Designing Agent Voice, Tone, and Boundaries for Trust-Driven Adoption
Designing Agent Voice, Tone, and Boundaries for Trust-Driven Adoption
How I used user mental models to guide language, escalation, and handoff in an AI system.
How I used user mental models to guide language, escalation, and handoff in an AI system.
Context
Expert Mock, an AI-powered mock interview platform, had an MVP but lacked clarity on brand positioning, voice, and user trust. Founder intent and user perception were misaligned, limiting differentiation and adoption in a crowded market.
My Role
Content Strategist & Workshop Facilitator
I led research, facilitation, and synthesis to define a clear brand narrative, voice system, and content direction grounded in user perception.
Methods
Brand Diagnostic Workshop (Founder Lens):
Designed and facilitated a structured workshop to understand how the founder perceived the brand, using:
Peaks & Valleys
Dreams & Nightmares
Golden Circle
Purpose articulation
Competitive landscape mapping
This surfaced assumptions around authority, differentiation, and value.
Co-Creation Workshop (User Lens):
Facilitated a participatory workshop to capture how users experienced the brand, using:
Brand DNA & personality exercises
“What if” messaging patterns
Archetype exploration
This / Not That value clarification
Content ideation frameworks (Crazy 8s)
This revealed emotional expectations users had from the product.
Founder vs. User Perception Mapping:
Synthesized workshop outputs to compare:
Founder view: Power, authority, strategic guidance
User view: Empathy, support, approachability, confidence-building
This gap became the core design problem.
Brand Archetype Definition:
Translated insights into a Mentor archetype, defining:Goals (guidance, growth, preparation)
Motivations (care, impact, legacy)
Pains (fear of failure, missed potential)
This archetype anchored tone, content depth, and trust signals.
Voice, Tone & Content System Design
Designed a cohesive content system including:Brand essence and value proposition
Voice & tone principles (confident, supportive, empowering)
Messaging pillars and content concepts
Visual and narrative examples for consistency across channels
This ensured the brand sounded human, credible, and encouraging, not intimidating.
Brand Diagnostic Workshop (Founder Lens)
Designed and facilitated a structured workshop to understand how the founder perceived the brand, using:
Peaks & Valleys
Dreams & Nightmares
Golden Circle
Purpose articulation
Competitive landscape mapping
This surfaced assumptions around authority, differentiation, and value.
Co-Creation Workshop (User Lens): Facilitated a participatory workshop to capture how users experienced the brand, using:
Brand DNA & personality exercises
“What if” messaging patterns
Archetype exploration
This / Not That value clarification
Content ideation frameworks (Crazy 8s)
This revealed emotional expectations users had from the product.
Voice, Tone & Messaging Frameworks
Designed content principles that balanced:Functional rigor (feedback, improvement)
Emotional safety (encouragement, reassurance)
Clear expectations around AI limitations and support
Delegation & Trust Strategy
Defined when the system should guide, when it should challenge, and how it should explain feedback to avoid discouragement or over-authority.
Key Insight
Users were more willing to engage when the brand felt like a partner in growth, not an evaluator. Trust emerged from how the system communicated, not just what it offered.
Outcome
Clear brand positioning rooted in user trust
Unified voice, tone, and messaging across touchpoints
Content strategies supporting acquisition and engagement
Foundation for scalable marketing, partnerships, and growth
This project demonstrates how content strategy, facilitation, and synthesis can transform ambiguous brand intent into a clear, trust-driven content system—critical for adoption in AI-assisted products.



Product concept & Brand attributes
Product concept & Brand attributes



Brand Diagnostic Workshop Plan with Founder
Brand Diagnostic Workshop Plan with Founder



Co-creation Workshop with Users
Co-creation Workshop with Users



Analysis & Brand Archetype
Analysis & Brand Archetype



Voice, Tone & Content System Design
Voice, Tone & Content System Design
Context
Rubenius Interiors, an award-winning interior design studio with global operations, faced friction in client onboarding. Every new prospect needed repeated explanations of services, sectors, and portfolio, and the manual requirements-gathering process slowed down awareness, lead generation, and conversion.
My Role
UX/Product Designer & Conversational Architect
I analyzed common chatbot failures,
Mapped user intent patterns,
Designed personas, and
Built an 8-step conversational framework that adapts to different prospect types while keeping data capture seamless.
Methods
UX Audit: Studied existing chatbots to identify typical drop-off triggers, long threads, unclear pathways, information overload.
User Segmentation: Defined 3 user types aligned with the funnel:
Window-Shopper – browsing for information
Inquisitive Researcher – comparing differentiators
Goal-Oriented Client – ready to share project details
Conversation Architecture: Created a branching flow that intelligently responds to each persona’s intent, from awareness to off-boarding.
Prototyping & Iteration: Built and refined a modular chatbot using free low-code tools.
Key Insight
Clients didn’t follow a linear journey. They entered with different intent levels and forcing them into a single script caused friction. Designing a branching conversation allowed the chatbot to educate, engage, and capture requirements without overwhelming users.
Outcome
The final chatbot delivered a fast, engaging onboarding experience that:
Increased qualified lead retention by 30% in beta
Eliminated repetitive manual explanations
Captured project details upfront (location, scope, budget, preferences)
Supported 24/7 automated lead qualification across all markets
It gave Rubenius a scalable way to understand client needs before the first call, reducing friction and accelerating conversions.
