Chatbots, static FAQs and automating low-value requests

Chatbots, static FAQs and automating low-value requests

I have discussed chatbots at length in my earlier posts (see Chatbots are coming your way: Start moving now!, Chatbots for wealth management, trade execution, and portfolio management? Yes, please!Why (chat)bots are so exciting?, and Chatbots: The future of customer service?). There is no lack of pretty good commentary and viewpoints when it comes to chatbots, and various chatbot platforms have spawned in Europe, US, and Asia-Pacific.

As I have argued in my earlier article around a week ago, Facebook has played a prominent role in the emergence of chatbots. There have been some news that Facebook has recently decided to scale down their initial plans in regards to chatbots as chatbots are not living up to initial expectations. Nonetheless, it has been remarkable that tens of thousands of chatbots have been created on the top of Messenger, and Facebook has allowed aspiring chatbot designers to experiment with a variety of ideas. Whole Foods, Pizza Hut, Domino’s Pizza, Wall Street Journal, TechCrunch, Trulia, TransferWise, Skyscanner, Burberry, and many other big corporates have already deployed chatbots as part of their (omni)channel strategy.

There is so much information available about chatbots, and various kinds of use cases have been discussed publicly so even the skeptics can now learn what chatbots are really about. It’s true that chatbots are not entirely mainstream yet, but there are so many interesting examples of their successful usage that this should encourage others to run more pilots and deploy them eventually into production. For example, Zendesk’s Answer Bot represents one very realistic way to take advantage of this new technology. [1] Answer Bot, although simplistic in its “seen” logic, is a beautiful illustration how to put self-service (that works) into action and addressing the root cause of ever-increasing (customer) service demand. The client sends a question, and the bot returns an answer, and if the issue cannot be resolved by Answer Bot, it will be escalated to a customer service agent. [2]

There are various problems with static FAQs, and we are now pretty used to get answers almost instantaneously. I have worked with front-line customer service myself, and there is no way to eliminate all the contacts per se. There isn’t a reason to remove everything, just the “dumb contacts” as Bill Price and David Jaffe point out in their great book. One of the biggest problems with static FAQs is that they are mostly populated with questions that are not contextualized, and as I noticed while working in customer service at Nordnet, FAQs actually interfere with the whole customer journey. Too often FAQs are not organically maintained but rather created out of thin air relying on the goods-dominant perspective, i.e. this is the product or service we offer, and these are the things that we assume are relevant to know.

Why then we still continue to rely on repeatedly failed tricks in self-service? “Our perspective is that they understand neither the need for self-service nor how to create self-service solutions that their customers will embrace,” Brice and Jaffe argue in their book. Answer Bot can’t produce content of its own so supervised learning, and content management is still critical. Thanks to the new forms of self-service technologies, Price and Jaffe argued already years ago, “new generations of organizations are shaping their whole market offering around the self-service capabilities that they provide, thereby disrupting the industries in which they compete.” Thanks to chatbots, clients and prospects can effectively self-serve themselves, and as technologies such as machine learning advance, knowledge bases will become more robust and can be intelligently combined to provide more contextualized answers.

Get used to chatbots

Chatbots have been proven to be an efficient and convenient form of self-service. Many companies, telcos, financial services and others, provide vast amounts of information via various online and offline channels to promote their service offering and educating their clients about their offerings. Yes, this is great as we want to be sure that we’ll get the best possible deal but usually, new issues and questions regularly arise during along the way. Paying a visit to the brick-and-mortar office, giving a phone call or sending out an email are not very convenient, and as we tend to value some things (like our family and leisure) more than being in a constant relationship with the service provider, interactions tend to be less of value than we though ex-ante. Advanced self-servicing, whether done with the help of chatbots, self-checkout or IVRs, is a significant trend that will eventually reach our homes too. We are so familiar with self-service that we don’t even seem to notice it when it’s done well, but we tend to take notice of the lack of service when self-service doesn’t work as we expect it to work.

Chatbots, on their own, can’t fix the service delivery systems but if they are build following consistent omnichannel customer service strategy, and seen as a way to address demand-side issues, they can be a vast improvement for customer service. In the end, while demand for human service is reduced, chatbots can be utilized to generate more valuable demand for service (e.g. upselling). Chatbots, equipped with machine learning, advanced analytics, and expert system capabilities, can actually be more like assistants rather than reactive typing boxes so that they can guide and direct the clients along their decision journey to give them a bit more personalized experience.

Yet another reason why businesses want to take advantage of chatbots is that there is need to prop up customer engagement and drive higher customer loyalty. Chatbots, which are in general relatively intuitive to use, can take many shapes and sizes, and if a client, for example, uses Facebook, providing a Messenger chatbot makes a lot of sense as people are already using Facebook (and Messenger), and there is an excellent chance that a person doesn’t need to actually change behavior as much as in the case of getting familiar with standalone chatbot somewhere out there.

A chatbot, deployed on a familiar (social networking) platform, is definitely something to think about as your clients are probably already using one of these platforms, and as we tend to mingle around these sites a lot, there is a less of a hustle to interact with a bot on well-known platforms that clients are familiar with rather than setting up a standalone solution (and we know both from research and experience that more active customers tend to be more satisfied). Chatbots, if done correctly, can attract new clients, build brand awareness, and most importantly, more people can be served with fewer resources. So chatbots are not just a neat way reaching out in a way that is not only convenient and cheap but possibly available 24/7/365. According to Ubisend’s 2016 report, 70% of all surveyed consumers prefer using email when making initial contact with a company and over half of the customers want prompt customer service every single day.

Figure 1. Emailing and messaging usage trends amongst three consumer cohorts.

Another reason any business that wants to increase customer engagement to use chatbots is because the platforms that offer them have a huge number of people already and there is a very chance that majority of your customers are already on Facebook and Twitter. Therefore, it only makes sense that you go to where your customers are and regularly frequent to better engage them and solve their issues. Through chatbots, brands can build and solidify their social media presence while reaching clients in a way that is not only accessible but also easy. It’s a win-win situation for everyone.

The use of chatbots can also save time and costs significantly. Instead of hiring support staff to take care of repetitive low-value questions that tend to make work boring and makes agents less productive, simple Q&A bots and more advanced chatbots can automate these tasks. The support staff can then channel his/her efforts into other areas of customer service that can improve the business.

Start automating now!

Kevin Sandlin brilliantly argues for following take on problems, “If it moves, automate it; If it keeps moving, crowdsource it; Once it stops moving, analyze it.” I would add that optimize, automate & outsource approach has been laid down in great detail in Ari Meisel and Nick Sonnenberg‘s book, Idea to Execution: How to Optimize, Automate, and Outsource Everything in Your Business. As Bill Price and David Jaffe write in their book,

[Being proactive] involves notifying customers proactively whenever contact-creating “events” are known, especially repetitive and critical situations. Being proactive produces other important value for companies, such as “Wow, you remembered” and “They’re taking care of me, so why should I take my business elsewhere?” both solid marketing and top-line benefits.

Allen Weiner argues that “chatbots are aimed at impatient millennials who are always on the hunt for instant gratification.” Because chatbots provide instant gratification by providing a potential to resolve various issues customers face almost immediately, live chats is not a viable option in the long run. Something everyone has experienced that situation where you ask something or give feedback without receiving an answer for days. As a former customer service agent, I regularly faced a difficult trade-off between customer inquiries, i.e. the scope and content of the questions varied unpredictably so that you could not focus on resolving more complex issues first, but you were forced to move from simple questions to complex inquiries (and this, of course, is also related to the principles of query management). In the end, it didn’t really matter how fast you were able to type, pick up a phone call or trying to locate relevant information, there was so much variation that it was impossible to predict the workflow. Also, issues that couldn’t be resolved for the first time caused additional problems in the workflow management.

The most simple way to acknowledge that the matter is being managed is to create a client-facing interface from which client can track down and follow through various queries that have been made (coupled with approximate time to resolution, updates on progress for the request, etc.). In fact, being more transparent and proactive reduces the need for clients to contact customer service multiple times in a row. Sending or displaying a notification that a specific request for service has been received, and proactively keeping clients informed about the completion of issues and requests, is not just respectful towards customers but it also improves customer experience. Such simple things do not even require a chatbot, but a chatbot can answer (simple) requests, track down particular issues and provide suitable solutions to the clients.


In conclusion, agile, proactive and intelligent organizations will always find ways to take advantage of emerging technologies to develop innovative and customer-centric solutions especially when it comes to serving, engaging and connecting with their clients, and providing them the best service possible. As Ubisend’s recent report reveals, almost half of the consumers surveyed said “they would rather use a messaging application to communicate with a business than a phone call” and “over 60% of consumers believe businesses should be available, contactable and responsive through messaging applications”.

Chatbots, as I have argued in this article and my earlier writings, is just one piece of the puzzle, and they must be connected with the overall operating model of the organization to improve the quality of the service. Chatbots are just one of the latest technology-driven trends in the transformation of customer service, and therefore the companies that want to stay ahead of their competitors need to grasp the opportunities that various new technologies provide.

[1] There are other similar solutions available too, e.g. AnswerDashQnA Maker, Indico, and Driftbot. It’s fairly straightforward to set up a Q&A bot (on the top of Slack).

[2] See Amir Shevat’s Designing Bots: Creating Conversational Experiences for more information.

Photo creditSimon Blackley via / CC BY-ND

Leave a Reply

Your email address will not be published. Required fields are marked *