Will Yang · Behavioral research / Human factors / AI workflows

Every problemis a technical problem.

I lead Consulting and Business Development at Noldus China, use research to inform product decisions, and explore how AI can reduce repetitive professional work. By technical, I mean taking a difficulty apart, building an approach, and checking whether it helps.

Enter the interactive home Explore the casesDiscuss your project

The claim in practice

A different question. A practical next step.

Three cases on organizing evidence, testing earlier, and simulating an unfinished interaction. Each explains my role, the work, and the limits of the results.

See all cases and experiments
  1. EVIDENCE ARCHITECTURE
    A route back to the evidence254GB of research records organized by tasks and observations into 86 traceable issues in six categories. Lines show relationships, not data distribution.RESEARCH RECORDS254 GBInterviews · TestsORIGINAL CONTEXTISSUE INDEX86issuesSIX CATEGORIESTask contextObservationSource recordEVERY FINDING HAS A SOURCE
    1. Source recordsKeep tasks and context
    2. Tasks & observationsSeparate actions and interpretations
    3. Traceable issuesEach finding has a source

    Records → Tasks & observations → Traceable issuesConceptual model

    DELIVERED2018Fintech · User researchFinance app: turning field records into issues a team can act onFrom organizing 254GB of records to finding grounds for product changes: review behavior by task and form 86 traceable issues.Delivery & progressDelivered 86 decision-ready issues in six categories, linked to tasks, observations, and priorities.
  2. DESIGN WITH FEEDBACK
    Build the check into the processInsight informs design, design and observation iterate, and confirmation follows. Tracks show feedback relationships, not iteration counts, durations, or project progress.DESIGN ⇄ OBSERVESTILL CHANGEABLEInsightConfirmFEEDBACK RETURNS TO DESIGN
    1. InsightDefine the assumption to examine
    2. Design ⇄ ObserveReturn feedback while changes are possible
    3. ConfirmationCheck whether the question was answered

    Insight → Design & observation in a loop → ConfirmationConceptual model

    Ongoing consulting practice2019—presentMedtech · Human factorsMedical devices: testing while there is still time to change the designFrom completing a test to checking design while change is possible: connect critical tasks, possible use errors, and test timing to development.Delivery & progressHelped establish study plans, recording templates, and evaluation processes used in work for 25+ medical-device companies and institutions.
  3. PROTOTYPE / TWO LAYERS
    Experience before implementationThe frontstage preserves the start, task, and end of the user experience. A researcher simulates unfinished responses backstage. Test understanding before choosing implementation.FRONTSTAGE / EXPERIENCEBACKSTAGE / SIMULATIONSTART · TASK · ENDWizard-of-Oz
    1. Frontstage · ExperienceStart, perform the task, end
    2. Backstage · SimulationA person follows response rules
    3. Observe & decideTest understanding before implementation

    Observe understanding · Simulate the responseConceptual model

    DELIVERED2019Intelligent hardware · HRIService robots: trying the interaction before building the featureFrom waiting for a finished feature to testing understanding now: separate interaction stages and simulate key responses before further development.Delivery & progressUsed a four-stage model and Wizard-of-Oz studies, and helped document repeatable research procedures.

Thinking about the question

Examine the question before looking for an answer.

Separate the goal from the proposed solution, and observations from interpretations. Use a study, a simulation, or a tool prototype to check the next step. The approach page covers the questions, worked examples, and tools.

Explore the full toolbox
Read the central argument: what I mean by “technical”Explore my approach and tools

About me

From understanding behavior to changing how work happens.

Psychology, four years of teaching, and field research taught me to watch what people actually do. When someone cannot use something, the cause may be a hidden entry point, unclear feedback, or an awkward process. I like to separate those conditions and try changing them. I bring the same approach to AI tools.

Read about my background and working habits

Technology & ideas

Why I think this way. What I question next.

Explore ideas
  1. AI systems & human collaborationMethod to be tested6 min readWhat has an agent automated if someone must keep watching it?Background execution is a beginning. Configuration, supervision, recovery, and interruptions must count when evaluating whether automation makes work easier.Read essay
  2. Behavior, measurement & interpretationAnalysis6 min readWhat survives when we turn behavior into labels?An action can be recognized correctly and interpreted incorrectly. What coding schemes preserve, what they exclude, and why behavioral analysis needs a route back to context.Read essay
  3. Core argumentAnalysis8 min readEvery problem is a technical problemBy technical, I mean examining goals, constraints, and relationships, then reorganizing what is available into an approach we can act on and test. How this shapes my research, use of AI, and collaboration.Read essay