A fundamental design choice in the creation of the AzoraOne API is to isolate the algorithm's learnings on a company level. This choice paves the way for hyper precise, company specific categorization of costs. This is important in bookkeeping since the purchase of the same product from a particular vendor can be categorized in many different ways, depending on how the purchasing company is using the particular product in their business. Another aspect is the level of detail; one company might want to categorize a certain purchase on one account, another company might want to split this purchase on many accounts.
Since AzoraOne's learning is based on two factors;
a) what can be extracted from a documents and;
b) the historical data-entry choices in similar, prior situations;
- the algorithm will understand the slight differences in data entry that separate different companies. This is an important difference between the AzoraOne API service and the typical Machine Learning solution based solely on mass population data.
One example would be to look at it from an accounting firm perspective; your customer, being an accounting firm, will benefit the automation of the base data extraction as well as the company specific categorization. As they bookkeep for different customers of theirs, your software will - thanks to AzoraOne's company specific categorization - learn the slight differences in the choice of categories/accounts for the different customers.
Another example of value for you as an integration partner would be to create value from the higher level of detail and clone this knowledge obtained by AzoraOne by using the resource Progenitors - for example you could create industry specific knowledge, shared between companies in your customer base.