IDA AI Storytelling Coach

StoriesSeen partnered with me to design and deliver IDA, the first contextually-aware storytelling coach built for underestimated founders. The product combines the StoryEdge Method™ with the Elaboration Likelihood Model (ELM) to provide AI-assisted narrative intelligence that understands who founders are, not just what they say.

Approach

Worked as a fractional technical lead (Apr through Jul 2025), partnering closely with the founder to translate qualitative storytelling frameworks into a structured AI product. The core challenge was building an analysis pipeline that could reflect her human coaching voice while delivering consistent, rubric-based feedback at scale.

Highlights

Role: Fractional Technical Lead & Full-Stack Engineer

• Delivered the end-to-end MVP enabling founders to upload recordings, get transcripts, and receive structured narrative feedback.
• Built the AI foundation: orchestrated speech-to-text with OpenAI Whisper and designed a GPT-4o analysis pipeline aligned with StoryEdge rubrics.
• Translated qualitative storytelling frameworks into structured outputs via a JSON-based scoring schema grounded in persuasion and communication theory (ELM).
• Co-designed the admin dashboard and data model for future prompt and rubric management, ensuring the MVP could evolve without rework.
• Iterated on prompt tone, rubric calibration, and UX with the founder to ensure IDA reflected her coaching methodology.

Tech Stack

Next.js 14, Supabase (Auth + Postgres), Vercel, Tailwind/ShadCN UI, OpenAI Whisper, GPT-4o, TypeScript

Outcomes

Delivered the complete MVP by July 2025, enabling StoriesSeen’s founder cohort to validate the product vision before a broader launch. Established a clear technical foundation for the roadmap toward an identity-aware storytelling intelligence platform.