When we buy a new product or subscribe to a service, we expect a great experience. As we should. That extends to customer support — if and when we need help, we expect to be able to find the information and answers we need. On our own and within an instant.

In fact, a recent Microsoft report found that two-thirds of consumers prefer self-service and solving issues on their own to other forms of support. And because of this, “self-service should be a cornerstone of any omni-channel (customer service) strategy.”

At Directly, we are fixated on helping companies (and their customers) increase the effectiveness of self-service support with our AI-powered CX automation platform.

With that fixation comes a fascination with failure. We are obsessed with understanding where self-service falls short. Through that understanding, we can identify opportunities and areas for improvement.

Recently, we co-authored research (“The Inner Circle Guide to AI, Chatbots & Machine Learning) with ContactBabel, an analyst firm for the contact center industry. One area we looked at was why people abandon self-service and seek other forms of support.

The top reasons web-self service fails

As part of the research, ContactBabel conducted a survey of call center agents to hear their perspective on the overall performance of self-service. The highlights (and lowlights) include:

31% of agents said that fewer than one-in-ten of their calls tried to answer questions via their self-service support channel before calling to speak with someone.


Takeaway: Self-service has fallen short of customer expectation for many years, with answers too hard to find or not robust enough to appeal to customers. Virtual agents and support automation gives enterprises a chance to re-imagine a much better self-service experience than traditional self-service options like online FAQs and support forums.  

66% of agents stated that the information and answers the customer called in about were not available online.


Takeaway: Companies face a difficult challenge in building and maintaining a robust knowledge base. Continuously delivering comprehensive answers at scale is a key part of Directly’s CX automation platform, unlocking the ability to automate many more support interactions over time.

62% of agents stated that they received calls about issues that could have been resolved online, but customers were unable or unwilling to do so.

Takeaway: Many support forums are littered with poorly organized answers and outdated information that can frustrate users. This is where machine learning and AI are offering tremendous value today, as they help surface the most accurate, helpful information in an instant.

85% of survey respondents agreed that web-to-phone escalation often involved a complex issue requiring a live agent to complete successfully. 82% of agents also felt that customers wanted the reassurance that a live agent brings to a conversation.

Takeaway: While support automation helps companies and users solve more problems quickly, having a clear, integrated escalation path from digital channels to an agent is still important.

How virtual agents can improve self-service support performance  

There’s little question that virtual agents and chatbots, as the front-end of self-service support, are poised for rapid growth. According to our research, web chat automation in support increased by 17% during 2018. And according to a recent report by Juniper Research, the retail segment alone can expect to reduce support costs in 2023 by up to $439 billion a year by using AI-powered chatbots.  

A well-trained AI-powered virtual agent increases the effectiveness of self-service and the speed that issues are resolved — especially as businesses scale and internally produced FAQs and help-desk materials can’t keep up. And fast resolution is ultimately what both businesses and their customers want.

Want to see for yourself how virtual agents can help improve your self-service support? Contact us to set up a Directly demo.

Download our report today

Want to learn more about how automation is being used in the contact center today? Make sure to download the full report here: “The Inner Circle Guide to AI, Chatbots & Machine Learning”