NYU Entrepreneurial Institute 2025 AI Tools · Coaching Systems

Team Compass: Turning Coach Intuition into a Structured Strategic Document

A GPT-assisted tool that converts coach insights and application data into personalized SWOT reports for every team entering the Sprint Accelerator. The first time the Institute delivered this level of team-specific strategic analysis at scale.

First ever Team-specific strategic analysis at this depth for the program
3 audiences Founders, coaches, and program leadership all benefit from the same artifact
Zero info decay Post-interview insights captured and structured instead of lost in notes

The Problem

After each Sprint Accelerator intake, coaches came away with rich, nuanced thinking about every team. They had views on each founder's strengths, the risks in their market, the behaviors that would hold them back, and the milestones they should be targeting. This diagnostic work was real and valuable.

But it lived in call notes, Notion pages, and hallway conversations. There was no shared artifact that captured it, no consistent format for communicating it to founders, and no structured way to align the coaching team around common expectations before the program began.

The result was information decay. Hard truths that coaches thought about but rarely said out loud. Goal-setting that happened informally, if at all. Founders entering the program without a clear picture of where they stood or what they needed to fix.

"We were already doing the diagnostic thinking. We just weren't doing anything with it. This tool was about closing that gap."

What I Built

A lightweight, GPT-assisted workflow that takes structured coach input and produces a polished, team-specific SWOT report. The report is delivered in print at Sprint Orientation, followed by a dedicated deep-dive session with the assigned coach.

The tool is deliberately human-centered. Coaches fill out a structured insight form that prompts them to articulate what they are excited about, what they are worried about, and what they would tell the founders directly if they were being fully honest. GPT assembles a first draft from that input. Coaches review, edit where needed, and sign off before delivery.

AI amplifies the coach's judgment. It does not replace it.

How It Works

From Coach Insight to Founder Report

01
Coaches complete a structured insight form

The form is designed to surface the thinking coaches already have but rarely write down. Three prompts capture everything needed for a meaningful report.

Excited about: Strengths and opportunities the team has going for them
Worried about: Weaknesses and risks, especially around customer discovery
Notes to founders: Clear-eyed truths about behaviors, patterns, or risks that need to be said directly
02
GPT assembles a first-draft SWOT report

Using pre-defined formatting prompts and tone guidance, GPT-4o converts the coach input and application data into a structured four-quadrant SWOT report. The tone is direct, empathetic, and professional. No generic advice. Everything is specific to the team.

03
Coach review and edit

The assigned coach reviews the draft, adds nuance, and edits where the first pass missed the mark. This step is where the human judgment layer matters most. The tool earns trust by getting close enough that review is fast, not by eliminating it.

04
Delivery at Sprint Orientation

Printed reports are handed to each team at orientation, followed by a structured deep-dive session with their coach. Founders receive honest, well-articulated feedback early enough to act on it before the program is underway.

Why It Matters

One artifact, three audiences

For founders
Honest, specific feedback delivered early enough to actually change behavior. A referenceable document they can return to throughout the program rather than a one-time verbal debrief.
For coaches
A shared artifact that aligns the coaching team around common expectations, performance milestones, and support strategy for each team before a single session takes place.
For the program
A structured method to track founder growth and program performance over time. Transforms post-interview insight from information that decays into a record that compounds.

Tools and Stack

What I built it with

ChatGPT (GPT-4o)
Core generation layer. Structured SWOT prompts and tone guidance convert coach input into polished first-draft reports.
Notion
Houses the coach insight forms and program documentation used to inform each report.
Google Docs
Final report format. Docs are reviewed, edited, and exported to PDF for print delivery at orientation.

Watch It in Action

A 4-minute behind-the-scenes walkthrough of how the tool works and what the output looks like.

Lessons Learned

01 The quality ceiling is coach honesty. This tool is a rare opportunity to deliver hard truths that often go unsaid in live coaching sessions. But not every coach felt comfortable going there. Getting full value from the system required coaching the coaches, not just building the tool.
02 Reception validated the design philosophy. The tool was exceptionally well received by both the coaching team and the founders. Marking the first time the Institute delivered this level of team-specific analysis was the proof point that justified the approach and opened the door to deeper iterations.
03 The next version needs richer external data. Coach intuition is strong but not complete. Plugging the tool into the NYU Library system to pull market research and external context would meaningfully elevate the strategic quality of each report.
04 Human review is a feature, not a workaround. Keeping coaches in the loop before delivery wasn't a concession to skepticism. It was what made the output credible. Founders trust the report because they know a real coach reviewed and endorsed it.