Hello.
I'm Mahsa.
I'm a User Experience Researcher. I turn open business questions into studies teams can act on — and make the evidence outlive the readout.
My own research looks at how AI systems interpret human emotion and the body, and I'm working towards a PhD in human–AI interaction.
See my case studies
Quick links
How I work
- 01
Frame the real problem
I start from the growth or trust problem the team is stuck on — an engagement ceiling, an abandoned checkout, an AI feature nobody trusts — and turn it into research questions a study can actually close.
- 02
Triangulate methods
Behavioural and quantitative data locate where things break; interviews, usability sessions and journey mapping explain why. I run them deliberately together, not sequentially, so each one shapes the next instrument.
- 03
Follow the contradiction
The most useful finding is usually where sources disagree — stated privacy concerns versus untouched defaults. I present the contradiction as the central insight instead of smoothing it into an average.
- 04
Turn evidence into specs
Findings land as validated design specs, principles and trust models engineering can build once and configure per market — with limits, scope and the decisions they informed written down.
Selected case studies
The ability seems to be there — the defaults just point somewhere else.
When the Body Speaks First: How LLM Chatbots Respond to Early Bodily-Emotional Distress in English and Persian
A small exploratory study of how LLM chatbots respond to everyday expressions of distress in English and Persian, and to body maps people drew of where they feel their emotions.
- Cross-lingual message sets (English & Persian)
- Multi-model comparison (OpenAI, Anthropic, Google)
- Body maps as stimuli
How I embedded behavioral research into sprint decisions, and what I found that no dashboard could surface, across a 3M+ user e-commerce platform
Beyond A/B Testing: Building a Real-Time Research Engine for a Live Platform Redesign
A six-month, mixed-methods research programme run alongside a live full-platform migration: a seven-stage continuous cycle that fed funnel data, heatmaps and usability sessions straight into bi-weekly product decisions.
- Mixed methods
- Funnel analysis
- Hotjar heatmaps & recordings
- Usability testing (36 participants)
- Surveys (5,000+ responses)
- Continuous metric optimisation
What 1,500 users across three countries revealed about why ‘more privacy controls’ is the wrong answer
Architecting Trust and Engagement in a Multinational Social Network
An 11-week parallel mixed-methods programme across Canada, the Netherlands and Italy that mapped the cultural architecture of trust and reframed the MVP from “more controls” to “better defaults”.
- Semi-structured interviews (39 total)
- Large-scale survey (1,500+ respondents)
- Trust-scenario usability tests
- Privacy mental model mapping
- Competitive & cultural analysis
How identifying 4 trust failures reduced call center contacts by 31% and increased checkout conversion by 19%
Building Trust in High-Stakes Hybrid Services: The iToll Vehicle Handover Story
A trust-focused research programme on a hybrid digital–physical service: users had to hand a high-value car to a stranger. Call center intercepts, heatmaps and a heuristic audit produced a three-layer handover model.
- Call center intercepts (35 users)
- Session recordings & heatmaps
- Heuristic evaluation
- Behavioral audit
- Literature review
- Concept testing
How a pre-launch ethics foundation kept a conversational AI a self-help companion, not a therapist
Designing a “Step Zero” for Responsible AI in a Mental Health App
A pre-launch ethics foundation for a conversational mental health AI: expert walkthroughs of scripted dialogs, simulated user perspectives, an ethical language checklist, and a conversation contract designed into the product.
- Expert walkthroughs of scripted conversations
- Simulated user perspectives (through experts)
- Literature & responsible-AI guidance review
- Public discourse scan (X/Twitter, Reddit)
- Ethical language checklist
A privacy-first HCI study in which methodological restraint was part of the research quality
Researching Trauma-Informed Social Media Without Compromising Participant Safety
A concise academic case study on conducting trauma-informed social-media research through interviews and a bilingual survey while minimising participant exposure and retaining no identifying data.
- Bilingual survey
- Semi-structured interviews
- Iterative thematic analysis
- Low-fidelity prototyping
- Expert walkthroughs
Research toolkit
- Interviews & Usability Testing
- Moderated sessions end to end — from multinational social-network research at Together to validating the One Thing habit MVP against primary interviews.
- Behavioural & Product Analytics
- Live platform data at Okala's real-time research engine, plus Microsoft Clarity, Google Analytics and Hotjar. Certified: Product Analytics Micro-Certification (PAC)™, Product School.
- Mixed Methods & Survey Design
- Quantitative instruments designed and analysed honestly, combined with qual — as in the iToll high-stakes handover study. Course: UX Research — Overcoming Data Anxiety (LinkedIn).
- Behavioural Science & Trust
- Cognitive psychology and human decision-making applied to trust architecture — the thread through my Step Zero AI and functional-friction work.
- CRO & Experimentation
- Conversion research, A/B testing and funnel evidence — applied in the Okala redesign and across experimentation programmes.
- AI-Assisted Research & Prompt Engineering
- Claude and Gemini as analysis partners — from macro-scale social-signal synthesis to iterating research instruments. Certified: Introduction to Prompt Engineering for Generative AI (LinkedIn).
Selected articles
View all articles“The problem arises when we design a solution for our specific user, but fail to design the path leading to that solution for them.”
We're Trying Hard to Design the Right Solution, But Did We Design the Right Path?
Onboarding & user understanding · 5 min read
“When a product outsources the user's decision-making process entirely, eliminating friction ceases to be a design achievement and becomes a product risk.”
When Friction Is Functional: Designing Intentional Boundaries in AI Products
AI products & boundaries · 6 min read
“The first step to becoming a professional researcher is accepting that the player's brain is not a flawless machine — it is an evolved system optimized for saving energy and resources.”
The Gamer's Brain: How Neuroscience and UX Make You a Better Game Researcher
Neuroscience & game research · 5 min read
“Peripheral vision affects our emotions, and our emotions affect our choices — every decision on a screen starts with watching.”
Have You Ever Thought About the Effect of Visual Perception on Your Design?
Visual perception · 5 min read