The four secrets of implementing bots

Bots are at a tipping point in customer care, and many large bands, including RBS and Vodafone, are already bringing AI and bots to life for millions of their customers. But the hype around bots has also introduced many misconceptions. If implemented poorly, bots could have a disastrous effect on the customer experience.

From our work with many of the world’s leading bot-makers and largest brands over the last few years, we have learned there are four ‘secrets’ to making bots successful in customer care, without which bot projects are more likely fail.

Secret #1: A bots is just another customer service agent

Customers want an efficient, productive, and satisfying user experience, regardless of whether they are interacting with a human or a bot. Interestingly, if a bot were just as accurate as a human customer care agent, a majority of global consumers (55 per cent) would prefer a bot over a human, according to a recent survey from LivePerson. Therefore, you need to hold bots to the same standards you would your human employees.

Consider where bots would fit in the org chart and how you’re going to track and measure their effectiveness. If the expectation for a good bot experience is lower than that of a human agent, then your investment in bots is really only creating frustrated customers. You wouldn’t be happy with an employee who was consistently passing work on to others, so it doesn’t make sense to turn a blind eye to bots if they're not serving their purpose either. If the bots you are using aren’t ‘smart’ enough to meet your standards, you should ‘fire’ them or further train them, just as you would your human employees.

Secret #2: Bots need to specialise

In the hype of AI and bots over the last few years, many bot-makers created generalist bots, which, in theory, were supposed to handle a variety of different jobs rather than a specific task. But they took on too much, and ended up doing many things poorly rather than one thing well.

A better approach for customer care is to implement bots with a specific purpose — such as welcoming a customer, closing out a conversation, assisting an agent, or specialising in FAQs.

·         Welcome bots help route enquires to the right agent or specialist bot.
·         Closer bots help wrap up conversations.
·         Agent assist bots help agents on the back end by queuing up potential responses and relevant information for an enquiry.
·         Specialist bots focus on answering frequently asked questions and standard business processes, such as changing an address or updating a credit card on file.

Bots that specialise in a specific task not only result in a much higher CSAT than generalist bots, but having an ‘area of expertise’ makes them easily measurable as well. In the same way you review the success of your human employees, you can evaluate a bot’s success against its responsibilities and decide which ones need to be improved or replaced.

Secret #3: Humans and bots should ‘tango’

Few things will frustrate a customer more than being in the middle of a messaging conversation and then being asked to call a 0800 number when the bot encounters a question that it cannot understand. To avoid a situation like this, it’s best to have a system in place in which bots and humans work together in a ‘tango,’ seamlessly passing the conversation between them. 

This tango enables human agents and bots to work hand in hand, delegating tasks to one another according to their specialities. As an example, a bot could help a customer pay their bill, but, when it comes to a question about the details of one of the charges, the bot can transfer the interaction to a human agent to appropriately handle that query. The human agent can then transfer the conversation back to a bot if the customer wants to change the address on their bill. 

It’s important to note here that customers value transparency and do not like feeling tricked. A majority of consumers want to be told if they are communicating with a bot. So, when these handoffs happen, it’s best to alert the customer.

Secret #4: Humans and bots should be managed side by side

For bots and humans to work seamlessly together, both need to exist on the same software platform, with the same metrics running for both, to judge effectiveness, escalate when needed, and so on. One of the biggest barriers that early bots have it is that they were running natively on a separate platform that didn’t allow for a human to intervene if the bot failed. 

Working in partnership means knowing when, where, and how to pass a consumer enquiry between bot and human agent without causing major disruptions. The only way to ensure this seamless transition is to manage all your agents under the same roof.

Investing in bots should be a decision you make after carefully considering the best interests of both your customers and agents. Deploying bots without a strategy will leave you with unhappy employees, frustrated customers, and rising costs — as ineffective bots create more customer service phone calls and agents pick up the pieces. This domino effect can be avoided by having a strong integration strategy that keeps these four secrets in mind.

Alan Ranger, EMEA Marketing Director, LivePerson
Image Credit: Praphan Jampala / Shutterstock