Activision AI Assisted DasHboard
Playtest Triage Dashboard for UX Researchers
COMPANY
Activision Capstone
ROLE
UX Designer/ Researcher
EXPERTISE in this project
UX Research/Design
YEAR
April 2026(Ongoing)
UX researchers have an AI that detects anomalies, but no interface to interpret, trust, or act on what it finds.
Who is Affected
UX researchers at game studios who run playtest sessions, analyze player behavior, and present findings to game designers and producers, often under extreme deadline pressure.
Why Required Now
Studios are rapidly deploying AI to detect player frustration and reduce cognitive load, but without researcher-facing interpretation tools, those insights are unlikely to be trusted or adopted.
Cost of Inaction
If nothing changes, researchers keep spending 2 weeks per study verifying AI claims they can't trust, and the AI investment goes unused. One bad experience is all it takes to abandon the tool entirely.
100’s flags per playtest session | 0 Tools expose AI reasoning | 15 Days Lost per study to manual work.
Research & Design Process
Every project begins with understanding people before designing interfaces.
User Interview Insights
We talked to various UX researchers in Activision to understand their thought process and workstyle to gain insights on current AI workflow.
User Journey Mapping
After collecting the user insights and interview data, we mapped our users current journey to better understand the fricition areas and pain points in their workflow.
Concept Building
Through our thorough, competitive analysis, data collection, insights from interviews, observation from user personas & journey maps, we built our design concepts.
Concept Overview 01 - Research Question Upfront
A collapsible prompt that gates the flag list behind a research question: only relevant flags surface, while out-of-scope ones dim.
Our core Innovation :
“Provides the user a filtered overview of relevant flags, dismissing the ones which are not useful, reducing the user's cognitive load of skimming through 100s of flags.”
Concept 1 Paper Prototypes
Concept Overview 02 - Conversational Challenge Panel
Override opens a dialogue space. Researcher haves a conversation in case of disagreement, AI produces suggestion, and flags are locked in.
Our core Innovation :
“Gives user more control rather to understand the flag better by having a conversation with AI for giving prompts to correct the log or override it.
Concept 2 Paper Prototypes
Concept Overview 03 - Smart Search Bar & Chat History
An always-visible search on top of the dashboard that navigates through multiple features in dashboard. Chat history lets researchers retrace decisions without starting over.
Our core Innovation :
“ Acts as a dashboard directory for users to find any information quickly and easily, without having to go through various complex information and pages
Concept 3 Paper Prototypes
UX researchers at Activision have an AI that detects, but no interface that helps them interpret, trust, or act.
What we Learned
Explanation panels are not a feature. They are the mechanism through which the dashboard earns the right to be used at all.
What we Built
Three connected concepts, context filtering, conversational override, and smart navigation, grounded in five interviews and zero assumptions.
Next Steps
Going ahead, we intend to deep dive in building the, wireframes, UI design for the proposed concepts and a complete AI-assisted dashboard. We plan to do user testing, and vibe coding as the final deliverable for the project.
Leveraged Figma Make for AI-assisted initial concept generation, alongside deep engagement with AI explainability principles as both a design subject and a working method.
Used Figma Make to rapidly generate initial interface concepts from research insights, accelerating the ideation-to-prototype pipeline
Applied AI-assisted tools throughout the research and synthesis process to structure findings and identify patterns across interview data
The project required understanding AI system architecture well enough to design a meaningful interpretive layer between model outputs and end users
Designed for an AI model's outputs, developing fluency in explainability, confidence communication, and human-AI trust dynamics