What was the scope of their involvement?
Lemay.ai undertook the data science to build AI/machine-learning models that can be trained on how to recognize and automatically classify business documents. We started by putting together a matrix of business functions that roughly translate to departments within a company. Within that department are types of documents they might use on a regular basis. For example, in the finance department or as a finance business function there would be invoices and purchase orders under procurement and meeting minutes under administration. As part of the proof of concept with Lemay.ai, we centered on 10–15 types of documents and collected a large sample set for each.
Once the model was built, Lemay.ai tested its capability. Together we went through probably 5–6 iterations. One of our senior software developers worked on understanding how the technology worked in order to be self-sufficient by the time the project wrapped.
What is the team composition?
Initially we worked with Mathieu (CEO and Co-Founder, Lemay.ai) on scoping the project. Once we were up and running, they assigned us to their chief data scientist. Daniel (CTO, Lemay.ai) worked with us on a regular basis on all of the technical aspects of getting the models built and viewing results.
How did you come to work with Lemay.ai?
I wasn't personally involved in their selection. I believe our companies first connected at a local technology forum.
How much have you invested with them?
We spent between $50,000–$100,000 CAD (approximately $37,500–$75,000 USD).
What is the status of this engagement?
Discussions began the first week of June 2019 and the project wrapped in mid-August.