About

WORK

CONTACT

BEHANCE

BEHANCE

LINKEDIN

LINKEDIN

DRIBBBLE

DRIBBBLE

Tej

( *** )
REBUILDING CONFIDENCE IN AI-DRIVEN WORKFLOWS

CHATGPT

TRUSTSCOPE

CHATGPT

TRUSTSCOPE

PROJECT OVERVIEW

In 2025, as AI systems became integral to business decision-making, a critical trust crisis emerged. The "confidence paradox" – where AI presents false information with unwavering certainty – was eroding user confidence and amplifying operational risks across industries. This case study presents TrustScope AI, a comprehensive interface design solution that addresses the $1.8 trillion opportunity to rebuild trust in AI-driven workflows through transparent, human-centered design.

In 2025, as AI systems became integral to business decision-making, a critical trust crisis emerged. The "confidence paradox" – where AI presents false information with unwavering certainty – was eroding user confidence and amplifying operational risks across industries. This case study presents TrustScope AI, a comprehensive interface design solution that addresses the $1.8 trillion opportunity to rebuild trust in AI-driven workflows through transparent, human-centered design.

In 2025, as AI systems became integral to business decision-making, a critical trust crisis emerged. The "confidence paradox" – where AI presents false information with unwavering certainty – was eroding user confidence and amplifying operational risks across industries. This case study presents TrustScope AI, a comprehensive interface design solution that addresses the $1.8 trillion opportunity to rebuild trust in AI-driven workflows through transparent, human-centered design.

ROLE : UX DESIGNER
DURATION : 3 MONTHS
( 01 )
( 01 )
PROBLEM
STATEMENT
PROBLEM
STATEMENT

The Challenge: AI systems are exhibiting a dangerous confidence paradox – they're more likely to assert incorrect statements with confident language ("definitely," "without doubt"), making hallucinations deceptively persuasive. A 2025 MIT study revealed a 34% higher incidence of confident language when models are wrong versus when they are right.

The Challenge: AI systems are exhibiting a dangerous confidence paradox – they're more likely to assert incorrect statements with confident language ("definitely," "without doubt"), making hallucinations deceptively persuasive. A 2025 MIT study revealed a 34% higher incidence of confident language when models are wrong versus when they are right.

The Challenge: AI systems are exhibiting a dangerous confidence paradox – they're more likely to assert incorrect statements with confident language ("definitely," "without doubt"), making hallucinations deceptively persuasive. A 2025 MIT study revealed a 34% higher incidence of confident language when models are wrong versus when they are right.

BUSINESS IMPACT


  • 47% of enterprise AI users acted on hallucinated insights (2025)

  • $67.4B lost due to AI misinformation (2024)

  • 41% of organizations made flawed strategic decisions


  • 47% of enterprise AI users acted on hallucinated insights (2025)

  • $67.4B lost due to AI misinformation (2024)

  • 41% of organizations made flawed strategic decisions


  • 47% of enterprise AI users acted on hallucinated insights (2025)

  • $67.4B lost due to AI misinformation (2024)

  • 41% of organizations made flawed strategic decisions

USER IMPACT


  • Users can’t tell correct vs. incorrect confident outputs

  • Transparency is lacking, eroding trust

  • No clear indicators of AI certainty → misplaced confidence


  • Users can’t tell correct vs. incorrect confident outputs

  • Transparency is lacking, eroding trust

  • No clear indicators of AI certainty → misplaced confidence


  • Users can’t tell correct vs. incorrect confident outputs

  • Transparency is lacking, eroding trust

  • No clear indicators of AI certainty → misplaced confidence

( *** )
GOALS

Primary Goal: Design an AI interface system that transforms uncertainty from a weakness into a strength by making AI confidence levels transparent and actionable.

Secondary Goals:
Reduce user reliance on incorrect AI outputs by 60%
Increase user confidence in AI decision-making by 45%
Build scalable design patterns for AI transparency.
GOALS

Primary Goal: Design an AI interface system that transforms uncertainty from a weakness into a strength by making AI confidence levels transparent and actionable.

Secondary Goals:
Reduce user reliance on incorrect AI outputs by 60%
Increase user confidence in AI decision-making by 45%
Build scalable design patterns for AI transparency.
( 02 )
RESEARCH INSIGHTS
( 03 )
COMPETITIVE ANALYSIS
( 04 )
PERSONA

ALIA

ALIA

Age: 37
Role: Senior Business Analyst (Financial Services)
Experience: 12 years in risk analysis and strategic decision-making
Location: Bangalore, India

Age: 37
Role: Senior Business Analyst (Financial Services)
Experience: 12 years in risk analysis and strategic decision-making
Location: Bangalore, India

Age: 37
Role: Senior Business Analyst (Financial Services)
Experience: 12 years in risk analysis and strategic decision-making
Location: Bangalore, India

PAIN POINTS

PAIN POINTS

  1. Finds it difficult to identify flawed AI outputs due to overconfidence and lack of transparency.

  2. Feels anxious about making high-risk decisions based on unreliable or unexplained AI insights.

  1. Finds it difficult to identify flawed AI outputs due to overconfidence and lack of transparency.

  2. Feels anxious about making high-risk decisions based on unreliable or unexplained AI insights.

  1. Finds it difficult to identify flawed AI outputs due to overconfidence and lack of transparency.

  2. Feels anxious about making high-risk decisions based on unreliable or unexplained AI insights.

Needs / goals

NEEDS / GOALS

  1. Make accurate, low-risk, data-driven decisions while maintaining professional credibility.

  2. Rely on transparent, controllable AI tools with clear confidence indicators, reasoning access, and verification options.

  1. Make accurate, low-risk, data-driven decisions while maintaining professional credibility.

  2. Rely on transparent, controllable AI tools with clear confidence indicators, reasoning access, and verification options.

  1. Make accurate, low-risk, data-driven decisions while maintaining professional credibility.

  2. Rely on transparent, controllable AI tools with clear confidence indicators, reasoning access, and verification options.

( 05 )
( 05 )
USERFLOW
USERFLOW

A clear, trust-focused flow that lets users interact with AI, view confidence levels, understand reasoning, and act or verify results easily.

A clear, trust-focused flow that lets users interact with AI, view confidence levels, understand reasoning, and act or verify results easily.

A clear, trust-focused flow that lets users interact with AI, view confidence levels, understand reasoning, and act or verify results easily.

( 06 )
( 06 )
HI-FIDELITY
WIREFRAMES
HI-FIDELITY
WIREFRAMES

These wireframes will display clear, color-coded confidence indicators for every AI output. Transparency options, such as “How was this calculated?”, will always be accessible but unobtrusive. Trust tools will surface based on user actions, and verification options—including quick access to sources and expert help—will be seamless.

These wireframes will display clear, color-coded confidence indicators for every AI output. Transparency options, such as “How was this calculated?”, will always be accessible but unobtrusive. Trust tools will surface based on user actions, and verification options—including quick access to sources and expert help—will be seamless.

These wireframes will display clear, color-coded confidence indicators for every AI output. Transparency options, such as “How was this calculated?”, will always be accessible but unobtrusive. Trust tools will surface based on user actions, and verification options—including quick access to sources and expert help—will be seamless.

( *** )
( 07 )

IMPACTS

IMPACTS

Increased User Trust and Confidence

Increased User Trust and Confidence

Clear visibility of AI confidence levels and reasoning helps users trust outputs and make informed decisions.

Clear visibility of AI confidence levels and reasoning helps users trust outputs and make informed decisions.

Reduced Errors and Operational Risk

Reduced Errors and Operational Risk

Warning cues and verification options prevent decisions based on uncertain or incorrect AI information, minimizing costly mistakes.

Warning cues and verification options prevent decisions based on uncertain or incorrect AI information, minimizing costly mistakes.

Improved Decision Efficiency

Improved Decision EfficiencY

Actionable high-confidence responses speed up workflows and reduce time spent on manual verification.

Actionable high-confidence responses speed up workflows and reduce time spent on manual verification.

© All rights reserved to Tejasvi Murmu

BEHANCE

BEHANCE

BEHANCE

BEHANCE

BEHANCE

BEHANCE

LINKEDIN

LINKEDIN

LINKEDIN

LINKEDIN

LINKEDIN

LINKEDIN

DRIBBBLE

DRIBBBLE

DRIBBBLE

DRIBBBLE

DRIBBBLE

DRIBBBLE

TURNING DIGITAL CHAOS INTO CHEF'S KISS EXPERIENCES.

TURNING DIGITAL CHAOS INTO CHEF'S KISS EXPERIENCES.

LET'S WORK

TOGETHER

LET'S WORK

TOGETHER

LET'S WORK

TOGETHER

HIRE ME NOW

HIRE ME NOW

HIRE ME NOW

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