Gylling Accounting

What was done?
Aiso’s team developed an AI-based solution to digitize PDF and paper invoices for Gylling Accounting’s business needs. In the solution, an independent web-solution and a digitizing engine were developed. The web-solution is used to archive and validate values found from individual invoices. The digitizing engine uses machine vision, advanced algorithms and AI to find values.

Why it was done?
Similar solutions used in the accounting company were not capable enough for the future needs. Outsourcing invoice digitalization, especially when volumes increase, is significant expense and there was no user friendly solution in the market that doesn’t use templates to digitize invoices. In solutions that use templates one needs to first create template for every single supplier to find values.

Main goal for Aiso was to make invoice validating process user friendly and remove the need for invoice templates entirely. The implemented solution brought a cost efficient and a pleasant way to digitize PDF and paper invoices in accounts payable process.

Technologies used:
Web:
React
NodeJS
PostgreSQL

Digitizing engine:
C++
Python
Keras
TensorFlow
Scikit-learn
OpenCV
Tesseract OCR

DevOps / Delivery:
Docker
GitLab CI/CD
Ansible
Apache2

Gylling Accounting’s comment about the project:

“Developed solution brings cost effectiveness to digitizing work and is clearly more user friendly than template based solutions found on the market. Creating templates would have brought continuing costs which can now be totally avoided. Aiso used many state of the art technologies which also brings other business opportunities. Aiso knows these technologies extremely well and understands how to use them effectively.”

– Max Gylling, CEO