Zena Zerai
GWI
In flight project · GWI

Designing self-serve AI-simulated research

GWI is exploring a move towards self-serve AI-simulated research, enabling customers to create surveys using respondents modelled on GWI’s consumer data.

The challenge is balancing a straightforward research experience with the consumption-based model behind it — from understanding results to managing credits, access and usage.

Working with User Management, Data and engineering across both teams.

Survey creation screen for self-serve AI-simulated research, showing a question, a sample result and organisation credits.

Different users, different needs

Needs being explored

Researcher creating a survey.

Researcher / survey creator

  • Create useful surveys independently
  • Configure audiences and questions
  • Understand and interpret results
An administrator reviewing usage, credits and access.

Organisation / account administrator

  • Understand consumption and usage
  • Allocate credits across users
  • Manage access and sharing

The challenge is making research creation feel effortless while giving organisations the visibility and control they need over a consumption-based product.

Understanding the underlying product

User Management × Data × Engineering

Working across User Management and Data, I’m connecting both sides of the product: understanding what customers need from an AI research experience, and what’s needed to make self-serve work within GWI’s wider platform and commercial model.

Understanding the AI

Working with Data and engineering to understand the model’s capabilities and limitations, and what those mean for survey creation and interpreting results.

Making self-serve work

Building on existing User Management discovery and working with engineers across both teams to explore usage caps, credit allocation and access controls.

This builds on earlier API/MCP usage dashboard work. Once the commercial model and requirements are agreed, there may be an opportunity to extend that experience for visibility of simulated survey consumption.

Design explorations

Prototype

  1. 01

    Survey management

    Prototype survey list with statuses and an organisation credit summary.

    Exploring how to make survey setup clear and manageable.

  2. 02

    Results and notifications

    Prototype home screen with a notification that survey results are ready.
    Prototype dialog that a published audience is ready to use in Platform.

    Exploring how to make simulated results understandable and useful.

  3. 03

    Credits and consumption messaging

    Prototype credit dialog over survey setup.
    Prototype survey setup with a message that free credits have been used.

    Credits and consumption messaging exploration.

Prototype walkthrough

A working prototype exploring the journey from panels and survey setup through to results.

Panels · Survey setup · Results

What happens next

Planned next steps

  1. 01

    Align on the commercial model

    Work with the business to understand how simulated survey consumption will work. Will usage be tied to existing product consumption or billed separately?

  2. 02

    Work through the platform implications

    Collaborate with engineers across User Management and Data to explore what the agreed model means for usage caps, credit allocation across users and access controls.

  3. 03

    Validate with customers

    Test the proposed experience with enterprise clients through GWI’s innovation partners programme, exploring survey creation, interpretation of simulated results, and expectations around consumption, sharing and cost visibility.

Validation focus

Customer testing will help us understand whether people can create suitable surveys independently, interpret simulated results appropriately, and understand how consumption and access should work.

Email copied to clipboard