Edition 5
25th September 2019

In this edition of SaaS@Scale, we take you through how Freshworks is envisioning ML/AI reshaping the business of business software and how we are applying for the advances in these areas to our products. We’ll talk about our thinking of six layers of Artificial Intelligence and how that helps us in organising and prioritising product features. We’ll discuss how choosing and executing on ML projects differ greatly from normal software projects. Diving into the specifics, we’ll also touch upon various aspects of planning, building, and launching one specific project - Answerbot, which is an intelligent answering system for customer support. We’ll cover the ideas, techniques, and data pipelines that went into building this.

Freshworks platform for ML/AI in customer engagement
We will present our platform vision for AI/ML capabilities in the customer engagement space. Our AI/ML platform comprises six layers wherein each layer incrementally adds newer and more sophisticated capabilities on top of the layers beneath it. Such a structure enables us to offer a wide variety of AI/ML features to our business customers; which they use to optimize their end-customer interactions and achieve higher LTVs. We will talk about the technical aspects of this platform touching upon concepts such as semantic vector spaces, ranking & relevance models, forecasting and recommender systems.
Building an intelligent answering system for customer support
Ideas and techniques that go into building an intelligent answering system that learns from information available from a customer's account and attempts to answer queries from their users over chat, email, or phone. As users interact with the system more, it learns more; it clusters, triages and ranks these user interactions and presents in a tool for customers to curate questions and answers even further, building a feedback loop that continually increases coverage of queries answered.