Lessons from Barrett Nash’s guest lecture at the University of Calgary’s ENTI 407 – NoCode Meets AI
“It’s not just about building a chatbot anymore—it’s about building a team.”
— Barrett Nash, CEO of InfiniteUp
🧭 Introduction
For decades, building advanced software meant wrangling code, backends, frameworks, and frustration. But something quietly revolutionary is happening in tools like FlutterFlow and OpenAI’s Assistant API: developers—and non-developers—can now build AI-driven product teams using drag-and-drop components and natural language.
In a recent guest lecture at the University of Calgary’s ENTI 407 course (Digital Entrepreneurship Technology – NoCode Meets AI), serial entrepreneur Barrett Nash broke down exactly how his startup InfiniteUp is doing just that—creating functional AI “employees” using FlutterFlow, a handful of APIs, and a lot of iteration.
This article unpacks that approach.
⚙️ What Are We Actually Building?
Let’s define terms first. This isn’t about adding a chatbot to your app. What Barrett’s team is doing through their product DOER Business is:
- Creating COO and CEO AI agents
- Automating task tracking, briefings, and sales strategies
- Using OpenAI’s function-calling + real-time inputs
- Building entire apps using FlutterFlow as the frontend, Firebase as the database, and OpenAI for intelligence
The result is a full-featured, responsive business tool where AI agents act like digital employees—coordinating, assisting, and even making strategic suggestions.
🧩 The Tech Stack: FlutterFlow + OpenAI + Firestore
Let’s zoom into the architecture.
1. Frontend: FlutterFlow
FlutterFlow is the visual IDE. Instead of writing Dart, Barrett’s team builds:
- Multi-screen responsive UIs
- Conditional logic via action trees
- Component libraries for repeatable UX patterns
- Visual links to Firestore and external APIs
This gives them a scalable, flexible way to test, redesign, and ship quickly.
2. Backend: Firebase / Firestore
Used for:
- User authentication
- Realtime data storage
- Structured AI feedback loops
It’s fast, serverless, and works cleanly with FlutterFlow’s native integration.
3. Intelligence Layer: OpenAI Assistants
Using the Assistants API, InfiniteUp builds “employees” with:
- Custom natural language prompts
- Embedded files and memory
- Function calling for executing real tasks in the app
- Optional web search and vision integrations
A single agent might handle generating sales pitches, sending legal contracts, or monitoring KPI changes.
🛠️ Use Case Walkthrough: Sales Team Built from AI
Here’s how a real-world use case works in DOER Business:
👩🏽💼 A user wants help selling to a local business.
- They choose a target (e.g. “Tina’s Nail Salon”) from Google Maps via API.
- AI assistant fetches salon info, current offers, and location data.
- A custom sales strategy is generated, including messaging.
- The app tracks whether that strategy is acted on.
- If not? The AI recommends follow-ups or asks if the pitch needs refining.
This flow uses:
- Firestore to store user state
- FlutterFlow for UI and logic
- GPT-4o-mini to generate dynamic text and strategy
- Function hooks to modify app state in real-time
🧠 Architectural Insight: Digital Twins & Agent Hierarchies
One of Barrett’s core takeaways is that AI works better in teams than solo. This inspired InfiniteUp’s digital org chart approach.
- A CEO agent oversees daily briefings and priorities
- A COO agent routes tasks to other agents
- Specialized agents do sales, legal, scheduling, etc.
Each assistant is trained in a narrow scope. They communicate through logic in the app—not through direct AI-to-AI messaging (yet). This helps maintain clarity and performance.
🔍 Why FlutterFlow?
Barrett compared FlutterFlow to platforms like Adalo and Bubble:
- ✅ Own your code – Apps export clean Flutter code
- ✅ Not locked-in – No vendor database dependency
- ✅ Highly customizable – Perfect for team-based collaboration (frontend, backend, UX, QA)
His caveat? It’s harder than Bubble or Adalo—but worth it. The team structure at InfiniteUp reflects this:
| Role | Tool Focus |
|---|---|
| Frontend / Designer | FlutterFlow |
| Backend Engineer | Firestore, API integration |
| QA / Testing | User testing, AI validation |
| Barrett (CEO) | System design + prompt engineering |
💡 Practical Takeaways
Thinking of building your own AI-powered tool in FlutterFlow? Here’s where to start:
- Design your AI org chart first – What roles will your agents play?
- Use OpenAI’s Assistant API – It’s made for integration.
- Rely on natural language – Write long, clear prompts. Use embedded documents.
- Use action trees in FlutterFlow to wire logic to real AI behavior.
- Iterate like a startup – Build–measure–learn still applies.
🎓 Final Thoughts
Barrett’s presentation was part technical demo, part philosophy session. His core belief?
“We’re not building apps. We’re building business partners. AI that works with you—not instead of you.”
The future isn’t just AI-enabled software. It’s AI-enabled teams. And for anyone with a laptop, a browser, and a few hours on FlutterFlow, that future starts now.
📘 Course Reference
This article is based on a guest lecture in:
ENTI 407 – Digital Entrepreneurship Technology: NoCode Meets AI
Professor: Andishe Ashjari, University of Calgary
Want to bring AI to your business? Setup an exploratory call with InfiniteUp CEO Barrett Nash here or reach out at nash@infiniteup.dev