Behind the Scenes: AI at Freshworks

We have a lot of exciting things rolling off people’s sleeves at Freshworks all the time. But there is a lot more work that goes into bringing out all that awesomeness that everyone sees. We thought we would give you a glimpse of what happens behind the curtains. We are introducing ‘Behind the Scenes’, a fortnightly series that talks about all the effort that is put into making Freshworks work for you.

It’s an open secret that AI is making inroads into every industry in the quest to offer simple solutions to complicated problems. But the main question for us is how relevant it is to our business

AI will be the next frontier for customer engagement – Swaminathan

It seemed like an obvious choice to approach Swaminathan, our Director of Data Science, to have a chat about AI at Freshworks, with his 13+ years of experience in Data Sciences and Machine Learning. He focuses on the commercial applications of AI, i.e building technology to enable businesses to derive value. Currently, he is building a Machine Learning team that is focussed on enabling AI based automations and smart features within the Freshworks customer engagement product suite.

AI is currently taking over most of the market by helping companies generate greater value. How is this happening?

A company can leverage AI to target the right audience and market their products. Further downstream, they can use it to optimize the sales process; i.e. to identify which customers have higher propensities to convert and engage with the product.  Once the product is sold, AI is once again used to derive a customer persona, which is a representation of the customer based on their background profile and in-product behavior. This persona is built over a period of time and can be used to offer customized in-product experiences, discover customer pain-points and adoption patterns – all of which can help make the product better.

This would increase customer loyalty and open up opportunities to cross-sell products to the existing customer base and also reduce churn.  Thus, AI can help unlock significant value all along the customer’s life cycle and enhance customer lifetime values (LTV) over a period of time.

Let’s take Freshworks for instance, how are we aligning ourselves to AI and bots?

With the recent Freshworks 360 launch, we have re-imagined traditional customer engagement software. It aims to bridge the gap between different customer engagement functions within a company and weave a seamlessly integrated experience around customer support, sales, and marketing. As part of this, we plan to introduce AI based automations and workforce optimization solutions across our customer engagement product suite.

The list includes intelligent chatbots that can intercept and answer customer queries, automated email answering systems which can resolve issues without requiring human intervention, noise cancellation systems that can help a business focus on what’s important, smart routing systems to help identify the right agent to work on a ticket, prospect scoring systems that help a sales team identify and prioritize the right opportunities at hand, sentiment analysis on text and voice to help assess the quality of customer interactions, intent detection systems and so on.

When we think of AI in recent times, we can’t help but wonder how it has risen to such an extent.

This revolution can be attributed to the rise of deep learning. This technology has fundamentally changed the way we approach problems in areas such as vision, speech and text processing.  Deep learning techniques are able to outperform traditional machine learning algorithms when some pre-conditions are met, e.g. when they have access to plenty of training data or when large amounts of training data can be synthesized by running simulations. The latter case applies to well-defined and closed problems, e.g. chess, where simulations can generate valid games for a model to learn from and get better on an continuous basis.

At Freshworks, we have now started investing in Deep learning tech for a variety of NLP applications such as vertical-specific text embeddings, recommender systems and context-aware question answering systems.

Our objective is to combine the best of both worlds, i.e. we are trying to make traditional ML techniques co-exist with Deep Learning in order to leverage their strengths appropriately.

Using AI, we can build highly reactive systems that can instantly respond to customer requests.  A context-aware AI powered bot can instantly sift through thousands of technical manuals and knowledge base articles, model the user’s association with the brand, interpret his/her query using NLP and provide an answer within seconds.  Thus, customer interaction becomes instantaneous; the low turnaround times would delight the customer. There are multiple downstream benefits to this as well. For example, faster resolution of issues would result in a better perception of the brand’s ability to handle future issues, thus reducing customer churn.

However, there are risks associated with prematurely exposing users to AI based services. Today, in customer support, direct human intervention (by a support agent) is preferred by the customer.  The reason for this is that the current crop of AI based bots are not quite upto the mark. More work needs to happen before AI becomes useful in customer engagement.  AI should never be unconditionally imposed on the user. It should be made to intervene only when there’s confidence that it can add value, otherwise it should fallback to a manual workflow.  For instance, in the customer support context, if a user’s query is unclear or too complex to handle, AI should transfer control to a human agent for resolution.

Over the last year or so, we have started making AI based automations and enhancements available to some of our customers. They have been able to derive tangible business value by adopting these solutions. We strongly believe that AI will continue to evolve over the next few years and gain even more significance, to a point wherein businesses would start opting for (customer) engagement software containing better AI.

AI would emerge as a determinant factor in the (business) decision on which customer engagement software to use.  At Freshworks, we have a head start and we are very much excited about the journey ahead.