What is COGO Genie?
A text to COGO accelerator
At COGO Genie, we believe in the power of making people’s lives easier.
We want AI and computers doing all the work while we enjoy life.
We would rather be making art, enjoying time with our kids, pets, and spouses, or riding motorcycles.
I have been mulling over this project for a while now. In my first full time position in GIS, I would enter COGO traverses into ArcMap.
I loved the work and routine of it. My fingers would fly over that number pad and slam the enter key with each entry.
But after seeing the patterns and consistencies in the legal descriptions, I knew there had to be a better way.
With some direction from a co-worker I began looking at Regular Expressions.
I’m not a coder. So this was all new to me. My coding experience to that point had been one Python course on Coursera. I think I still owe that professor my final project.
Seeing how I could tell a script to look at an input text and extract text based on a rule, made me think, COGO extract is possible.
Fast forward a few years of living and learning, and I started working on a language model that could pull the instructions from a legal description and convert it into COGO traverses.
As I began to learn more data manipulation techniques and machine learning became easier to deploy, I changed my approach.
I began developing a simple language model that can pull specific lines of text. This has been developed into our first tool, the Text to COGO tool.
A mess goes in.
Clean COGO instructions come out.
This is what we want our tools to do for you.
We want our hard work to make your lives easier.
Our goal for COGO Genie is to save you time. You should be working on more important things then transcribing a legal description from a recorded document into an electronic entity.
We want you to be able to use that data with confidence to maintain a historic record.
Our model is being trained on public documents from different agencies. They range from easements to deeds to quitclaim. From the year 2024 to 1990.
Handwritten to typed.
But we will need your help to improve our tools so they can do more for you. If you would like to share your unique documents that our tools are stumped by, please do.
We will use these to improve and give you a better product in the end.
Follow us on LinkedIn and join our newsletter for all the latest news and developments.
Thank you for being a part of our story. Let us help you. So you can enjoy yourselves.
Greg Kaganyuk, Co-Founder
gregkaganyuk@gmail.com