Niabot: Conversational AI Portfolio Assistant
An interactive conversational guide that reimagines portfolio exploration using a hybrid AI UX approach, showcasing my expertise.
Role
Conversational AI Designer
Tools
Voiceflow, Miro, FigJam
Duration
2 months
Project type
Webchat-based Conversational AI Interface
Target Audience
Hiring Managers, Recruiters, AI Design Leads & Portfolio Explorers
Tools Used | Methods |
|---|---|
|
|
“Designing Niabot let me wrestle with the same questions conversation designers face dailly - just in new contexts. I stay grounded in the craft through curiosity, community, and a genuine love for shaping how people and technology speak to each other.”
TL;DR
Created an interactive chatbot that demonstrates my conversational AI skills in action while helping visitors navigate my portfolio
Built a hybrid architecture combining deterministic flows with LLM flexibility, demonstrating technical implementation skills
Achieved 20% increase in recruiter engagement and generated 4 interview opportunities
Designed personalized conversation paths for different user types: hiring managers seeking specific qualifications vs. explorers browsing my work
🎯 The Challenge: Show, Don't Tell
As a designer expanding into conversational AI and UX, I faced a challenge:
Portfolio employers and clients struggled to quickly understand my diverse skillset.
Traditional static case studies weren't demonstrating my abilities effectively. Research revealed hiring managers spend just seconds scanning portfolios, and most conversational AI positions required demonstrated experience.
I needed a solution that would:
Create immediate clarity and connection
Demonstrate my Conversational AI design skills through real implementation
Provide a personalized navigation experience for different user types
Serve as a learning ground for my own skill development
🔍 UX for AI Integration
One of the most critical aspects of implementing NiaBot was designing for seamless discovery and adoption. I focused on three key points:
UX Principle | Design Challenge | Implementation Approach |
|---|---|---|
Findability | How would users discover NiaBot on my portfolio site? | Strategic positioning in bottom right corner with subtle animation to draw attention |
Discoverability | How would users understand NiaBot's purpose and value? | Clear welcome message explaining capabilities with visual cues that leverage familiar mental models (bot icon, chat interface, buttons) |
Learnability | How would users learn to effectively interact with NiaBot? | Progressive disclosure of features with conversational guidance and quick-reply options |
🛠 My Approach: Bridging Human Needs with Technical Implementation
1. User-Centered Planning
I began by creating detailed personas for both hiring managers and portfolio explorers, then mapped a comprehensive generic user journey tracking emotions, pain points, and opportunities across the entire conversation flow.
2. Systems Thinking Approach
Rather than designing the chatbot in isolation, I mapped the entire ecosystem, considering:
How NiaBot would integrate with my portfolio website? How users would move between chatbot interactions and portfolio content?
Key interconnections between responses and navigation patterns
Environmental, social and technological factors influencing user-chatbot interactions
3. Hybrid Architecture Design
I developed a custom hybrid conversational architecture that addresses the fundamental challenge in conversational AI: balancing structured guidance with natural flexibility. My approach centered on creating differentiated experiences based on user type and intent, combining deterministic elements with AI-powered flexibility. This architecture enabled personalized journeys while maintaining system reliability and conversational coherence.
The design prioritized different experiences based on user needs:
For exploratory users: An experience optimized for discovery and engagement
For goal-oriented users: An experience optimized for efficiency and direct information access
This approach resulted in better user satisfaction scores and increased engagement metrics compared to purely deterministic or purely generative approaches.
4. Bot Persona Development
I crafted NiaBot with a warm, professional tone—designed to be approachable and engaging without ever overshadowing the content. Its voice guidelines established a clear perspective: speaking in the first person as the bot to foster connection, while referring to me in the third person to maintain clarity and context.
5. Dialog Flow Implementation
I designed tailored conversation flows for two key user groups—hiring managers and explorers—beginning with the primary ‘happy paths’ before expanding into edge cases. Following the pareto principle that around 80% of users typically engage with the most frequent 20% of dialog paths, I focused on optimizing those high-impact journeys first:
Hiring managers: Greeting → Skills overview → Project highlights → Contact options
Explorers: Greeting → Portfolio introduction → Interactive exploration → Deeper dives
The strategic approach prioritized common paths while ensuring graceful handling of unexpected inputs.
💡 Technical Implementation
Multi-Modal Contact Methods:
Hiring Managers: Structured form submissions
Explorers: Conversational messaging
All Users: Calendly integration for direct meeting scheduling
🧠 Context-Aware Conversations
I implemented memory variables to preserve user preferences, with context awareness that varied by
user type:
Explorer route: Full conversational flexibility, allowing natural topic shifts
Hiring Manager route: Deterministic flows with context awareness for error recovery
🚧 Technical Roadblocks
Key Challenges: Had to work within token limits on LLMs, which restricted the length and complexity of some conversational flows.
🔒 Privacy & Design Principles
My development approach integrated privacy considerations and accessibility standards from the beginning, rather than treating them as afterthoughts. Key focus areas included:
User-centered data practices that balance personalization with privacy
Clear information architecture supporting multiple interaction preferences
Accessibility-driven design choices that improve the experience for all users
🧪 Testing & Iteration: From Prototype to Refined Solution
After developing and deploying an initial, simplified version based on the journey map, I gathered real user interactions and feedback that shaped NiaBot’s evolution—analyzing utterance patterns to identify common intents and edge cases.
Feedback Theme | User Insights | Implementation Changes |
|---|---|---|
Text Density | "Responses are too long and text-heavy" | Broke complex information into multiple conversation turns |
Navigation Clarity | "Not sure how to find specific portfolio sections" | Added clear navigation options at key decision points |
Persona Consistency | "The bot feels robotic and switches between first and third person" | Refined NiaBot's personality and standardized conversational style |
Content Coverage | "Information about your best work is missing" | Expanded content to cover frequently requested topics |
Technical Topics | “I’m not sure where NiaBot is hosted, what data it stores, or how secure it is” | Added proactive explanations, context-sensitive help options, and transparency around hosting, data handling, and privacy safeguards |
🔍 Ongoing Usability Testing & Evaluation…
The evolution of NiaBot continues through rigorous testing. My current evaluation focus includes several key metrics:
Testing Focus | Key Questions & Measurement Areas |
|---|---|
Journey Completion |
|
Conversion Metrics |
|
Error Analysis |
|
Popular Pathways |
|
User Satisfaction |
|
I'm implementing a feedback mechanism that presents at the end of each conversation:
For measuring performance, I developed a comprehensive evaluation framework that combines platform analytics with qualitative assessment methods.
✨ Key Discoveries
Through this project, I identified fundamental principles that separate great conversational experiences from merely functional ones:
1. Grice's Conversational Maxims Matter
Quality: Providing truthful, accurate information
Quantity: Balancing informativeness with brevity
Relation: Staying relevant to queries
Manner: Ensuring clear, well-structured responses
2. Emotional Intelligence > Information Delivery
The most engaging interactions acknowledge user context before providing information
This creates rapport that purely informational exchanges lack
3. Structure Creates Safety
Users navigate conversations more confidently with clear pathways
Even AI-driven interfaces benefit from predictable patterns
4. Error Recovery Defines Trust
A bot's response to misunderstandings is more important than handling expected inputs
Graceful error recovery builds confidence in the overall experience
I developed a multi-layered approach to error handling:
Intent Recognition Monitoring
Contextual Fallbacks
Graceful Recovery Paths
Continuous Learning Framework
“Designing Niabot let me wrestle with the same questions conversation designers face dailly - just in new contexts. I stay grounded in the craft through curiosity, community, and a genuine love for shaping how people and technology speak to each other.”
🔄 My Multidisciplinary Foundation
My experience across different disciplines has shaped my approach to conversational AI:
Brand design background provides a foundation for crafting compelling, consistent bot personalities
Creative technology experience instilled a "learning by doing" mentality that allowed me to quickly master new tools
UX/UI design skills ensure I create conversations that are both functional and delightful
Linguistics understanding informs natural dialogue construction, particularly how context affects meaning
Human interaction patterns form the foundation of conversation design, drawing from principles like turn-taking mechanics
Cognitive psychology insights guide decisions about memory limitations and attention management
🚀 Why This Matters for Your Team
My approach to conversational AI design combines creative thinking with technical implementation:
User-First Design: Starting with human needs, not just technical capabilities
Evidence-Based Refinement: Using conversation data to drive improvements
Strategic Flexibility: Balancing structured flows with the adaptability of AI
Outcome-Oriented Thinking: I measure success through outcomes that drive business value and enhance the user experience
I bring a hybrid perspective that bridges brand experience, technical architecture, and UX design, a combination that helps create conversational experiences that truly connect and convert.
You can experience and interact with NiaBot in the bottom right corner of this page.
For more detailed information about specific implementation techniques or to discuss collaboration opportunities, please feel free to reach out directly. I'm always open to connecting with professionals in the field :)
Other projects
Niabot: Conversational AI Portfolio Assistant
An interactive conversational guide that reimagines portfolio exploration using a hybrid AI UX approach, showcasing my expertise.
Role
Conversational AI Designer
Tools
Voiceflow, Miro, FigJam
Duration
2 months
Project type
Webchat-based Conversational AI Interface
Target Audience
Hiring Managers, Recruiters, AI Design Leads & Portfolio Explorers
(For optimal experience, this case study is best viewed on desktop devices 💻)
Tools Used | Methods |
|---|---|
|
|
“Designing Niabot let me wrestle with the same questions conversation designers face dailly - just in new contexts. I stay grounded in the craft through curiosity, community, and a genuine love for shaping how people and technology speak to each other.”
Created an interactive chatbot that demonstrates my conversational AI skills in action while helping visitors navigate my portfolio
Built a hybrid architecture combining deterministic flows with LLM flexibility, demonstrating technical implementation skills
Achieved 20% increase in recruiter engagement and generated 4 interview opportunities
Designed personalized conversation paths for different user types: hiring managers seeking specific qualifications vs. explorers browsing my work
TL;DR
🎯 The Challenge:
Show, Don't Tell
As a designer expanding into conversational AI and UX, I faced a challenge:
Portfolio employers and clients struggled to quickly understand my diverse skillset.
Traditional static case studies weren't demonstrating my abilities effectively. Research revealed hiring managers spend just seconds scanning portfolios, and most conversational AI positions required demonstrated experience.
I needed a solution that would:
Create immediate clarity and connection
Demonstrate my Conversational AI design skills through real implementation
Provide a personalized navigation experience for different user types
Serve as a learning ground for my own skill development
🔍 UX for AI Integration
One of the most critical aspects of implementing NiaBot was designing for seamless discovery and adoption. I focused on three key points:
UX Principle | Design Challenge | Implementation Approach |
|---|---|---|
Findability | How would users discover NiaBot on my portfolio site? | Strategic positioning in bottom right corner with subtle animation to draw attention |
Discoverability | How would users understand NiaBot's purpose and value? | Clear welcome message explaining capabilities with visual cues that leverage familiar mental models (bot icon, chat interface, buttons) |
Learnability | How would users learn to effectively interact with NiaBot? | Progressive disclosure of features with conversational guidance and quick-reply options |
🛠 My Approach: Bridging Human Needs with Technical Implementation
1. User-Centered Planning
I began by creating detailed personas for both hiring managers and portfolio explorers, then mapped a comprehensive generic user journey tracking emotions, pain points, and opportunities across the entire conversation flow.
2. Systems Thinking Approach
Rather than designing the chatbot in isolation, I mapped the entire ecosystem, considering:
How NiaBot would integrate with my portfolio website? How users would move between chatbot interactions and portfolio content?
Key interconnections between responses and navigation patterns
Environmental, social and technological factors influencing user-chatbot interactions
3. Hybrid Architecture Design
I developed a custom hybrid conversational architecture that addresses the fundamental challenge in conversational AI: balancing structured guidance with natural flexibility. My approach centered on creating differentiated experiences based on user type and intent, combining deterministic elements with AI-powered flexibility. This architecture enabled personalized journeys while maintaining system reliability and conversational coherence.
The design prioritized different experiences based on user needs:
For exploratory users: An experience optimized for discovery and engagement
For goal-oriented users: An experience optimized for efficiency and direct information access
This approach resulted in better user satisfaction scores and increased engagement metrics compared to purely deterministic or purely generative approaches.
4. Bot Persona Development
I crafted NiaBot with a warm, professional tone—designed to be approachable and engaging without ever overshadowing the content. Its voice guidelines established a clear perspective: speaking in the first person as the bot to foster connection, while referring to me in the third person to maintain clarity and context.
5. Dialog Flow Implementation
I designed tailored conversation flows for two key user groups—hiring managers and explorers—beginning with the primary ‘happy paths’ before expanding into edge cases. Following the pareto principle that around 80% of users typically engage with the most frequent 20% of dialog paths, I focused on optimizing those high-impact journeys first:
Hiring managers: Greeting → Skills overview → Project highlights → Contact options
Explorers: Greeting → Portfolio introduction → Interactive exploration → Deeper dives
The strategic approach prioritized common paths while ensuring graceful handling of unexpected inputs.
💡 Technical Implementation
Multi-Modal Contact Methods
Hiring Managers: Structured form submissions
Explorers: Conversational messaging
All Users: Calendly integration for direct meeting scheduling
🧠 Context-Aware Conversations
I implemented memory variables to preserve user preferences, with context awareness that varied by
user type:
Explorer route: Full conversational flexibility, allowing natural topic shifts
Hiring Manager route: Deterministic flows with context awareness for error recovery
🚧 Technical Roadblocks
Key Challenges: Had to work within token limits on LLMs, which restricted the length and complexity of some conversational flows.
🔒 Privacy & Design Principles
My development approach integrated privacy considerations and accessibility standards from the beginning, rather than treating them as afterthoughts. Key focus areas included:
User-centered data practices that balance personalization with privacy
Clear information architecture supporting multiple interaction preferences
Accessibility-driven design choices that improve the experience for all users
🧪 Testing & Iteration: From Prototype to Refined Solution
After developing and deploying an initial, simplified version based on the journey map, I gathered real user interactions and feedback that shaped NiaBot’s evolution—analyzing utterance patterns to identify common intents and edge cases.
Feedback Theme | User Insights | Implementation Changes |
|---|---|---|
Text Density | "Responses are too long and text-heavy" | Broke complex information into multiple conversation turns |
Navigation Clarity | "Not sure how to find specific portfolio sections" | Added clear navigation options at key decision points |
Persona Consistency | "The bot feels robotic and switches between first and third person" | Refined NiaBot's personality and standardized conversational style |
Content Coverage | "Information about your best work is missing" | Expanded content to cover frequently requested topics |
Technical Topics | “I’m not sure where NiaBot is hosted, what data it stores, or how secure it is” | Added proactive explanations, context-sensitive help options, and transparency around hosting, data handling, and privacy safeguards |
🔍 Ongoing Usability Testing & Evaluation…
The evolution of NiaBot continues through rigorous testing. My current evaluation focus includes several key metrics:
Testing Focus | Key Questions & Measurement Areas |
|---|---|
Journey Completion |
|
Conversion Metrics |
|
Error Analysis |
|
Popular Pathways |
|
User Satisfaction |
|
I'm implementing a feedback mechanism that presents at the end of each conversation:
For measuring performance, I developed a comprehensive evaluation framework that combines platform analytics with qualitative assessment methods.
✨ Key Discoveries
Through this project, I identified fundamental principles that separate great conversational experiences from merely functional ones:
1. Grice's Conversational Maxims Matter
Quality: Providing truthful, accurate information
Quantity: Balancing informativeness with brevity
Relation: Staying relevant to queries
Manner: Ensuring clear, well-structured responses
2. Emotional Intelligence > Information Delivery
The most engaging interactions acknowledge user context before providing information
This creates rapport that purely informational exchanges lack
3. Structure Creates Safety
Users navigate conversations more confidently with clear pathways
Even AI-driven interfaces benefit from predictable patterns
4. Error Recovery Defines Trust
A bot's response to misunderstandings is more important than handling expected inputs
Graceful error recovery builds confidence in the overall experience
I developed a multi-layered approach to error handling:
Intent Recognition Monitoring
Contextual Fallbacks
Graceful Recovery Paths
Continuous Learning Framework
“Designing Niabot let me wrestle with the same questions conversation designers face dailly - just in new contexts. I stay grounded in the craft through curiosity, community, and a genuine love for shaping how people and technology speak to each other.”
🔄 My Multidisciplinary Foundation
My experience across different disciplines has shaped my approach to conversational AI:
Brand design background provides a foundation for crafting compelling, consistent bot personalities
Creative technology experience instilled a "learning by doing" mentality that allowed me to quickly master new tools
UX/UI design skills ensure I create conversations that are both functional and delightful
Linguistics understanding informs natural dialogue construction, particularly how context affects meaning
Human interaction patterns form the foundation of conversation design, drawing from principles like turn-taking mechanics
Cognitive psychology insights guide decisions about memory limitations and attention management
🚀 Why This Matters for Your Team
My approach to conversational AI design combines creative thinking with technical implementation:
User-First Design: Starting with human needs, not just technical capabilities
Evidence-Based Refinement: Using conversation data to drive improvements
Strategic Flexibility: Balancing structured flows with the adaptability of AI
Outcome-Oriented Thinking: I measure success through outcomes that drive business value and enhance the user experience
I bring a hybrid perspective that bridges brand experience, technical architecture, and UX design, a combination that helps create conversational experiences that truly connect and convert.
You can experience and interact with NiaBot in the bottom right corner of this page.
For more detailed information about specific implementation techniques or to discuss collaboration opportunities, please feel free to reach out directly. I'm always open to connecting with professionals in the field.