Northline

Inspiration

A lot of businesses are still doing repetitive work manually without really knowing where the problem is. Enquiries get copied into spreadsheets, follow-ups are done by hand, information is moved between different tools, and the same questions get answered repeatedly. I wanted to build something that could look at how a business actually operates and point out where automation would make the biggest difference.

What it does

Northline analyzes a business through its website, social profile, CSV data, or a simple description of how it works. It identifies manual processes, highlights what should be automated first, and generates a practical operations snapshot. It can also find real businesses using better workflows so users can see what is possible beyond generic AI advice.

How I built it

Northline was built with Python and FastAPI, with a research pipeline for collecting, filtering, and analyzing public business information. I separated deterministic processing from the parts where AI is useful for interpretation and drafting. The system also includes reporting, email delivery, search providers, SQLite storage, and an admin research workflow.

Challenges we ran into

The biggest challenge was messy and inconsistent business information. Websites and social profiles rarely describe their internal processes directly, and search results contain a lot of noise. We had to build filters and validation around that instead of assuming the information would be clean or complete.

Accomplishments that we're proud of

We’re proud that Northline is more than a chatbot or a generic AI recommendation tool. It takes real business information, turns it into an actionable view of manual work, and connects that analysis to real examples. We also made a conscious decision to never invent contact information when researching businesses.

What we learned

We learned that building a useful AI product is not just about adding an LLM. A good system needs reliable inputs, clear rules, filtering, validation, and a useful experience around the AI. The hardest part was deciding where AI actually added value and where a deterministic approach was better.

What's next for Northline

The next step is making Northline better at turning its findings into concrete automation plans. We want it to go beyond identifying manual work and help businesses understand exactly how that process could be redesigned, what tools could be involved, and what the implementation could look like.

Built With

• css
• fastapi
• html
• javascript
• llm
• openai
• python
• restapi
• webscraping