How We Built Soren: An AI Market Researcher
Discover how we built Soren, an AI-powered survey research tool that analyzes responses and generates 100+ insights per hour — fast, scalable, and zero human input. Learn about the tech stack, business model, and what’s coming next.
Posted by

Related reading
URL Parameters: Pass Custom Data Into Your Surveys
Improve your surveys with URL Parameters. Pass custom data to personalize experiences, track attribution sources, and streamline analysis
Introducing Dynamic Conditional Surveys for Better Research
Upgrade your research with Survey Sync's new conditional logic surveys. Design adaptive questionnaires, get AI-powered insights, and create engaging research workflows
Seamless Survey Integration for Mobile Apps
Boost user engagement and gather real-time feedback with Survey Sync’s React Native component. Perfect for developers aiming to optimize their mobile apps!

🚀 How We Built an AI Survey Tool That Generates 100+ Insights per Hour
Manual survey analysis is slow, expensive, and doesn’t scale.
So we built Soren — an AI-powered survey researcher that analyzes responses and delivers insights autonomously. 🤖✨
In this post, I’ll break down the system, tech stack, and business model — no fluff.👇
🧩 The Problem
Traditional survey workflows rely on human analysts to read through responses, find patterns, and write summaries.
That process is:
- 🐢 Time-consuming (days, not hours)
- 💸 Expensive (consultants, researchers)
- 🎲 Inconsistent (depends on who's doing the analysis)
We wanted something that could scale — without sacrificing quality. 💡
💡 The Solution: Soren
Soren is an AI market researcher.
You submit a survey and responses — it returns a detailed, structured insight report.
Fast, accurate, and zero human analysts. ⚡📊
⚙️ How It Works
- 📤 Upload or create a survey
- 📬 Collect responses (or let us handle distribution)
- 🧠 Soren analyzes using a high-reasoning OpenAI model
- 🔍 Extracts sentiment, key themes, and actionable insights
- 📄 You get a clean, readable report within minutes
🧱 Tech Stack
- Frontend: Next.js + React + DaisyUI 🌐
- Backend: Java + Spring Boot ☕
- Database: PostgreSQL 🐘
- Infrastructure: AWS ECS, RDS, S3 ☁️
- Authentication: JWT 🔐
- AI Engine: High-reasoning OpenAI model, enhanced with custom logic, prompt engineering, and post-processing 🧬
💳 Credits-Based Billing System
Soren runs on a credit system:
- 🎟 Clients buy credits in advance
- 🔄 Each request uses credits based on complexity
- ⏸ When credits run out, processing pauses
- ✅ No subscriptions, no contracts — just pay-as-you-go
Simple. Predictable. Scalable.
📈 Results
-
🥇 First client: €500/year (early design partner)
-
💰 New pricing: 4× higher for new customers
-
☁️ Monthly cloud cost: ~$160
-
🗣 Feedback:
“Feels like having a full research team in one click”
🔮 What’s Next
We're working on:
- 📝 Automatic survey brief generation
- 🗺️ Visual insight heatmaps
- 🔗 Integration with LinkedIn Ads
🚪 Want Access?
We’re selectively opening early access to Soren.
If you're building something similar — or want to try it out —
📩 Contact us here or DM me on X.
Let’s automate better decisions, together. 🙌