Jeff EpsteinJeff EpsteinOctober 16, 2019
art-blank-cardboard-1111368-1280x851.jpg

10min2219

If you’ve ever visited a website and been greeted by a human-like pop-up asking “How may I help you?”, you’re not alone.

According to Comm100, nearly 50 percent of consumers already engage in automated conversations with chatbots. And, according to Gartner, these numbers are growing. 

Gartner predicts that 25 percent of customer service and support operations will integrate chatbot technology across customer service channels by 2020. The same source reports that in 2017 fewer than two percent did so, marking a huge jump in adoption of this technology in a relatively short amount of time.

With any business trend, organisations can feel pressure to adopt quickly, fearful that they will miss out on revenue and engagement opportunities if they do not use the same technologies as their competitors. However, simply deploying the latest technology does not guarantee companies will immediately begin delivering a great customer experience.

Organisations should be thoughtful in the way they strategically plan before implementing chatbots – AI-powered or not – to ensure that they are contributing to a positive customer experience, rather than just masking existing CX flaws.

How digital body language can guide when – and when not – to deploy a chatbot

A recent report by Juniper Research estimates that chatbots could help lower annual business costs by more than $8 billion by 2022. Chatbots also increase efficiency. By using AI-powered chatbots to process simple requests – account balances, due dates, etc. – agents have more time to have more personal, in-depth interactions with the customer via live chat.

These in-depth interactions also include successful sales conversions: the American Marketing Association reported that live chat increased sales by up to 20 percent.

With chatbots increasing efficiency and live chat boosting sales, bringing technology into customer interactions seems even more enticing than ever before. However, companies must consider how they will design their bot strategy so that it helps – rather than harms – Customer Experience.

I spoke with Tim de Paris, CTO at Decibel, a Digital Experience intelligence company based in Boston and London. He shared his thoughts on how chatbots can actually damage a customer’s experience if deployed ineffectively.

“To make a chatbot successful, organisations must have insight into how users are feeling about their experience,” he said.

“If a chatbot pops up asking the user if he/she needs help during an experience where they clearly don’t need help – like right when the page opens, or when he/she is already engaged in a positive experience – the chatbot interruption will only irritate the user, pushing them away rather than serving as a helpful assistant.”

According to de Paris, bots should be equipped to understand users’ digital body language: is the user engaged and ready to purchase? Or showing signs of confusion through scattered mouse behaviour? By being able to identify user pain points, brands can determine the best time to interject with a chatbot.

For example, if a customer is bouncing from page to page on a website and showing frantic mouse movements, clearly showing signs of frustration, the chatbot should step in to help, and even potentially pass the interaction off to a human agent who might be better positioned to help.

Conversely, if a customer has viewed one page for a significant amount of time and is flipping between shirt colours, he/she could be toward the end of the funnel, about to make a purchase and just contemplating last minute details. In this case, the chatbot should stay quiet, avoiding interrupting the customer’s decision.

“Only when organisations have insight into users’ digital body language with the right digital experience technology can chatbots be deployed at the most effective time,” said de Paris.

Best practices for implementing chatbots

Chatbots are immensely useful in boosting the efficiency of a company’s contact centre, but they are not a ‘one size fits all’ tool. Some bots can analyse text with natural language processing (NLP), whereas others only offer predetermined response options for users to interact with.

To successfully bring chatbots into the contact centre, companies should begin by being honest about the chatbot’s capabilities. By being up front about what a bot can and can’t do, customers will know right away what they can achieve in their interaction with a bot, and companies will understand when it’s time to transfer an interaction to a human agent.

Some chatbots are most helpful with basic questions – generating account balances, sharing business hours, etc. – but an agent should be brought in whenever the customer’s needs go beyond the bot’s capabilities. It is important that contact centres identify when a customer’s needs would be better serviced by a live agent based on a range of other criteria such as status, shopping cart value, geography, or the relative value of their query – every company will have a threshold above which they would prefer the question gets handled by a human.

Once a company deploys a chatbot, it should take advantage of all the metrics that the service provides to fine-tune the applications as needed to offer the best interaction. This includes feedback from post-chat surveys, recorded wait times, conversation lengths and customer satisfaction scores. This data can be used to identify trends as well as areas of strength and areas that need improvement.

While continuously using the metrics that the chatbot provides, businesses should be prepared to maintain the chatbots for best performance outcomes. Implementing a chatbot is not a ‘set-and-forget’ solution, but requires constant monitoring and improvement to best serve the agent and the customer, leading to a better interaction across the board.

After all, positive Customer Experience leads to more customer conversions.

Chatbots are here to stay, and if companies are going to use them, they need to know how to do so successfully and efficiently. By bringing a personal, customised experience to prospective buyers on digital channels, companies can improve the customer experience and increase revenue.


Jeff EpsteinJeff EpsteinMarch 18, 2019
blog-3591860_1280-1280x1280.png

7min1189

To create the best possible Customer Experience, organisations must think strategically about implementing the tools that will support their omnichannel strategy.

As adoption of live chat increases, simply implementing the technology won’t be enough to set a business apart in terms of the CX they offer.

According to Walker, today’s consumers expect consistent and user-friendly experiences from any brand they interact with, which is why CX is set to surpass price as a key differentiator for consumers by 2020. A business’s CX offerings are no longer just measured against their competitors, but against what consumers know is possible with CX. To meet these high expectations, organisations must be mindful of how to properly service their customers while making good use of their agents’ time. It will also be important to prioritise quality over quantity and ensure customers are serviced on the digital channels they prefer.

Customers are making the shift to mobile in waves and, as this occurs, businesses must ensure they provide seamless and consistent CX regardless of device or channel. The fourth annual live chat benchmark report highlighting the future of live chat and its impact on CX from Comm100, a global provider of omnichannel CX solutions, found that last year, chat queries sent from a mobile device increased to nearly 52 percent, representing an almost eight percent increase from 2017.

As customers continue to pivot their primary device usage away from desktop to mobile, mobile chat optimisation is becoming a critical strategy for all brands, but particularly for those in the consumer services and recreation industries.

Focusing on improving CX metrics alone may not be the best decision for brands. Brands that scored 90 percent or higher for customer satisfaction had an average wait time of 46 seconds, while customers that reported the lowest satisfaction ratings had an average wait time of 25 seconds. While many organisations strive for short wait times and quick conversations, these metrics do not necessarily indicate more efficient agents and increased customer satisfaction. It is easy to sacrifice the quality of the Customer Experience for efficiency, but organisations who emphasise quality of service over arbitrary targets will have an easier time meeting overall business goals.

The report indicates that companies and agents are close to achieving the right balance between speed and quality. On average, chat duration saw a decrease of four percent, with chats lasting an average of 11 minutes and 53 seconds. This continued the trend of shorter chat times, following the nearly 15 percent drop in 2017.

Just as with wait time, companies with a 90 percent or higher customer satisfaction rating had an average chat duration of 12 minutes and 26 seconds – 13 percent longer than organisations with lower satisfaction scores. Having meaningful, personalised experiences that address customer needs is more important than only attempting to lower metrics like wait time or chat duration.

For longer chat durations that take up your agents’ time, AI can step in to help balance out the workload. To ensure resources are used efficiently, organisations can route chats through AI-powered chatbots to offset chat volume and free up their agents for more complex queries. Chatbots with Natural Language Processing (NLP) and machine learning capabilities are now involved in over half of all chat interactions. They’re proven to be able to handle nearly 27 percent of those interactions without an agent, almost a seven percent increase from 2017.

Another tool that agents can rely on for optimising their workload is co-browsing. When an agent can view and interact with a customer’s web browser in real-time, it allows them to troubleshoot issues more efficiently, making co-browsing one of the quickest and most well-received ways for agents to solve customer problems. 

The report found that co-browsing sessions have an average satisfaction rating of 89 percent – six percent higher than the overall 2018 customer satisfaction rating of 83 percent. Customers may complain that canned messages are robotic or impersonal, but when used correctly it can help decrease an agent’s workload without sacrificing quality, which is why the use of canned messages has increased nearly 70 percent in one year.

The benchmark report’s findings indicate that consumers are readily embracing live chat, so long as the focus remains on improving their experience. To stand out from the competition and exceed customer expectations, brands need to focus on strategically implementing their omnichannel customer experience solutions in a way that prioritises personalised, consistent service without putting a strain on their resources.


Jeff EpsteinJeff EpsteinJanuary 7, 2019
chatbot-3589528_1280-1280x1106.jpg

6min1588

Forrester has predicted that this is the year “the backlash against chatbots will begin”, but you could say that it has already begun.

Headlines painting chatbots as a disappointment, and articles highlighting the funniest chatbot fails are easy to find, yet implementation isn’t slowing down. Despite early missteps, brands are eager to learn how to get it right because the upside is considerable: scalable, 24/7 customer service and a modernised, accessible experience.

So, how can companies capitalise on all that chatbots and AI have to offer while ensuring that they adhere to their core customer experience values? The key is to understand what’s keeping customers from having positive experiences with chatbots in the first place.

There are essentially two main obstacles to chatbot success: failing to define the scope of what a bot is expected to do, and striking a balance between bots and human agents. If companies want to succeed where others are failing, they must address these two key areas.

Defining scope and communicating functionality

Bots often fail because of poor design and a lack of defined scope. In order to have a successful, well-received chatbot, companies need to have a clear purpose in mind for a bot, understand what that role entails and then ensure the bot is equipped to execute.

Internally, companies must effectively plan and map their bot’s content and then make sure that externally, customers understand what the chatbot can do – especially if it has limited functionality. A bot should introduce itself and give customers a rundown of its functions so customers understand its capabilities. By giving users clear instructions on how to interact with a bot, companies can keep their customers on track and asking appropriate questions. 

If companies clearly communicate what a bot is capable of and equip it accordingly, escalation to a human agent is needed less frequently, leaving agents available to handle more complicated queries that require a human touch.

Striking the balance between bots and human agents

Even with a well-defined scope, bots can’t stand in for a human in every scenario, and inevitably a chatbot will be asked to do something beyond its capabilities. A primary reason why customers find chatbots to be barriers – rather than gateways – to easy communication is a lack of clear and effective escalation from chatbot to human agent.

We’ve all experienced an annoying phone menu that didn’t offer the option we were looking for and found ourselves repeating the phrase “speak with a human!” until at long last, we were connected with an operator who could help us in a way the automated system could not.

But solving this problem effectively isn’t as simple as including a “chat with an agent” option. While that might seem like an easy fix, it will likely result in too many chats escalating to agents and compromise the goal of automating a meaningful portion of frontline customer service.

The unpredictability of customer interactions makes it impossible to design a frustration-proof bot, but allowing customers to chat with an agent whenever they want isn’t the answer. Chatbot developers need to give brands the tools they need to add a layer of intelligence to the bot-agent relationship.

Bots need to know when to seek help from a human agent – for instance when a customer starts to get frustrated (detectable via sentiment analysis) or when a conversation starts to veer into uncharted territory (e.g. repeated attempts at asking the same question).

Once agents are notified, they can keep tabs on the conversation and take it over if necessary. This way, brands can have confidence that important conversations are routed to agents, and those that aren’t are being handled comfortably by the bot. 

Chatbots have a bad reputation due to poor design and lack of intelligent escalation, not because the technology doesn’t work. If companies get these aspects of chatbots right, they can prevent backlash against their bot. Scoping bots correctly and implementing an escalation process that provides a positive customer experience will help brands avoid frustrating customers while maximising the ROI on their chatbot investment – and those are outcomes that help everyone win.




Inform. Inspire. Include.
A free way to improve your business.

Customer Experience Magazine is the online magazine packed full of industry news, blogs, features, reports, case studies, video bites and international stories all focusing on customer experience.


CONTACT US

CALL US ANYTIME



Contact Information

For article submissions:
Editor
Paul Ainsworth
paul@cxm.co.uk

For general inquiries, advertising and partnership information:
alexandar@awardsinternational.com
Tel: 0207 1932 428

For Masterclass enquiries:
antonija@cxm.co.uk
Tel: 0207 1937 483

Awards International ltd
Acacia Farm, Lower Road,
Royston, Herts, SG8 0EE
Company number: 6707388

JOBS IN CUSTOMER SUPPORT

Find a job in customer support with Jobsora


Newsletter