We are living in an age where chatbots and bots are in use in many areas of our lives. Bots have been around for a long while. I have experimented with Twitter bots for years, and as a gamer, I have played against bots as AI characters, and used them in messaging apps. Even as you read this, there is probably an army of botnets (groups of hijacked computers) bombarding your email provider with spam.
Most recently there have been some very public cases were bots were used to influence the political agenda. Here is a post from Wired that sums up the influence that bots can have over an extended period of time. In this second example, Brexit Twitter bots posted 65,000 messages then vanished after the EU referendum. The bots operated as a ‘supervised network of zombie agents’, according to researchers Dr Marco Bastos and Dr Dan Mercea from City, University of London.
Let’s focus on chatbots and discover the success factors that will make chatbot pilots and projects successful. Chatbots can boost digital customer service levels in B2B and B2C.
Chatbots are more like digital assistants. Many of us have used the popular chatbots like Cortana and Siri. Chatbots are being used in many different industries but there is a common thread when you look at what chatbots are actually helping us with. The travel industry is one example. Research from Kayak “Mobile Travel Report 2017” conducted by Opinium Research LLP in May 2017 found that 49% of respondents aged 18+ were using chatbots to answer customer service questions. In second place was searching for products while shopping online, then in third place doing research for booking flights, hotels etc.
Starting with Chatbots?
I am a firm believer in trying out new technology as an early adopter. I experimented with Intranets in 1997, Extranets in 1999, forums and eMarketing in B2B from around 2003, and social media from 2009.
Having a sponsor that has suggested a pilot and is willing to fund it is a very useful place to start. My advice is to have a low impact business case as a first pilot for chatbots. Having fun is good but I would strongly suggest aligning that first pilot to a business case that has a solid and well devised user story. I prefer a hybrid model, where chatbots are supplemented by human support. This makes it easier to rescue situations where the chatbot has reached it’s limit or it might be engineered to had off to a human in certain circumstances.
In my industry, I am tracking the early use of chatbots and bots to determine the most used scenarios. I am deep diving into the social media conversations world-wide to track early use cases and influencers in both B2B and B2C. I use the command center from Sprinklr to display those trends in real-time in our head office building in Eindhoven. Learn how to set up your own command center.
Testing your chatbot for normal line of business situations is important. Testing the chatbot for many more situations than you would possibly think of is super critical in my opinion. Most people try to probe a chatbot to see the extent it can go to with its responses. If the chatbot can deal with these probes and still maintain a dialogue to get the user back on track is very impressive. Trust me. Your bot will be tested! Ensure that you have logs to see how the bot is being used so that you can see how it is coping. A hybrid scenario is a good back up.
Here is one of my favorite examples of a chatbot being thoroughly tested. The Comedy Channel and the Domino’s Pizza chatbot discussion.
A Forrester Research paper by Amit Bhatia (October 18, 2017) suggested that it is important to “humanize your conversational bot” and also to “use bot personality and engage better to connect with humans”. I fully endorse this aspect. The research lists situations where brands forced customers to learn terms not previously known to the user or where users had to learn its unique language.
Going to the next level?
There are often industry pressures which force companies to start pilots with chatbots. A common need is to expand customer service levels; bedding in a more professional omni-channel approach to customer service. In situations like these there are higher customer expectations for real-time always on customer service. Chatbots in a hybrid model would help to expand multi-lingual global coverage to the next service level. For example, expanding human 18/7 coverage to 24/7 with chatbots initially picking up low traffic hours. Another use case might be languages that have usage figures below the level where it is feasible to employ a full-time person.
The attitude of millennials is very positive towards the use of chatbots (source; Retale, Feb 21, 2017) but the reception towards chatbots does very based on the country in question. A report by LivePerson “How consumers view bots in customer care” May 12, 2017, found that users aged 18+ in France were more positive about their chatbot experience than users based in Germany, UK, Australia, Japan or the USA, which was last in the list.
Panel debate about chatbots at DMEXCO
I was invited to be a part of a panel to debate chatbots at the recent DMEXCO event in Cologne. My fellow panelists were from Facebook, Microsoft and Valora. I represented my company, Philips lighting. Here are the details:
Watch the video here;
What will you do?
Bots have come a long way from since the days of Joseph Weizenbaum and his ‘bot’ called ELIZA in 1966. We have seen fictional bots in high-profile films like “Jarvis” from the Iron Man film. However, the bots that you and I are more likely to encounter are chatbots that help us get things done. These chatbots replace our efforts using many apps to just one request.
The question now is “what will you do if you are in a business that wants to expand its digital customer service levels in B2B or B2C?” I focused on this aspect in my contribution in the DMEXCO panel event about automated communications using chatbots. Now is the time to see what is happening in your industry. Are you ahead or behind the curve? What are your competitors doing? What business aspects could be automated? Define a pilot and KPIs to track the chatbot. Fully test it and check all the logs for improvement actions.
I will end with an example from my own company, Philips lighting. The Singapore sales office introduced a chatbot and piloted it and gathered some marvelous lessons and insights. Here is a review by LEDInside which publishes daily updates on the global LED industry.
What will you do with chatbots? Want more ideas? Have a look here..