Skip to content

Chapter 2: The Persona Mental Model

Learning Objectives

By the end of this chapter you will be able to:

  1. Explain what a persona is in the FCC framework and why it matters.
  2. Name all ten components of the R.I.S.C.E.A.R. specification and describe each in one sentence.
  3. Distinguish between core personas, plugin personas, and champion personas.
  4. Describe how persona dimensions, the Discernment Matrix, and Design Target Factors add depth beyond R.I.S.C.E.A.R.

The mindmap below groups a sample of FCC's persona catalog by category so newcomers can see how the 102 core plus 23 plugin personas are organised.

mindmap
  root((FCC Personas<br/>102 Core + 23 Plugin))
    Core
      Research Analyst
      Strategy Lead
      Technical Writer
    Integration
      API Designer
      Data Pipeline Engineer
    Governance
      Ethics Auditor
      Compliance Officer
    Stakeholder
      Product Owner
      Business Sponsor
    Champion
      Research Champion
      Delivery Champion
    Data Engineering
      Data Curator
      Data Quality Analyst
    ML Lifecycle
      ML Engineer
      Model Evaluator
    DevOps
      CI/CD Engineer
    Protocol Engineering
      A2A Skill Designer
      MCP Tool Architect

Notice that categories cut across FCC phases: Data Engineering and ML Lifecycle personas, for instance, can appear in Find, Create, or Critique roles depending on the workflow graph that activates them.

Why Personas?

A persona is a structured identity that constrains an AI agent's behavior. Without personas, agents operate with unbounded latitude -- they decide for themselves what to research, how to write, and which standards to apply. That latitude is a liability in multi-agent systems because it makes behavior unpredictable and outputs inconsistent.

FCC personas solve this by making every behavioral constraint explicit. A persona definition answers ten questions before the agent generates a single token:

  1. What is your Role?
  2. What Input do you expect?
  3. What Style should your output follow?
  4. What Constraints must you respect?
  5. What is your Expected Output?
  6. What Archetype do you embody?
  7. What are your Responsibilities?
  8. What Skills do you bring?
  9. Who are your Collaborators?
  10. What is your Adoption Checklist -- the steps a human or agent takes to "become" this persona?

These ten questions form the R.I.S.C.E.A.R. specification (Role, Input, Style, Constraints, Expected output, Archetype, Responsibilities -- plus Role Skills, Role Collaborators, and Role Adoption Checklist).

R.I.S.C.E.A.R. in Detail

Role

A concise statement of what the persona does. Example: "Analyzes market data to identify trends and competitive positioning relevant to the project's strategic goals." The Role field is the persona's identity in one sentence.

Input

What the persona needs to begin work. This might be "a set of market research reports" or "the output of the Requirements Engineer persona." Declaring inputs explicitly prevents the common failure mode where an agent starts generating content without adequate context.

Style

How the persona communicates. A Research Analyst might use "structured, data-driven prose with citations." A Software Architect might use "technical documentation with diagrams and decision rationale." Style constraints ensure outputs are consistent across simulation runs and compatible with downstream consumers.

Constraints

Hard boundaries on behavior. Constraints might include "do not make recommendations -- only report findings" or "all code examples must include error handling." These are the guardrails that keep the persona within its lane.

Expected Output

The concrete artifact type the persona produces: "a Markdown report," "a Python module with test coverage," "a scored review with pass/fail determination." Expected Output tells the workflow engine what to look for when deciding whether the persona has completed its work.

Archetype

The behavioral archetype the persona embodies. FCC uses a vocabulary of 37 archetypes (e.g., "The Analyst," "The Builder," "The Guardian") that map to common professional roles. The archetype provides a human-readable shorthand for the persona's behavioral tendencies and is used by the simulation engine to generate persona-appropriate prompts.

Responsibilities

A list of specific duties, more granular than the Role. Where the Role says "reviews code," the Responsibilities might list: "identify security vulnerabilities," "check for style guide compliance," "verify test coverage meets threshold," and "flag architectural violations."

Role Skills

The technical and domain skills the persona brings. These are used by the cross-reference matrix to determine which personas can collaborate effectively and by the simulation engine to select appropriate personas for a given task.

Role Collaborators

Explicit upstream and downstream dependencies. The Research Analyst's collaborators might include the Data Curator (upstream) and the Strategy Lead (downstream). These declarations are compiled into a queryable cross-reference matrix (see Chapter 6).

Role Adoption Checklist

A sequence of steps for activating the persona. For a human adopting the persona, the checklist might read: "1. Read the project brief. 2. Load the relevant dataset. 3. Apply the analysis template." For an AI agent, it might read: "1. Retrieve the latest findings. 2. Set the temperature to 0.3. 3. Use the structured output schema."

The Persona Catalog

The FCC framework ships with 102 core personas, 45 vertical personas, and 23 plugin-contributed personas across 20 core categories and 6 vertical packs. The categories reflect the breadth of the framework's applicability:

Category Count Examples
Core 5 Research Analyst, Strategy Lead, Technical Writer
Integration 8 API Designer, Data Pipeline Engineer
Governance 4 Ethics Auditor, Compliance Officer
Stakeholder 5 Product Owner, Business Sponsor
Champion 4 Research Champion, Delivery Champion
Data Engineering 6 Data Curator, Data Quality Analyst
ML Lifecycle 9 ML Engineer, Model Evaluator
ML Models 11 Feature Engineer, Model Architect
DevOps 3 CI/CD Engineer, Infrastructure Analyst
App Development 2 Frontend Developer, Backend Developer
Others 50+ Open Science, Responsible AI, Privacy, Knowledge Graph, etc.

All personas are defined in YAML files under src/fcc/data/personas/ and loaded at runtime by the PersonaRegistry.

Champions

Four personas in the catalog carry a special designation: champion. A champion persona has two additional fields:

  • champion_of: The category it leads (e.g., "data_engineering").
  • orchestrates: A list of persona IDs it coordinates.

Champions are elevated coordinators. They do not replace the personas they orchestrate -- they sequence them, resolve conflicts between them, and escalate issues to governance. The champion architecture enables hierarchical team structures within the flat persona catalog.

Persona Dimensions

Beyond R.I.S.C.E.A.R., each persona can carry a 56-dimension profile organized into 9 categories:

  1. Core -- fundamental identity attributes
  2. Behavioral -- working style, risk tolerance, communication preferences
  3. Communication -- verbosity, formality, persuasion style
  4. Cultural -- collaboration orientation, hierarchy preference
  5. Decision-making -- analytical vs. intuitive, speed vs. thoroughness
  6. Professional -- domain expertise, seniority, breadth vs. depth
  7. Market -- customer orientation, competitive awareness
  8. Innovative -- creativity, experimentation tolerance
  9. Advanced -- meta-cognition, learning rate, adaptability

Each dimension is rated on a 1--7 scale. The profile enables nuanced behavior differentiation: two personas with the same Role but different dimension profiles will approach the same task differently.

Discernment Matrix and Design Target Factors

Two additional rating systems complement the dimension profile:

  • Discernment Matrix: Rates each persona on 6 traits (Humility, Professional Background, Curiosity, Taste, Inclusivity, Responsibility) across 7 rating dimensions. This captures the persona's judgment quality.
  • Design Target Factors: Rates 6 factors (Optimism, Social Connectivity, Influence, Diversity Appreciation, Curiosity, Leadership) across 7 rating dimensions. This captures the persona's collaborative orientation.

Together, R.I.S.C.E.A.R. + Dimensions + Discernment Matrix + Design Target Factors give the framework a rich, multi-layered model of each persona's identity, behavior, and collaborative fitness.

Key Takeaways

  • A persona is a structured identity that constrains AI agent behavior through ten explicit components.
  • R.I.S.C.E.A.R. stands for Role, Input, Style, Constraints, Expected output, Archetype, Responsibilities, plus Role Skills, Role Collaborators, and Role Adoption Checklist.
  • The catalog ships with 102 core + 45 vertical + 23 plugin personas across 20 core categories and 6 vertical packs.
  • Champions are elevated coordinators that orchestrate teams of base personas.
  • Persona Dimensions (56 attributes), the Discernment Matrix (6 traits), and Design Target Factors (6 factors) add behavioral depth beyond the R.I.S.C.E.A.R. specification.

Cross-References

← Chapter 1: What Is FCC? | Next: Chapter 3 -- Workflow Thinking →