Use Case

ATHENA

Venture Support & Scoring Platform — UX Design, Prompt Engineering, Full-Stack Development

Athena is a WordPress-based venture support and investor hub with an AI scoring engine that evaluates startup ideas using structured LLM reasoning and real-time Perplexity research. I designed the UX and built the scoring workflow, prompts, and logic integrations that turn slow, manual research into fast, consistent, insight-driven evaluations.

Use Case

ATHENA

Venture Support & Scoring Platform — UX Design, Prompt Engineering, Full-Stack Development

Athena is a WordPress-based venture support and investor hub with an AI scoring engine that evaluates startup ideas using structured LLM reasoning and real-time Perplexity research. I designed the UX and built the scoring workflow, prompts, and logic integrations that turn slow, manual research into fast, consistent, insight-driven evaluations.

Use Case

ATHENA

Venture Support & Scoring Platform — UX Design, Prompt Engineering, Full-Stack Development

Athena is a WordPress-based venture support and investor hub with an AI scoring engine that evaluates startup ideas using structured LLM reasoning and real-time Perplexity research. I designed the UX and built the scoring workflow, prompts, and logic integrations that turn slow, manual research into fast, consistent, insight-driven evaluations.

Use Case

ATHENA

Venture Support & Scoring Platform — UX Design, Prompt Engineering, Full-Stack Development

Athena is a WordPress-based venture support and investor hub with an AI scoring engine that evaluates startup ideas using structured LLM reasoning and real-time Perplexity research. I designed the UX and built the scoring workflow, prompts, and logic integrations that turn slow, manual research into fast, consistent, insight-driven evaluations.

My Role

Lead & Solo Designer Full-Stack Developer Testing & QA

My Role

Lead & Solo Designer Full-Stack Developer Testing & QA

My Role

Lead & Solo Designer Full-Stack Developer Testing & QA

Client

R-LABS Canada

Client

R-LABS Canada

Client

R-LABS Canada

Team

CEO - George Carras Head of Technology - Roy Liu Venture Lead - Jameel Somji Brand Designer - Mae Tsui Venture - Siobhan Byrne

Team

CEO - George Carras Head of Technology - Roy Liu Venture Lead - Jameel Somji Brand Designer - Mae Tsui Venture - Siobhan Byrne

Team

CEO - George Carras Head of Technology - Roy Liu Venture Lead - Jameel Somji Brand Designer - Mae Tsui Venture - Siobhan Byrne

Year

2024

Year

2024

Year

2024

Project Details

The Problem

Although I designed and built the entire venture platform, this case study focuses on the most impactful feature: the AI-driven scoring engine powered by the Perplexity API. The team needed a way to evaluate more than 100 startup ideas quickly—something their current process was having difficulty with.

The key question became:
Could AI make early-stage venture evaluation scalable and consistent?

The Solution

I designed an AI-powered scoring system that evaluates venture ideas in minutes, using structured reasoning, research citations, and R-LABS criteria.

It enables faster decision-making, consistent evaluation, and a shared space to compare and validate opportunities.

Project Details

The Problem

Although I designed and built the entire venture platform, this case study focuses on the most impactful feature: the AI-driven scoring engine powered by the Perplexity API. The team needed a way to evaluate more than 100 startup ideas quickly—something their current process was having difficulty with.

The key question became:
Could AI make early-stage venture evaluation scalable and consistent?

The Solution

I designed an AI-powered scoring system that evaluates venture ideas in minutes, using structured reasoning, research citations, and R-LABS criteria.

It enables faster decision-making, consistent evaluation, and a shared space to compare and validate opportunities.

Project Details

The Problem

Although I designed and built the entire venture platform, this case study focuses on the most impactful feature: the AI-driven scoring engine powered by the Perplexity API. The team needed a way to evaluate more than 100 startup ideas quickly—something their current process was having difficulty with.

The key question became:
Could AI make early-stage venture evaluation scalable and consistent?

The Solution

I designed an AI-powered scoring system that evaluates venture ideas in minutes, using structured reasoning, research citations, and R-LABS criteria.

It enables faster decision-making, consistent evaluation, and a shared space to compare and validate opportunities.

Project Details

The Problem

Although I designed and built the entire venture platform, this case study focuses on the most impactful feature: the AI-driven scoring engine powered by the Perplexity API. The team needed a way to evaluate more than 100 startup ideas quickly—something their current process was having difficulty with.

The key question became:
Could AI make early-stage venture evaluation scalable and consistent?

The Solution

I designed an AI-powered scoring system that evaluates venture ideas in minutes, using structured reasoning, research citations, and R-LABS criteria.

It enables faster decision-making, consistent evaluation, and a shared space to compare and validate opportunities.

The Journey at a Glance

Problem

Early-stage investors spent hours manually validating startup ideas using documents, scattered research, and gut instinct.

Problem

Early-stage investors spent hours manually validating startup ideas using documents, scattered research, and gut instinct.

Opportunity

Strong ideas were slipping through, slowing decisions and reducing team alignment.

Opportunity

Strong ideas were slipping through, slowing decisions and reducing team alignment.

Concept

R-LABS asked: What if AI could generate and score ideas to handle 80% of early validation automatically?

Concept

R-LABS asked: What if AI could generate and score ideas to handle 80% of early validation automatically?

My Role

I led the UX design, prompt architecture, and full-stack development of the Athena AI scoring system.

My Role

I led the UX design, prompt architecture, and full-stack development of the Athena AI scoring system.

Build

Using the Perplexity API, I created a structured LLM workflow, stored insights in WordPress database, and built real-time scoring with consistent output logic.

Build

Using the Perplexity API, I created a structured LLM workflow, stored insights in WordPress database, and built real-time scoring with consistent output logic.

Result

Idea evaluation dropped from days to minutes, giving the team fast, research-backed insights for better decision-making.

Result

Idea evaluation dropped from days to minutes, giving the team fast, research-backed insights for better decision-making.

Architecture

Athena is a full venture support platform, but this sitemap focuses on the flow that powers the AI idea-generation and scoring system. It visualizes how users move from:

inputs → AI processing → structured scoring outputs

and how the LLM logic integrates into the surrounding application screens.

Score Details

1. Sketch

Explored early layouts for venture scoring.

Introduced repeatable cards for Problem, Solution, and Viability.

Tested basic actions like Delete for managing AI results.

2. Wireframe

Refined hierarchy and scan-ability of AI-generated content.

Added a Rerun action to regenerate scores.

Removed Like/Thumb actions to avoid misleading AI authority.

3. Mockup

Added color-coded score indicators for quick evaluation.

Introduced breadcrumbs for navigation.

Grouped primary actions to reduce cognitive load.

4. MVP

Expanded actions based on early testing.

Removed Rerun/Delete below navigation to encourage report review.

Separated AI features from Data Room content.

5. Final

Added a “?” tooltip explaining AI scoring criteria to improve transparency and trust.

Score List

1. Sketch

Explored repeatable score-card layouts for AI ideas.

Sketched Problem, Solution, and Viability cards to match the scoring model.

Included Like/Thumb actions for early team voting.

2. Wireframe

Defined hierarchy and interactions with lo-fi wireframes.

Added a Rerun action to regenerate scores during prompt testing.

Removed voting once ideas moved to the venture evaluation process.

Improved separation to make ideas easier to compare.

3. MVP

Moved to a sortable, filterable table for faster comparison across many ideas.

Shifted inputs into a modal to reduce clutter.

Added color-coded score indicators for quick prioritization.

4. Final

Added a circular progress state to show AI processing.

Introduced quick Delete actions for low-value ideas.

Grouped the table in a card and separated AI tools in navigation.

Final Design

Generate Idea Modal

Users select how many venture ideas to generate, with limits to control API usage and maintain system stability.

Idea List

AI-generated ideas appear in a table with checkboxes for selecting which ideas to score.

Ideas are visible only to the logged-in user and are removed once scoring begins to keep the list clean.

Scored Idea List

Scored ideas appear in a sortable, filterable table for fast comparison.

Search and color indicators (green/yellow/red) help users quickly identify stronger concepts.

The Details link is disabled while the AI processes results to prevent early interaction.

Score Idea Modal

A scoring modal allows team members to run AI evaluations.

Supports scoring ideas individually and is designed to later allow external submissions.

Venture Details Page

Displays the full scoring breakdown based on the R-LABS evaluation model.

Users can re-score, print, or copy the report.

Citations provide transparency and allow teams to verify AI-generated insights.

Key Takeaway

95% Faster Validation

Reduced research and idea-scoring time from 3 days to 10 minutes using structured LLM reasoning and automated scoring.

95% Faster Validation

Reduced research and idea-scoring time from 3 days to 10 minutes using structured LLM reasoning and automated scoring.

Adopted by the Industry Issue + Transformation (I+T) Council

Used internally to demonstrate platform innovation and strengthen the company’s AI strategy.

Adopted by the Industry Issue + Transformation (I+T) Council

Used internally to demonstrate platform innovation and strengthen the company’s AI strategy.

Faster Team Alignment

Standardized scoring eliminated inconsistencies and helped the venture team compare ideas and make decisions more quickly.

Faster Team Alignment

Standardized scoring eliminated inconsistencies and helped the venture team compare ideas and make decisions more quickly.

Future-Ready for Investors

The platform now supports a roadmap for deeper validation, multi-team access, and external investor onboarding.

Future-Ready for Investors

The platform now supports a roadmap for deeper validation, multi-team access, and external investor onboarding.

Let’s Connect

Open to product design opportunities and collaborations.

© 2026 Sae-Hee Shin. All rights reserved.

Let’s Connect

Open to product design opportunities and collaborations.

© 2026 Sae-Hee Shin. All rights reserved.