Ten talking points on chatbots: design, deployment and use


At the recent SAS Forum in Paris, I was part of the team that demonstrated the contextual value of chatbots. Our discussions during preparation and questions after the demo were wide-ranging and extremely interesting. It would be hard to claim that there was a real thread running through it, perhaps because the use of chatbots is very much a developing area. I did, however, take away a number of points which may be helpful and interesting to others, even just as a round-up of the current state of play.

1. Chatbots are freely available, with many operating through Facebook Messenger

Facebook Messenger is rapidly increasing in popularity, and many organisations now provide services using Messenger via chatbots. Whether you want reminders, weather forecasts, share price updates, or sports news, there is a chatbot app that can provide it, and many more. All customers have to do is install the chatbot app into Facebook Messenger, and it will be up and running and providing a service. Many are pretty simple in their scope, but if that’s the service you want, then it is an efficient way to get it.

2. Chatbots have reached a level of development where users may not even notice them

For fairly straightforward interactions, it is entirely possible that users may not be aware that they are interacting with a chatbot, rather than a person. Indeed, these are likely by definition to be the ‘best’ uses of chatbots, because they can solve problems without needing any human input. This is extremely efficient for both customer and company, because it saves customers from waiting for responses from staff, and therefore improves the customer experience.

3. Chatbots are often the first ambassador of your brand

When you are talking with someone from a company, you would expect him to embrace and present the company values. When you talk with a chatbot, you should experience the same and really avoid a chatbot without personality. A brand having young people as a target needs a chatbot that uses expressions relevant to this group to make a real connection. It is quite difficult to provide a good chatbot who can simulate an “intelligent behavior” with a good brand personality. 

4. A modern chatbot is much more than just a conversational bot

Interaction with a chatbot is made by speaking and natural language processing is used to give the best possible reply. Most of the chatbot are only doing a research using the keywords of the previous sentences to make the reply but some of them are using more advanced analysis techniques. Artificial Intelligence brings a real game changer to the chatbot field and a modern chatbot will provide contextual answers based on a lot of information. However, it still easy to understand the underlying rules and impact the bot’s behaviour.

5. There is an ethical question about whether companies should disclose their use of chatbots

At what point do companies using chatbots have an obligation to disclose that use? This is a tough question, especially if some suggest that the ideal chatbot interaction is one that is not recognised as such. This may be a particular issue in banking, for example, where some interactions may involve personal issues as well as personal data, and trust and reputation are major selling points.

6. Counterintuitively, customers may be prepared to trust chatbots more than humans

While trust is important in banking, it may be possible to ‘sell’ chatbots to customers as more secure than talking to a human. After all, a chatbot is not going to make a note of your credit card details, and then run off to go shopping with them. Setting sensible constraints on the activities of chatbots, or allowing ‘guided AI’ may also be important steps on the way to acceptance.

7. Customer acceptance appears directly linked to usefulness

People are pragmatic. Research shows that most of us have no objection to robots, chatbots and AI being used as long as we are able to get accurate and rapid answers to our questions. In other words, we’re not, as a rule, bothered about the jobs of customer service representatives, how chatbots will change the world of work, or the ethics of interacting with robots, we just want the information.

8. There is little agreement about how to measure the effectiveness of chatbots

There is some consensus that chatbots are likely to improve efficiency, and save customer service representative time, but precisely how to measure this is a challenge. Some commentators suggest that chatbots should be measured against all the same metrics as their human equivalents, such as engagement and conversion rates. My own view is that perhaps measuring seamless interactions is the most helpful metric at present, as this remains a real challenge.

9. Humans need to be ready to step in when necessary

AI is not what you would describe as mature yet. Seamless interactions are still something to aim for, rather than a routine reality. Until that changes, humans need to monitor chatbots closely, and step in as soon as the conversation gets a bit circular, or the customer shows signs of frustration. Without that close attention, chatbots will not improve customer experience; quite the reverse, in fact.

10. Ultimately, chatbots are likely to improve customer experience—but we’re not quite there yet

We can all see that chatbots have the potential to speed up customer interactions, provide routine information in a reliable way, and reduce the pressure on human customer service staff. It is, however, clear that chatbots will not be replacing humans any time soon. The reality is more likely to be a developing partnership between AI and humans, enabling better service and improved customer experience.



About Author

Kevin Mazille

I work as the SAS Global Technical Lead for Accenture in France. This position give me the opportunity to spread and leverage our message around the SAS Platform in the Alliance ecosystem to see real effects and implementations. My curiosity keeps me busy with data visualization, programming (C, JAVA, Python, etc.), statistics, Hadoop, cloud (AWS and others) and BI. I love sharing what I learn and my previous role as the technical BI spearhead for France also allows me to be a seasoned speaker who enjoys transmitting both high-level and technical messages.

1 Comment

  1. Very well written - bots will be the primary interaction channel of the future and holds a lot of promise against current and traditional formats. The challenge is to make the interaction design more intuitive, demystify the tech / engineering side to bring it into more acceptance of the masses by displaying value, and increase utilization for business cases. The 24x7 self service and intelligent nature of the format is already being used for customer service, conversational commerce, health tech and education. We have started on the journey at Engati, do visit us to give us feedback on http://www.engati.com - you can also read our collection of blogs on http://blog.engati.com/

Leave A Reply

Back to Top