Chatbots are hot property among internet marketing agencies and web design firms. You’ll probably be familiar with them as the latest generation of voice-operated virtual assistants like Siri, Alexa and Cortana. However, chatbots also come in text-based forms, and these non-verbal tools are slowly changing our web browsing habits.
A chatbot is essentially a conversational robot which scans human input strings and attempts to find a suitable response from a populated database of replies. As interaction levels increase, the algorithms powering this system become more intuitive. A human is generally on standby for complex input strings with no obvious response, but bots may run entirely autonomously.
That makes chatbots ideal for websites, appearing as a discreet pop-up in the corner of your browser. It might invite people landing on a restaurant’s website to make a reservation, or encourage someone browsing a health portal to ask a medical question. Often depicted as an avatar and a made-up name, these floating text boxes are becoming increasingly commonplace. But could your company benefit from using database-driven communications?
The advantages of chatbot tools
One of our biggest pain points as consumers is a lack of free time. Chatbots can instantly simplify basic processes like scheduling an appointment, as opposed to being put on hold in a call centre queue or sending a web form into the ether. Quick responses at any time of day or night suggest efficiency, implying a commitment to streamlining the user experience.
If your business revolves around appointments or common enquiries, automating this can free up customer-facing staff for other duties. Imagine a front-of-house employee in a restaurant attempting to seat diners while having to deal with booking enquiries. Automating the latter via a chatbot would enable the staff member to direct their full attention to paying customers. This improves productivity, and also contributes to a better client experience.
Chatbots may even help to reduce support staff levels, freeing up income for reinvestment elsewhere. Automation has improved business efficiency levels ever since the Industrial Revolution, and some analysts suggest that bots represent an inevitable efficiency as they’re endlessly scalable at very little additional cost.
Chatbot tools are readily available with dozens of WordPress plugins. These can handle everything from product enquiry resolutions to interactive lead generation, becoming an integral part of your website. You don’t need to understand how artificial intelligence works to offer this to customers – equally they don’t need any specific understanding in order to interact with it. Indeed, many site users will assume they’re talking to a real person throughout any correspondence.
The drawbacks of chatbot tools
One of the biggest drawbacks of chatbot tools is the risk of data siloes. Without a robustly-designed method of recording interactions, communications could be lost or misplaced. And few people will accept “the bot didn’t work” as an excuse for being let down. They’ll simply migrate to competitor companies, badmouthing your firm along the way for its poor customer service – the very opposite of a bot’s supposed value.
Until machine learning software becomes more sophisticated, chatbots are also going to be limited in their scope and inability to improvise. Slang or mistyped comments typically trigger a response along the lines of “I’m sorry, I didn’t understand that”, which creates a poor impression and might lead to site abandonment. Bots remain fairly primitive, and a lack of tracking tools mean companies may not realise their chatbots are losing them custom.
This is also a problem for firms where a large number of data strings may be input. A chatbot would struggle to book a multi-stage plane journey, for instance – the sheer complexity of options would baffle a conventional cell-structured database. There might also be regulatory issues in certain industries, where judgement is needed to determine best practice.
Finally, chatbots are inherently unemotional. Their lack of empathy beyond pre-programmed platitudes is wholly inappropriate for companies specialising in counselling or medical issues, and they tend to perform poorly when complaints are received. Natural language processing technology has a long way to go before a computer can respond to emotional cues, and it may be advisable to avoid any automated responses to criticism or complaints.