What can I help you with today?

So, what can I help you with today? If you get a question like that from your spouse or partner, consider yourself very lucky. Let’s face it, though, unless you are just married or freshly in love, you no longer expect such questions from your loved ones.

Luckily for us, humans, robots are here. And they don’t mind asking us such polite questions over and over again. It is very likely you may have had a close encounter with one of them already as you surfed the web - a friendly little pop-up chat window we usually call a chatbot, talkbot or interactive agent.

However, it’s very likely your experience with the chatbot wasn’t quite what you expected. Your initial enthusiasm with “the little man behind the scenes” quickly faded away when you wanted something real from him. Maybe you got through the first two questions OK, but you quickly realised the limitations of the chatbot.

Why does the chatbot user experience suck?

I see two reasons: the first one is the technology behind the chatbot is widely accessible and for so-called command-based chatbots, it is also not expensive, which is why these bots are widely used. Keep in mind that command-based chatbots cannot create any new text, since they are fuelled by a databank of replies and heuristics. This means the bots reply by selecting an answer that matches the context of a query. They can only answer a limited set of questions and will require human assistance for any atypical questions.

The second reason for a bad user experience with bots, from both, the user and business perspective is the same as for any other technology: if a technology is used without a strategy or at least a defined use case for the business you are trying to solve or the benefit you are trying to realize, there is no expected benefit from the technology. This is true whether you are rolling out a CRM solution or implementing a chatbot into a web page.

The “other” type of bots

The second type of chatbots is much more interesting and it is here that technology advances every day. I am talking about AI-based or machine learning chatbots. These chatbots can answer ambiguous, indirect questions since they create replies from scratch using natural language processing. Because they learn from past questions and answers these chatbots become smarter over time.

When a user interacts using a device or messaging platform, their question is processed through natural language processing. If the bot can answer the question using his database/API, it does so, if not, it can “hand over” the query to a human. With machine learning, a bot can improve and train – the more messages it receives, the better it gets at responding. And this is the basis of machine learning – the machine should be able to learn and adapt through experience. 

You can start using a chatbot today

Recently, Microsoft invited me to do a demo of a chatbot integrated within Dynamics 365 CRM. The chatbot technology we use at Adacta is from our partner CafeX, the preferred Microsoft’s omnichannel provider. CafeX offers platform called Live Assist – real-time support for Dynamics 365 that provides a fully integrated experience for customers and agents that brings together a chatbot, live chat, co-browsing and video and voice conversations.

In my demo, I demonstrated how to create a database of questions and answers using Microsoft’s QnA Maker and then integrated a chatbot into a web page to assist users as they search for the right holiday. It also offers additional discount and provides additional information regarding the desired holiday destination. And when the process gets too complex for the bot, it can hand over the interaction to a human in the call centre who can easily continue the conversation with the customer since the entire process is tracked in the Dynamics 365 CRM. It was great to see this new technology in action.

By the way, one of the things you should definitely check out is QNAmaker, a tool for creating interactive FAQs for chatbots in just a few minutes. This approach is called supervised learning because humans help chatbots to create a database with fluid conversations – this is the technology we use at Adacta.

A chatbot could be a recipe for success

You really should consider using a chatbot in your business. The number one reason is to scale up your operations – while your agents can handle only 2 or 3 conversations at a time, a chatbot can operate without a limit. Another good reason is that a chatbot can reduce the burden of your first line of support. If your company receives a lot of inquiries, chatbots can lead off your customer support team. Also, a chatbot can assist customers, such as your web site visitors, and direct them to the right product, persuade a customer to finish the purchase when they hesitate or assist with additional offerings.

This is a fast-moving market. A recent analysis has shown that the market will grow by 27.8 percent annually until 2025. This means investments are already happening and enterprises of all sizes are investing into automating a part of their marketing and customer service efforts. Particularly the area of AI-powered bots is changing quickly, reflecting the rapid advance of technology.

The best strategy is to start evaluating chatbots now to see where it fits in your operations and how it could deliver an improved customer experience.