Back to portfolio
Decision Support · Cross-functional

Persona Prioritizer

Persona-Based Decision Simulator

Generic advice lacks context → this AI applies a defined persona to evaluate decisions → users get recommendations aligned with a specific mindset.

01The Problem

Advice and recommendations are often too generic to be actionable because they ignore the decision-maker’s unique context, values, and priorities. This creates friction when users need perspective-specific guidance rather than broad best practices. Without structured persona application, outputs lack consistency and alignment with how a real individual would think and decide.

02What the AI Does

* Analyzes a structured persona description (background, values, preferences, decision style, goals). * Interprets subject matter through that persona’s lens. * Generates decisions, priorities, and recommendations consistent with the persona. * Enforces alignment by avoiding traits not explicitly stated. * Adapts tone and communication style to match the persona. Built on GPT-5.3 with custom system instructions that: * Define a two-input framework (persona + subject matter). * Constrain outputs to persona-grounded reasoning. * Prohibit assumption-making beyond provided data. * Structure outputs into decisions, priorities, and recommendations.

03Design Decisions

01 · Choice

Two-input structure (persona + subject matter)

Why

Ensures clear separation between context (who is deciding) and problem (what is being decided), enabling consistent application of perspective.

Constraint

Prevents vague or blended prompts that dilute persona fidelity.

02 · Choice

Explicit persona decomposition (background, values, preferences, decision style, goals)

Why

Breaks down “persona” into actionable variables the model can systematically apply rather than treating it as a loose narrative.

Constraint

Forces structured reasoning instead of surface-level mimicry.

03 · Choice

Strict prohibition on adding unstated traits

Why

[Creator: add rationale]

Constraint

Maintains fidelity and prevents hallucinated personality traits.

04 · Choice

Output framework (decisions, priorities, recommendations)

Why

Standardizes outputs into actionable categories rather than freeform text.

Constraint

Ensures clarity and usability across different use cases.

05 · Choice

Tone alignment requirement

Why

[Creator: add rationale]

Constraint

Prevents mismatch between persona identity and communication style.

06 · Choice

No personal opinions beyond persona

Why

Reinforces that the system is a simulation tool, not an independent advisor.

Constraint

Eliminates model-default general advice that conflicts with persona.

07 · Choice

Instruction to avoid unnecessary jargon unless persona supports it

Why

Keeps outputs consistent with persona communication style and accessibility.

Constraint

Prevents overcomplication or artificial sophistication.

08 · Choice

Emphasis on clarity and actionable outputs

Why

[Creator: add rationale]

Constraint

Avoids abstract or purely descriptive responses.

05Key Insight

AI becomes more useful for decision support when constrained to a specific perspective rather than optimized for generic correctness.