Virtual Personas Overview (Beta)

Have more questions? Submit a request

Virtual Personas is an AI-powered feature that allows teams to create and interact with simulated representations of their target audience segments.

This feature is designed to support early discovery work by helping teams explore ideas, test assumptions, and refine hypotheses before launching formal research.

Virtual Personas complement human research. They do not replace interviews, surveys, or live studies.

What Is a Virtual Persona

A Virtual Persona is an AI-generated representation of a defined audience segment. It is built using:

  • Background information provided by the researcher

  • Summaries of past Discuss research (human-led interviews and self-paced activities)

  • Uploaded knowledge documents (PDF format)

  • Structured personality and attribute frameworks defined by the researcher

Once created, users can chat with the persona in natural language to explore reactions, perspectives, areas of agreement, and areas of divergence within the segment.

The persona responds as a collective representation of the underlying data, not as an individual participant.

Virtual Personas are intended for:

  • Early idea exploration

  • Message and concept refinement

  • Hypothesis generation

  • Identifying alignment, disagreement, or uncertainty within a segment

All responses are grounded only in the data used to build the persona. If information is not present in the source material, the persona will state that it is unavailable.

Why Virtual Personas exist

During early discovery, teams often need directional insight before committing time and budget to live research. Traditional research cycles can take weeks, which makes them difficult to justify when ideas are still evolving.

Virtual Personas introduce a pre-research layer. They help teams:

  • Pressure-test early concepts

  • Refine assumptions

  • Identify gaps in understanding

  • Prepare more focused questions for human research

The goal is to improve decision confidence before launching formal studies.

Who can create and access personas

Creation permissions

  • Researchers and Organisation Admins can create personas.

Access permissions

  • Personas are private by default.

  • The creator can make a persona public for other Researchers.

  • Organisation Admins can view and edit all personas within the organisation.

  • Viewers and Contributors do not have access.

Creating a Virtual Persona

  1. Open the Virtual Personas tab from the main navigation.

  2. Select Create Persona.

  3. Enter:

    • Persona name

    • Background description of the segment

  1. Optionally select past research to include:

    • Human-led interviews

    • Self-paced activity responses

  1. Generate the persona.

Persona generation may take several minutes while the system processes inputs and builds the internal AI profile.

Editing a Persona

Only the persona creator and Organisation Admins can edit a persona.

Editors can:

  • Update the name and background description

  • Add or remove selected research

  • Add or remove uploaded knowledge documents

When edits are made, the internal AI profile is updated. Future chats will reflect the revised inputs. 

Chatting with a Persona

Users with access can start a new chat from the Virtual Personas dashboard.

Each chat:

  • Is a separate conversation with its own history

  • Is private to the user who initiated it

The persona:

  • Responds as a collective representation of the data

  • Highlights dominant themes where alignment exists

  • Acknowledges divergence where opinions differ

  • Signals uncertainty where evidence is limited

  • States when requested information is not present in the data

Responses are generated in English only.

A disclaimer in the chat interface clarifies that outputs are simulated and based on provided data.

Psychology and behavioural foundations

Virtual Personas are informed by established behavioural science frameworks, including the Big Five personality model:

  • Openness to Experience

  • Conscientiousness

  • Extraversion

  • Agreeableness

  • Emotional Stability

Traits are scored on a 1–5 scale only when supported by evidence in the source data. If sufficient evidence does not exist, the trait is omitted.

These frameworks help ensure personas remain consistent and interpretable across teams.

Validation approach

Virtual Personas are evaluated to ensure responses remain faithful to the human data used to create them.

Validation focuses on:

  • Alignment with source themes

  • Clear signalling of uncertainty

  • Avoiding fabricated or unsupported detail

  • Consistency across responses

Because personas evolve as new data is added, validation is ongoing.

Data handling, privacy, and retention

All data used to build Virtual Personas remains organisation-owned.

Data usage

Data is used only to:

  • Generate analytical summaries

  • Create and update the AI Persona Profile

  • Produce persona responses

No external datasets are introduced.

Model usage

  • Discuss uses the OpenAI API via secure, encrypted connections.

  • Organisation data is not used to train AI models.

  • Discuss operates under a Zero Data Retention agreement with OpenAI.

Retention

  • Derived summaries are anonymised and exclude PII.

  • Source data retention follows organisation-configured policies.

  • If source data is deleted, anonymised derived insights may persist in accordance with contractual and regulatory standards.

Was this article helpful?
0 out of 0 found this helpful