It’s About the Destination, Not the Journey

Mar 24, 2021 | Autonomous Insights

By Prosenjit Sen

Co-Founder and CEO,

“Autonomous” is a term applied most famously to vehicles. It is generally if colloquially, understood. As AI-infused self-driving cars become more and more independent of manual operation, they become more and more autonomous.

Much the same is true of the very different domain of Customer Support, particularly for mission-critical technology and industrial products. But while Autonomous Vehicles and Autonomous Customer Support have a number of AI-related commonalities, there is one critical difference. For vehicles, “autonomous” pertains primarily to the journey. Customer Support, in contrast, is all about the destination – i.e., the right resolution to a specified problem.

The migration toward autonomy in vehicles is well-known. First, our car simply beeped a warning if we were about to back into a telephone pole. Later, it semi-autonomously demonstrated that it could park itself. Now, it can drive itself for miles on a busy highway. At some point, perhaps sooner than we might think, manual functions will be all but superfluous — and the driver will become essentially a passenger.

The roadmap so far appears to be somewhat similar in customer support.

Today, we are at a standing start, as there is no autonomous customer support in widespread use.

This is a particular problem for technology and industrial products, which are known for their complexity, high velocity of releases, and fragmented but frequently updated reference data. When the toolbox is limited to search, the right resolution is like a needle in a haystack.

For all intents and purposes, customer support today is performed manually. When a complex issue arises, a customer typically first tries self-service search in the customer support portal. The response usually consists of a list of documents, some of which are not relevant, that the customer needs to read to find the resolution. At this point, according to a recent Gartner survey, nearly nine out of ten customers abandon the self-service functionality and create a support case with the company’s Case Management System.

The case is routed to a support engineer, whose first task is to interpret the complex issue.  Then, armed only with a common search tool, the support engineer searches read and sifts through the same voluminous documents that were digitally dumped on the customer shortly beforehand. Hours later, the support engineer provides a resolution – and hopes that it actually resolves the issue.

This largely manual process is inefficient, time-consuming, expensive, and frustrating – often resulting in low CSAT. And as the company adds customers — and customer inquiries – the manual process performs worse. And the low CSAT plummets even further. is currently helping corporate customers navigate the road to autonomous customer support. Early results are extremely positive, eliminating hours of document reviews, enabling significantly lower costs and higher CSAT, and improving customer-support productivity by more than 40%.

The roadmap consists of four steps:

1. Reactive Autonomous. This first phase of autonomy combines Deep Learning, NLU, and Computer Vision to interpret the customer issue accurately and deliver the resolution from the internal documents promptly.

When using Self Service, the customer goes to the support portal, describes the issue, and receives resolutions directly.  When a Support Engineer is involved, he or she opens a support case (in Salesforce, ServiceNow, or others) and receives the resolutions from the reference documents automatically. What would have taken hours of reading now takes only a few minutes. 

2. Proactive Autonomous. This next step essentially enables auto-response. When a customer submits a case, interprets the inquiry and generates pertinent resolutions. If the resolutions have a high confidence score they are sent automatically to the customer, a process called Auto-Response. If even 10% of the support cases are resolved with Auto-Response, it represents substantial savings to the company. is developing the third and fourth steps, which will be detailed in subsequent writings. The third step is End-to-End Autonomous, which includes a more comprehensive case-preparation functionality. The fourth step is Preemptive Autonomous, which aggregates and analyzes symptoms preemptively to anticipate the problem and come up with a resolution.

We will explore End-to-End Autonomous and Preemptive Autonomous more in forthcoming writings. While it is an eventful journey, the destination is on the horizon.

This is one of a series of periodic observations of the customer-support sector by Prosenjit Sen, Co-Founder and Chief Executive Officer of

About is the technology vanguard provider of Autonomous Customer Support.’s multi-channel platform combines Deep Learning, NLP, and Computer Vision to interpret complex customer cases and automatically provide resolutions at scale with unsurpassed accuracy and speed. The result is unrivaled efficiency and scalability in customer support, with lower escalations, higher CSAT, and significant cost savings.   More information may be found at