The science behind Afiniti
The Afiniti process starts with a customer’s caller ID or unique identification number. Upon placing a call or making contact with a business through chat, email, or another medium, Afiniti uses this key to start its data-chaining process using a set of pre-determined sources, including call history and CRM data for the caller.
In this process we combine interaction-level outcomes data from our clients with hundreds of internal and external databases through highly complex data joins in order to build a rich contextual data source for each interaction. Using this data, Afiniti deploys specialized machine learning techniques to identify behavioral patterns in customer-representative interactions that lead to success.
We look at all of the interactions from our client’s databases and determine whether or not they were successful. We then analyze all of the contextual data that we have access to (e.g.: demographic data, previous interactions, and any available internal analytics) and use algorithms we have developed to look for patterns which can in some way predict the outcomes (e.g.: if a customer has had multiple unsuccessful calls with customer care, they are very likely to want to cancel).
The number of interactions, however, is relatively low when compared to the number of data points generally required for most AI algorithms, meaning that relying solely on machine learning techniques can lead to unreliable results. In order to solve this, Afiniti has developed highly sophisticated Bayesian analytical methods combined with proprietary heuristics to further refine our algorithms and adapt them to the scarce data environment.
Isolating the causes of a successful interaction is only the first step. We cannot know when a customer will call in, and equally we cannot know which agents will be available. This requires us to run the algorithm in real-time whenever a customer contacts our clients, in order to determine which of the available pairs of customers and reps is most likely to lead to a successful outcome. Afiniti is able to execute this process in under 200 milliseconds making us imperceptible to clients and customers alike.
After determining the most optimal pairing, Afiniti assigns a pair on this basis.
The outcomes of these calls are recorded, and at the end of each day, the ever-improving Afiniti algorithm is updated with our inputs in order to create increasingly successful interactions in the future.