AI Developer Accelerator
Log In
Community
Classroom
Calendar
Members
Leaderboards
About
Log In
5
Patrick Chouinard
13d •
General discussion
AI Developer Accelerator — Coaching Call - July 21st
AI Developer Accelerator — Coaching Call - July 21
VIEW RECORDING - 122 mins (No highlights)
Meeting Purpose
Reviewing AI projects and sharing development progress.
Key Takeaways
AgentTask Pitch Refinement: Alireza's AgentTask, an AI-native task manager, needs a clearer value proposition and user workflow demo to differentiate it from competitors like Notion and Linear.
AI Booth Success: Juan's AI photo booth generated 859 images for only $52, but revealed significant UX friction for non-tech-savvy users, highlighting the need for intuitive design.
Open-Source Training: Patrick open-sourced his "Training Generator," a tool that creates interactive, self-serving Claude Code training courses directly from a project's GitHub repo.
Notary Client Win: Patrick secured a notary client after Claude Cowork found critical errors in their legal documents, proving AI's value in highly regulated, text-based fields.
Topics
AgentTask Demo & Feedback
Alireza demoed AgentTask, an AI-native task manager built to be simpler than Notion or Jira.
Core Features:
Claim IDs: Logs which agent made each change for full auditability.
Crews: Agent blueprints (e.g., for lead generation) that sync with AI harnesses like Claude.
Notes: A knowledge base for reusable instructions and skills.
Pricing: A generous free tier (3 seats, 1k tasks) and affordable paid tiers.
Feedback: The group advised Alireza to refine the product's pitch.
Problem: The demo lacked a clear, step-by-step user workflow, making it hard to grasp the value proposition.
Solution: Use AgentTask's own AI to define an Ideal Customer Profile (ICP) and create a focused pitch that highlights unique value, not just price.
AI Booth Deployment & UX Learnings
Juan successfully deployed his AI photo booth at a quinceanera.
Performance: The system generated 859 images for only $52 in token costs.
UX Challenges:
User Confusion: Many adults and teens struggled with the workflow (e.g., scanning QR codes), requiring manual assistance. Kids, however, intuitively understood it.
Solution: Add a short, animated video loop to the display screen demonstrating the user flow.
Next Steps:
Stabilize the image-to-image pipeline.
Integrate a video generation pipeline (inspired by Scott's Higgs field work) to create a highlight reel of the event.
Open-Source Training Generator
Patrick open-sourced his Training Generator plugin for Claude Code.
Functionality: Generates an interactive, self-serving training agent from a project's GitHub repo.
Key Features:
Adaptive Content: Tailors the course for different audiences (e.g., "developer new to this repo" vs. "user").
Progress Tracking: Lets users resume training where they left off.
Live Exercises: Includes interactive demos that users can run within the training.
Use Case: Patrick uses it for developer onboarding, as the training auto-updates with repo changes.
Other Project Updates
Ty: Built a meal prep app for his daughter in one morning, using a USDA API for macro tracking.
Ryan: Migrated his business to Google Cloud and is fixing ~60 security issues found by a custom agent-based review tool.
Shakur: Is refining a website redesign agent to move beyond "AI slop" by using a feedback loop to train the model on desired vs. generic outputs.
Morgan: Is setting up EINs for his "Heritage Plot" (cemetery management) and "Cluster Curb" (school pickup) SaaS products, which will target annual PO contracts.
Morgan: Automated a publishing workflow, reducing the time to output a book in four formats from hours to 15 minutes.
Next Steps
Alireza: Refine the AgentTask pitch with a clear value proposition and a step-by-step user workflow demo.
Juan: Add an animated video loop to the AI booth display to guide users; stabilize the image-to-image pipeline.
Shakur: Implement a feedback loop to train the website redesign agent on desired vs. generic outputs.
Morgan: Use Patrick's Training Generator to create a course on using Capterra/G2 for competitive analysis.
Patrick: Evaluate the "Omni-gent" meta-harness for managing multiple AI harnesses.
Action Items
Draft ICP-aligned comms/pitch; record short workflow clips; update site; share in forum -
WATCH (5 secs)
Run meal-prep app through phishing scanner; fix Dashlane AI warning -
WATCH (5 secs)
Email Ryan C pitch outcome update -
WATCH (5 secs)
Use Training Generator to update internal training materials -
WATCH (5 secs)
Stabilize AI booth image-to-image pipeline; fix DB issues -
WATCH (5 secs)
Create 10s animated usage loop for AI booth display -
WATCH (5 secs)
Add cost calculator to Cluster Curb landing page -
WATCH (5 secs)
Use Training Generator to compile G2/Capterra pros/cons for Heritage Plot & Cluster Curb -
WATCH (5 secs)
Create skill to generate HTML course from Training Generator output -
WATCH (5 secs)
Like
1
0 comments
5
AI Developer Accelerator — Coaching Call - July 21st
AI Developer Accelerator
skool.com/ai-developer-accelerator
Master AI & software development to build apps and unlock new income streams. Transform ideas into profits. 💡➕🤖➕👨💻🟰💰
11.3k
Members
60
Online
8
Admins
JOIN GROUP
Leaderboard (30-day)
1
Ty Wells
+6
2
Tom Welsh
+5
3
Keano Reurink
+4
4
Shakur Abdullah
+3
5
Ali S
+2
See all leaderboards
Powered by