How to get the most out of virtual agents

by Sean Baker
15 February 2017
Get the most out of virtual agents
With 89% of marketers believing their companies differentiate themselves primarily on the basis of customer experience (CX), it is becoming the key battleground for organisations of all types.

The customer journey is no longer a set of linear steps, but a series of micro-moments where organisations interact with their customers.

By now, we’re familiar with arguably the most famous chatbots: Apple's Siri, Amazon's Alexa and the new Google Home. Virtual agents, such as Virtual Agent Online, are like Siri for your brand, using your brand’s language. They’re always polite, always on and businesses are increasingly using them to fill the demand for around-the-clock support agents.

Virtual agents can live in many places – whether on your website, social media channels, or within mobile apps, among others – and encourage your users to ask a question. They are integrated into a business' databases so they have the ability to draw from static information such as FAQs, variable data such as pricing, and your customers' personal information, as well as richer media such as YouTube. The experience they provide, however, is far superior to simply serving up an FAQ based on a keyword or two.

Virtual agents interpret the intent of each customer’s question. They use natural-language processing to derive real value from questions. Take National Rail Enquiries (NRE), which is the source of all information on the National Rail network in the UK. Acknowledging a need for more agent support to help alleviate customer frustration, NRE rolled out virtual agent 'Lisa' in 2007.

Ask Lisa – National Rail Enquiries

A hastily typed query into the Lisa app, such as: “Late, want my money back”, will quickly return a result on claiming compensation for a delayed or cancelled service, for example. Customers can use these agents to find an answer to their query in their own time and using their own words rather than the organisation's language.

Although 67% of customers prefer self-service over getting in touch with a company representative, that self-service experience needs to be valuable and meet their needs. With this in mind, providing an easy way for customers to answer their own questions quickly is essential to growing buyer loyalty.

Here are some of the ways brands around the world are using virtual agents to provide the answers their customers are looking for and improve customer experience.

Take care of the little things

Many businesses implement virtual agents to reduce the load on human agents. For example, when one US financial services company noticed that many of its agents were answering repetitive, non-account questions, it opted for a virtual agent to help address these simple queries.

The company used transcripts of conversations between customers and agents to build the virtual agent, and included a prompt so that every virtual agent response directs the user to the relevant FAQ section of the site. Within a year of implementation, the company had reduced live chat volume by more than 80%.

Identifying these repetitive, straightforward questions gives your business an opportunity to deliver more consistent customer service. It also frees up your human agents to address more complex queries.

Integrate the human touch

Sometimes, a customer engaging with a virtual agent will need to be escalated to a human agent, particularly when the customer is planning to buy large or complex products or services.

An example of this integration is Verizon's virtual agent, which is integrated with live chat. This ensures that when trigger phrases are entered, such as “What is the price?”, the chat is handed off from the 'Ask Verizon' function to a live agent. Within 12 months of introducing the feature, Verizon significantly boosted sales while saving overall business costs involved with staffing the live support service.

Take the temperature

Improving customer service is one of the key motives behind deploying virtual agent technology, and it's a great way to capture real-time feedback.

Ask Sara – Commercial Bank of Dubai

Consider including a one-question survey at the end of each virtual agent response, like the Commercial Bank of Dubai’s (CBD) virtual agent, 'Sara'. Sara is a "smart, witty" chatbot with a distinct personality, and captures the sentiment of CBD’s customers by requesting feedback on each interaction. For your organisation, this could be as simple as asking “Was my answer helpful?” after each query.

This data can help identify pain points or content gaps in your knowledge centre. Not only that, but it allows you to scan the language customers use to improve chat responses and keep track of satisfaction in real-time.

Match the personality to the platform

Most businesses know that you need to alter the language you use depending on where your customer engages with you. The general rule: colloquial and less detailed on social platforms, and a more conservative personality with more detail on the main website. This rule also applies to virtual agents.

Ask Lisa – National Rail Enquiries

In NRE's case, the virtual agent 'Lisa' is on both its site and its Facebook page. When asked the same question, her response on each platform is tailored to the environment. This ensures Lisa’s answers meet customers' expectations, contributing to a more positive customer experience.

Find out more about Salmat’s Virtual Agent offering here or call us on 1300 725 628.


About the author
Sean Baker
Head of Sales – MicroSourcing

Sean Baker is a Philippines outsourcing specialist with extensive experience in creating operational excellence and building dynamic teams offshore. Today Sean represents MicroSourcing, a Salmat company, by helping companies in Australia and New Zealand set up offshore teams to reduce wage cost and onshore operational expenses in order to create competitive advantage. If you are interested in establishing your own team in the Philippines, Sean is the guy that will make your venture successful.

More articles by Sean Baker