robot toy hack
With so many communication channels already in play, how can a business decide if it needs a chatbot? A way to decide when and where you might want to deploy one. PHOTO: Rob Cruickshank

If you're in customer service or marketing, you hear a lot about chatbots and messaging, but you might ask yourself: Does my company need this? We have a website, a mobile app, a social presence, and phone and email support — do we really need another channel?

It comes down to understanding what a bot really is, and what the strengths and weaknesses of each channel are. There is no one channel to rule them all. Different types of enterprise content and modes of interaction require different avenues for engaging your customers. The trick is finding the best fit for the various scenarios. Let’s explore what those defining characteristics are for different channels and see if it helps answer the question of “Do I need a chatbot?”

The Power of Messaging Channels and Chatbots

A chatbot is a means of giving your customers answers to simple, frequently-asked questions in the fastest and easiest way possible. That alone should suffice as a reason to want one. A chatbot is essentially a digital employee that can answer simple customer service questions autonomously, but can also seamlessly hand a customer over to live staffers when necessary. And it does so on a ubiquitous and easy-to-use channel: messaging. Everybody knows how to text.

You might ask how much faster and easier messaging is compared to other channels you already offer. It helps to understand the characteristics of the various channels, of which the following are the most relevant today:

  • Email
  • Phone calls
  • Smart voice assistants
    • Devices such as Amazon Echo or Google Home, which are able to have live conversations or automated conversations (I am referring to voice-only systems here).
  • SMS
  • Stand-alone messaging
    • Messaging on chat channels that come with their own app/client, e.g. Facebook Messenger, WhatsApp, Twitter.
  • Embedded messaging
    • Messaging/chat experiences that are embedded in a corporate mobile app or website (e.g. web chat as a pop-up on a company’s home page).
  • Video
    • Face-to-face video chat, with or without screen-sharing.
  • Mobile Apps
  • Mobile Web
    • Web pages optimized for mobile devices.
  • Desktop Web
    • Traditional web content.

For each of these channels, we can distinguish between “live” and “automated” versions. Some of them aren’t available in certain modes. For example, there is no way to have a mobile app or website reliant on live staffers, and “video,” as  defined above, is a channel inherently based on live staff support.

How to Compare Customer Communication Channels

The characteristics you would want to compare the channels against are the following:

Characteristics Description Channel Impact
General Channel Characteristics
Bandwidth of information throughput How much information can you convey per time unit? A phone call is the lowest-bandwidth channel, due to the nature of verbal communication – you can only speak so fast before things get incomprehensible. Contrast that with a visual representation of information (“a picture says more than a thousand words”).
Richness of experience How “rich” does the experience need to be, on a scale between bare bones access to simple information, and a rich experience combining acoustic with visual representation, and allowing for various methods of data entry? Websites and mobile apps are the richest with their support for image, audio, video, complex forms, all embedded in large screen real estate. SMS, on the contrary, only allows 160 characters at a time as well as text only, and a phone call or Amazon Echo dialog typically only allow verbal conversations.
Time to answer How important is constant accessibility and availability of information? Is it acceptable to wait half an hour, or until the next day, or is an answer needed instantly? A contact center agent might only be reachable 10 hours on the phone or through social media during the day, whereas a website or chatbot is there 24/7/365.
Ease of use What environment does the customer find themselves in when needing help, and how easy is it then to access the channel of choice? Customers might have needs while at home, with access to a computer and thus the full web, or while driving, with a phone call being the only thing possible. While at home, you might have the need to show a technical support agent what you’re seeing, which would be a great use of a mobile app with its access to the smartphone’s camera.
Emotional neutrality How “emotionally neutral” is the channel in conveying information? Is it quite personal, like voice channels, or more pragmatic and neutral like a website? This matters. Certain topics might be of a nature that customers are embarrassed about, such as their inability to pay a bill, or medical conditions. Showing text on a website is the most neutral form of conveying information, but talking live over the phone can feel quite personal and close. Engaging in a dialog in the first place, be it automated or live, is more personal and emotional than reading data in a mobile app. Interacting with an AI about embarrassing topics might be more tolerable to some people than doing the same with a live person.
Customer Service Touchpoint Characteristics
Frequency of contact How often does your customer need to engage with you? If you're a bank, probably several times a week for balance checks, money transfers or bill pays. If you're in insurance, maybe twice a year to pay a bill, or once every few years for a claim. The more frequent the typical touchpoint, the more effort you should spend to make things easy for your customer. Giving them only a desktop-optimized website to check their balance, which requires access to a computer, logging in, navigating through menus etc., is a direct violation of the mandate to make things easy for the customer. Letting them send a simple text message and providing the answer within a second or two makes it easy. Make it possible to provide certain information that isn’t sensitive without any authentication besides verifying the number the request is coming from.
Clarity of need How “obvious” is the customer question? How well-formed is the question in the customer’s mind? There are times in the customer journey where a customer knows exactly what they need, e.g. the business hours or address of a store, or making a purchase of a product they decided to buy. At other times, they are just shopping around, browsing, with no clear idea of what they really need or want. In a phase of searching, browsing, shopping around, we prefer rich interfaces. Imagine finding new shoes merely by talking with a store clerk over the phone … you will want to see things, maybe even “try” them, you want to be “wowed” by the provider of the product. The richer the environment, the better that can be done. Shopping through an IVR or even chatbot with their limited bandwidth would not be a satisfying experience.
Complexity of need How many steps are required to accomplish the task? What level of “intelligence” needs to be applied (by both customer and business) to solve a need? This is different from the previous factor, as in some cases it might be very clear what is needed, but the inquiry isn’t simple, e.g. when needing to discuss a complex service issue. A channel such as chat, even when interacting with a human agent, can quickly get tiresome if the resolution of the issue requires a lot of back-and-forth dialog.

Where Do Chatbots Stand?

We can now assess how well both the live and automated flavors of a channel are suited for the characteristics listed above. Using a simple scale of 1 to 3, with 1 being poorly suited, 2 acceptably suited and 3 well-suited, we arrive at the following results when we add up the various channels’ ratings:

channel's suitability for chatbot evaluation

(Note that this assumes the same importance/weighting across all customer service touchpoint characteristics. You will want to adjust this for your particular industry.)

Based on these results, the following ranking emerges:

live or automated ranking

As you can see, stand-alone messaging emerges as an important channel, whether customer communication is automated or live. But if you are a believer in automation and customer self-service, then both chatbots and voicebots (the systems in use on the likes of Amazon Echo or Google Home) are the clear winners in this analysis.

What is interesting to see is that this analysis clearly shows the challenges of using interactive voice response (IVR) systems (or automated telephony). With its low bandwidth for information throughput, its general interaction slowness, the lack of any richness inherent in the channel due to its reliance on voice alone, IVR struggles to keep up with the other automated service channels.

For more details on this analysis, drop a line in the comments below or contact me on Twitter (@tpgoebel).

Give Customers a Choice

One of the biggest mistakes you can make when designing your customer experience is to assume a single-dimensional customer with never-changing needs or circumstances. Offering multiple channels to address changing environments is one way to address the many facets of people’s lives.

As you can see from this analysis, some channels are better suited than others when it comes to the different factors that determine the best experience. As stated before, no one channel rules all (which is why no channel will ever truly replace any other).

What’s worse, the general task that is undertaken in response to a need might be best done in Channel A the first time, but once customers have come to understand the business process, they might choose Channel B for all subsequent interactions. Consider booking a room in a hotel for a business trip: If you have never been to a particular hotel or area, you will want to research all your options before settling on a choice. Once you know what you want, the next time you need a room you might just contact the hotel directly and book a room via your channel of choice.

You can apply that to any shopping experience, where replacing something you have already used does not require going through the entire customer journey from scratch. For example, chatbots are the simplest, most convenient and most accessible way to reorder previously purchases items. Amazon knows this, and innovated in that area early, introducing 1-Click Ordering in 1997, when Amazon itself was three years old (and then obtaining a patent for 1-Click in 1999). More than 20 years later, 1-Click is still in use, but the technologies that support it have evolved. Now you just say “Alexa, re-order shampoo” to your Amazon Echo, and you’re just one verbal confirmation away from completing a purchase.

An Important Addition to the Channel Mix

Looking at customer service channels through the lens of different characteristics, and how your customer touchpoints differ in terms of complexity and frequency, you will realize how messaging is emerging as an important addition to the channel mix of today’s contact center. Add to that the general tendency to automate access to information, and you can conclude that no business that wants to stay competitive and relevant in 2018 and beyond can do without chatbots or voicebots.