revealing severe impairment in basic wayfinding tasks.

Where Users Get Stuck in Current Apps

Route Selection

Route Selection

Route Selection

87%

87%

87%

Address Entry

Address Entry

Address Entry

74%

74%

74%

Help Activation

Help Activation

Help Activation

62%

62%

62%

Payment Confirm

Payment Confirm

Payment Confirm

41%

41%

41%

Decision Time (Secs)

4s

Sober

18s

Tired

31s

Tipsy

3

Safety is personal and context-dependent



Users defined "safe route" differently based on personal experiences and priorities. A one-size-fits-all approach would fail 68% of users.


Route personalization factors: Lighting quality (89%), Population density (76%), Crime statistics (71%), Familiarity with area (83%), Distance from transit stops (54%), Presence of cameras/security (67%)


Route personalization factors: Lighting quality (89%), Population density (76%), Crime statistics (71%), Familiarity with area (83%), Distance from transit stops (54%), Presence of cameras/security (67%)


“I want well-lit streets even if longer ”


“I want well-lit streets even if longer ”


“I want fastest route to minimize exposure”

“I want fastest route to minimize exposure”

“I want populated areas where help is nearby”


“I want populated areas where help is nearby”


Willingness to Share Location or Use AI-Assisted

Navigation between Africans & Westerns

Willingness to Share Location or Use AI-Assisted

Navigation between Africans & Westerns

Willingness to Share Location or Use AI-Assisted Navigation between Africans & Westerns

Patterns Behind Night time Behaviour

Patterns Behind Night time Behaviour

Patterns Behind Night time Behaviour

97%

97%

97%

Prefer voice navigation when coordination is impaired.

73%

73%

73%

Forget to share location with friends before heading home.

64%

64%

64%

Make riskier choices when tired or impaired.

89%

89%

89%

Want automatic check-in with trusted contacts.

56%

56%

56%

Have gotten lost or confused returning home at night.

82%

82%

82%

Feel anxious about safety during late-night travel.

By mapping friction points, we saw how mental

fatigue leads to chain navigation failures.

By mapping friction points, we saw how mental

fatigue leads to chain navigation failures.

By mapping friction points, we saw how mental fatigue leads to chain navigation failures.

Can't type address accurately



Motor Skills Impaired

Can't evaluate route safety



Judgment Impaired


Forget to activate safety features



Planning Impaired


Can't explain location clearly


Communication Impaired


Ensuring every vulnerable user reaches

home safely through AI assistance .

Ensuring every vulnerable user reaches

home safely through AI assistance .

Proactive Safety

Proactive Safety

AI takes action before problems escalate, monitoring behaviour patterns and environmental factors to intervene at the right moment.

AI takes action before problems escalate, monitoring behaviour patterns and environmental factors to intervene at the right moment.

Transparent Logic

Transparent Logic

Shows why decisions are made, building trust through clear explanations of AI reasoning and safety recommendations.

Shows why decisions are made, building trust through clear explanations of AI reasoning and safety recommendations.

User Override

User Override

Always gives final control, respecting user autonomy while providing intelligent suggestions and safety guardrails.

Always gives final control, respecting user autonomy while providing intelligent suggestions and safety guardrails.

Building the AI Apps (MoonRide)

structure with observed data

Building the AI Apps (MoonRide)

structure with observed data

Building Trust Through Onboarding

Building Trust Through Onboarding

Luna adapts to your state in real-time

Luna adapts to your state in real-time

Luna's Contextual Awareness States

Luna's Contextual Awareness States

Normal Monitoring

Subtle blue glow indicates passive monitoring. Luna tracks journey progress and updates ETA without interrupting.

Attention State

Amber glow when unusual behavior detected (stopped moving, off-route, hesitant navigation). Luna asks gentle questions.


Alert State

Red glow for safety concerns (extended unresponsiveness, unsafe area after hours, distress signals). Luna takes proactive action.


Luna's Contextual Awareness in Action

Luna's Contextual Awareness in Action

Luna's Contextual Awareness in Action

Luna acts proactively, but you're always in control.

Luna acts proactively, but you're always in control.

Luna acts proactively, but you're always in control.

Luna's interventions can be stopped or reversed at any time

Luna's interventions can be stopped or reversed at any time

Luna's interventions can be stopped or reversed at any time

Instant Cancel

Instant Cancel

Instant Cancel

Every AI action displays a prominent cancel button.

One tap stops the intervention immediately.

Every AI action displays a prominent cancel button.

One tap stops the intervention immediately.

Every AI action displays a prominent cancel button.

One tap stops the intervention immediately.

30-Second Undo

30-Second Undo

30-Second Undo

Even after an action completes, you have 30 seconds to undo it and reverse the decision.

Even after an action completes, you have 30 seconds to undo it and reverse the decision.

Even after an action completes, you have 30 seconds to undo it and reverse the decision.

Luna Learns

Luna Learns

Luna Learns

When you override Luna's decisions, she learns your preferences to make better suggestions next time.

When you override Luna's decisions, she learns your preferences to make better suggestions next time.

When you override Luna's decisions, she learns your preferences to make better suggestions next time.

MoonRide gives you full control over Luna's autonomy level

MoonRide gives you full control over Luna's autonomy level

MoonRide gives you full control over Luna's autonomy level

AI Auto-Actions

AI Auto-Actions

AI Auto-Actions

Control whether Luna takes automatic

actions or just provides suggestions.

Control whether Luna takes automatic actions or just provides suggestions.

Auto Mode: Luna acts automatically when detecting safety concerns

Manual Mode: Luna only suggests - you choose every action

Switch anytime: Change modes instantly based on how you feel

Auto Mode

Luna takes proactive safety actions

Faster response in emergencies

Minimal cognitive load for user

All actions can be cancelled

Best for vulnerable states

Auto Mode

Luna takes proactive safety actions

Faster response in emergencies

Minimal cognitive load for user

All actions can be cancelled

Best for vulnerable states

Toggle the mode based on your current state. Feeling tired? Switch to Auto.

Want more control? Choose Manual. The choice is always yours.

Scenario: Luma automatically ordered a ride home

Scenario: Luma automatically ordered a ride home

Scenario: Luma automatically ordered a ride home

Movement Pattern

Erratic & Stopping

Erratic & Stopping

Erratic & Stopping

You stopped 4 times in 3 minutes, suggesting difficulty walking steadily.

Time of Day

2:14 AM

2:14 AM

2:14 AM

Late hour increases vulnerability and reduces available support options

User Preferences

Auto-assist enabled

Auto-assist enabled

Auto-assist enabled

You've enabled automatic ride calling for late night difficulty

Location Safety

Medium Risk (Score: 58)

Medium Risk (Score: 58)

Medium Risk (Score: 58)

Area has moderate crime rate and poor lighting after midnight

Distance to Home

1.2 miles (22 min walk)

1.2 miles (22 min walk)

1.2 miles (22 min walk)

Distance requires sustained walking ability in current state

Past Behaviour

Similar pattern 3x before

Similar pattern 3x before

Similar pattern 3x before

In similar situations, you previously accepted ride suggestions

Luna's Decision Logic

Luna's Decision Logic

Luna's Decision Logic

Based on erratic movement + late hour + medium-risk location + your preferences + past acceptance patterns, ordering a ride provides the safest outcome with 94% confidence.


Based on erratic movement + late hour + medium-risk location + your preferences + past acceptance patterns, ordering a ride provides the safest outcome with 94% confidence.


AI Transparency Commitment

AI Transparency Commitment

AI Transparency Commitment

Every AI decision Luna makes can be expanded to show the complete reasoning process, data sources, and confidence levels. We believe transparent AI builds trust and helps users understand how their safety companion works.


Every AI decision Luna makes can be expanded to show the complete reasoning process, data sources, and confidence levels. We believe transparent AI builds trust and helps users understand how their safety companion works.


Testing assumptions through controlled experiments

Testing assumptions through controlled experiments

Testing assumptions through controlled experiments

Testing Methodology

We conducted rigorous A/B testing with 120 participants in a controlled environment simulating tired and tipsy states. Each participant completed navigation tasks using both manual and AI-assisted interfaces. We measured three key metrics: task completion time, navigation error rate, and subjective safety confidence (1-10 scale). Participants were randomly assigned starting conditions and asked to rate their preference after experiencing both variants.


We conducted rigorous A/B testing with 120 participants in a controlled environment simulating tired and tipsy states. Each participant completed navigation tasks using both manual and AI-assisted interfaces. We measured three key metrics: task completion time, navigation error rate, and subjective safety confidence (1-10 scale). Participants were randomly assigned starting conditions and asked to rate their preference after experiencing both variants.


We conducted rigorous A/B testing with 120 participants in a controlled environment simulating tired and tipsy states. Each participant completed navigation tasks using both manual and AI-assisted interfaces. We measured three key metrics: task completion time, navigation error rate, and subjective safety confidence (1-10 scale). Participants were randomly assigned starting conditions and asked to rate their preference after experiencing both variants.


Test 1: Manual vs AI-Assisted Navigation

Test 1: Manual vs AI-Assisted Navigation

Test 1: Manual vs AI-Assisted Navigation

Manual Navigation


Traditional navigation interface requiring active user decision-making

at multiple points throughout the journey.

Manual route selection: User must review and choose from multiple route options

Active safety monitoring: User responsible for watching surroundings and identifying risks

Manual intervention: User decides when to call for help or alternative transport

Data entry required: User manually enters ride-share details and emergency contacts

Decision Point 1:

Select route from 3 options (fastest, safest, or scenic)

Decision Point 2:

Monitor journey progress and safety indicators

Decision Point 3:

Notice potential issue and decide on action

Decision Point 4:

Open ride-share app or call contact manually

Decision Point 5:

Enter destination and confirm booking

A

AI-Assisted (MoonRide)


Intelligent navigation with Luna AI proactively managing safety decisions and reducing cognitive load for impaired users.

AI route recommendation: Luna automatically selects safest route with clear explanation

Proactive monitoring: AI continuously assesses journey safety and user state

Predictive intervention: Luna suggests help before problems escalate

Auto-filled information: Pre-populated ride details and emergency contacts ready to go

Luna Decision:

Selected safest route (well-lit, 2 min longer) - User notified

Luna Decision:

Monitoring journey - all systems normal

Luna Alert:

"You've stopped moving for 3 minutes. Would you like me to call Sarah?"

User Override Available:

Tap to approve Luna's suggestion or choose different action

Luna Action:

Uber pre-ordered to current location, Sarah notified automatically

B

Winner

Test Results: Variant B Outperformed Across All Metrics

Test Results: Variant B Outperformed Across All Metrics

Task completion Time

Manual (A)

4.2

min

AI-Assisted (B)

2.1

min

50.0%

Faster completion

Navigation Errors (avg per session)

Manual (A)

3.7

errors

AI-Assisted (B)

0.9

errors

76.0%

error reduction

Subjective Safety Confidence (1-10 scale)

Manual (A)

6.2

/10

AI-Assisted (B)

8.9

/10

43.0%

confidence increase

Task completion Time

Preferred AI-Assisted (B)

84%

Preferred Manual (A)

16%

50.0%

Faster completion

Key Finding: AI assistance reduced cognitive load significantly, especially as simulated impairment increased. Users reported feeling "taken care of" rather than "having to manage" their journey. The reduction in decision points correlated directly with decreased stress levels and improved safety outcomes.


Key Finding: AI assistance reduced cognitive load significantly, especially as simulated impairment increased. Users reported feeling "taken care of" rather than "having to manage" their journey. The reduction in decision points correlated directly with decreased stress levels and improved safety outcomes.


“I didn't have to think, Luna just handled it. That's exactly what I need when I'm tired and just want to get home safely. ”


“"I felt safer knowing something was watching out for me. It's like having a responsible friend who stays sober.”


"I liked that I could override if I wanted to. Luna makes suggestions but I'm still in control. That balance is perfect."


Test 2: Finding the balance between helpful and annoying

Test 2: Finding the balance between helpful and annoying

Test 2: Finding the balance between helpful and annoying

Voice Heavy

A

Luna speaks frequently with voice prompts every 30 seconds on average. Constant reassurance and guidance.

"Turn left in 100 feet"

"You're doing great"

"Still 8 minutes from home"

"I'm here if you need me"

Average: One prompt every 30 seconds

Silent Notifications

B

Minimal audio. Visual-only notifications appear as cards. Luna only speaks for critical alerts or when user initiates.

AI route recommendation: Luna automatically selects safest route with clear explanation

Proactive monitoring: AI continuously assesses journey safety and user state

Predictive intervention: Luna suggests help before problems escalate

Auto-filled information: Pre-populated ride details and emergency contacts ready to go

Voice only for critical alerts

Adaptative

C

Winner

Luna speaks frequently with voice prompts every 30 seconds on average. Constant reassurance and guidance.

Luna is in quiet mode

You're navigating smoothly

You're navigating smoothly

Adapts to user needs in real-time.

Test Results helpful and annoying analysis on individuals

Test Results helpful and annoying analysis on individuals

Annoyance Rating (Lower is better)

Variant A

6.8/10

Variant B

3.2/10

Variant C

2.1/10>

Appropriate Timing

Variant A

4.9/10

Variant B

6.110

Variant C

9,2/10>

Reassurance Feeling

Variant A

8.1/10

Variant B

5.4/10

Variant C

8.7/10>

User Preference

Variant A

12%

Variant B

18%

Variant C

70%>

Annoyance Rating (Lower is better)

Variant A

6.8/10

Variant B

3.2/10

Variant C

2.1/10>

Reassurance Feeling

Variant A

8.1/10

Variant B

5.4/10

Variant C

8.7/10>

Appropriate Timing

Variant A

4.9/10

Variant B

6.110

Variant C

9,2/10>

User Preference

Variant A

12%

Variant B

18%

Variant C

70%>

Users wanted to feel Luna's presence without constant interruption. Adaptive communication provided reassurance during uncertain moments while staying quiet during smooth navigation.



"It was like Luna knew exactly when I needed to hear her voice and when I just needed to follow the line on the map."


— Study Participant, Test Group C

A world where no one has to worry about getting home safely, where technology acts as a guardian angel, anticipating needs and taking action to protect those in vulnerable moments.


A world where no one has to worry about getting home safely, where technology acts as a guardian angel, anticipating needs and taking action to protect those in vulnerable moments.


A world where no one has to worry about getting home safely, where technology acts as a guardian angel, anticipating needs and taking action to protect those in vulnerable moments.


Feature Pipeline/what to expect in the future

Feature Pipeline/what to expect in the future

Feature Pipeline/what to expect in the future

AR Wayfinding

Augmented reality navigation for clearer guidance in complex urban environments



Wearable Integration

Apple Watch and Fitbit support for completely hands-free operation and health monitoring


Public Transit Integration

Comprehensive multi-modal routing including buses, trains, and micro-mobility options


Community Safety Reports

Real-time crowdsourced safety data for dynamic risk assessment and routing



Health Monitoring

Advanced biometric tracking for more accurate state detection and personalized safety thresholds


Accessibility Features

Specialized support for visually and hearing impaired users with adaptive interfaces


AR Wayfinding

Augmented reality navigation for clearer guidance in complex urban environments



Wearable Integration

Apple Watch and Fitbit support for completely hands-free operation and health monitoring


Public Transit Integration

Comprehensive multi-modal routing including buses, trains, and micro-mobility options


Community Safety Reports

Real-time crowdsourced safety data for dynamic risk assessment and routing



Health Monitoring

Advanced biometric tracking for more accurate state detection and personalized safety thresholds


Accessibility Features

Specialized support for visually and hearing impaired users with adaptive interfaces


AR Wayfinding

Augmented reality navigation for clearer guidance in complex urban environments



Community Safety Reports

Real-time crowdsourced safety data for dynamic risk assessment and routing



Wearable Integration

Apple Watch and Fitbit support for completely hands-free operation and health monitoring


Health Monitoring

Advanced biometric tracking for more accurate state detection and personalized safety thresholds


Public Transit Integration

Comprehensive multi-modal routing including buses, trains, and micro-mobility options


Accessibility Features

Specialized support for visually and hearing impaired users with adaptive interfaces


Strategic Partnerships to ensure all roundedness in performance.

Strategic Partnerships to ensure all roundedness in performance.

Strategic Partnerships to ensure all roundedness in performance.

Uber / Lyft

Seamless ride booking integration for instant safety options


Cities

Access to real-time safety data and smart city infrastructure


Health Services

Direct emergency response coordination for critical situations


Universities

Campus safety programs protecting students during late-night activities


Venues / Bars

Patron safety initiatives ensuring responsible service and safe departures

Uber / Lyft

Seamless ride booking integration for instant safety options


Health Services

Direct emergency response coordination for critical situations


Venues / Bars

Patron safety initiatives ensuring responsible service and safe departures

Cities

Access to real-time safety data and smart city infrastructure


Universities

Campus safety programs protecting students during late-night activities


We envision a future where AI-assisted safety is as commonplace

as GPS navigation, where every night-time journey is protected by

intelligent, compassionate technology that puts human wellbeing first.

We envision a future where AI-assisted safety is as commonplace as GPS navigation, where every night-time journey is protected by intelligent, compassionate technology that puts human wellbeing first.

We envision a future where AI-assisted safety is as commonplace as GPS navigation, where every night-time journey is protected by intelligent, compassionate technology that puts human wellbeing first.

1M

1M

1M

Safe Trips/Month

50K

50K

50K

Emergencies Prevented

100+

100+

100+

Cities Worldwide

Let's Connect

linkedin.com/in/gloriachukwu

linkedin.com/in/gloriachukwu

linkedin.com/in/gloriachukwu

gloriachukwu566@gmail.com

gloriachukwu566@gmail.com

gloriachukwu566@gmail.com

GLORIA_ONYEKERE_CASSIDY

GLORIA_ONYEKERE_CASSIDY

GLORIA_ONYEKERE_CASSIDY

More of my works

Designed and developed by Your Gloria Chukwu • 2025 • Ebonyi, Nigeria

Create a free website with Framer, the website builder loved by startups, designers and agencies.