In today’s world, the majority of businesses have some sort of an e-shop or are e-commerce oriented (meaning: good portion of business can be done online). However, what if your Customer has an issue, and she/he needs to contact the Support, and it is outside of your working hours? Well – in that case, let us introduce you to Chatbots.

Chatbots (shortened from ‘chat robots’) are computer programs that can assist your business, and your customers, by ‘catching’ some of the key-words that Customers ask, and may provide the needed response to the questions, outside of your business hours. Sounds good, doesn’t it? You, perhaps, already have heard of some of the chatbots: Siri, Alexa, Google Home? Ow, yes – all of them are marvelous examples of it. So – chatbot doesn’t, necessarily, must be a software that will respond to your questions in written, but also – in voice.

These chatbots do enable your business, and expand your potential support, almost, to 24×7. Yes – ‘almost’, and not necessarily ‘always’ because, let’s face it – the AI applied here needs to be ‘trained’ – it needs a human input, before it is being put out ‘in the wild’ to help, and assist. Chatbots are usually trained to catch particular words, or phrases, and connect the dots between them, so they can assist as effectively as possible. Although not perfect, AI element here can offload significant volume of questions from your employees and, at the same time, ensure more extended support, even once the business hours are done with. Happy employees, and happy customers – sounds like a dream and, in a big portion of the cases, that dream has come true.

In this article we can focus on two, most commonly used, types of chatbots: the ones that do respond to the written queries in written and – the ones that respond via voice to the verbal questions.

Chatbots That Respond in Writing

First ones are, perhaps, the early adopters of AI where certain chat program can respond to the query ‘What kind of dishwashers you sell?’ with the list of the dishwashers available. Keywords here may be ‘sell’ and ‘dishwashers’ based off of which the chat program can give back the appropriate result, alongside with the adequate sentence that would make sense, and give potential Customer more ‘humane approach’, rather than confirming that she/he is talking to a software.

Chatbots That Respond Verbally

Second ones are becoming more and more adopted on an everyday basis – majority of us already has multiple voice devices across our homes (Google Home, Amazon Alexa, etc.), and we do use them to control the lights, cameras, doorbells, heating, and similar. These devices are even more capable of responding to slightly complex questions like: ‘Alexa, who won the Oscar?’, ‘Google, what is the forecast for today?’ These present questions that make sense. These present questions that one human being might ask another…except for the part where one human being is asking a….device? Yes – that is the world that live in today. These devices, although still only capable to understand relatively basic questions, are enabled to control portions of our home, access to it.

However – no need to be scared. If you’re a tech geek, and enthusiast, and you did your portion of the research, before ‘letting such beast in your home’ – you know what AI implemented there is, and is not, capable of. We are far, far away from the ‘Terminator’ scenario, as such intelligence, implemented in these devices is still dependent of internet-related database, and can only function when connected to that, oh so glorious, worldwide database. However – all of this gives us a very good insight of how much, most of us even, uses the chatbots, and today’s AI, and perhaps not even thinking of it.

Chatbots do make our lives easier. They make our lives easier as an individual, and as a potential customer. You can create your shopping lists, create to-do lists, get to the bedroom, and shout at the device to lock the doors, and turn off the lights, close the garage doors – handy, isn’t it?

Also, let’s focus our attention to frequently used term in chatbots universe: ‘canned responses’. These, usually, mean that, based off the certain question, chatbot will provide a predefined response (and again – that response will be based off the key words in your question). You might saw these in some of the companies who have business on Facebook also where, usually, when you open a chat, you have a suggested questions, such as: ‘What are your working hours?’, ‘What kind of services are you offering?’, and similar. Once you click on those suggestions, you get a response immediately, and it boils down to simple math: predefined question = predefined response.

However, you as a potential customer, did get the desired response, and no human was engaged (or harmed) in that process. Although this may be a rudimentary example, we need to look back to just a few years ago where these kinds of interactions, with no human involved in real-time weren’t even possible. Once business is done with the working hours – that’s it until tomorrow, or after the weekend. AI continues to evolve, and progress further and it does make our lives easier – may that be on example of chatbots that this article is focusing on, or in other areas of business: healthcare, financials, construction, and others.

Value of Chatbots in Business

As a starting point – measure the volume of chat requests that you are receiving, and your staffing/employee levels, and put it in a sheet, and compare the costs. Once you have the cost, or manpower needed, check on the costs of the solutions, payment models, and what you need to implement. Depending on your business needs it might be one variation of chatbot, or multiple – depending on your online presence, and platforms used (WhatsApp, Viber, Facebook, LinkedIn, etc.)

At the end of the day – you, as a business owner, customer, or an enthusiast know the best if this solution may be of interest to you. Do the needed research, ask around, or consult the experts in that field, and make as informed decision about it as possible.

Now, since we’re also talking about the business side of things, before doing any of that, of course, you need to make the finance plan and choose whether a few employees, or just chatbots, are better solution for you. Also, this depends on the type of support that you’re providing: if the questions tend to be unique, or specific, with lower volume – perhaps going with the full-time ‘human’ employees might be the better solution for you, and as your operation continues to evolve, you might introduce chatbots, and continue to train them.

Conclusion

Chatbots are one of many examples of AI usage in business and, as you could see from the article, they have its advantages, and disadvantages. Main ‘disadvantages’ here could be if this approach is right for your type of business, or some other solution will do better. Of course, if chatbots are better solution for your business – they do get better over time, as they are being trained, and learned based off of the data input which, in this case, are phrases/keywords, to which they can respond more accurately.

If this topic interests you, you might find something business related in our other article, also – Using ML in SLA.

Deep Learning for ProgrammersLearn how to use software development experience to become deep learning superstar!
  • Why should you care about deep learning?
  • Learn just enough math to be dangerous.
  • Get familiar with Python and TensorFlow.
  • Use familiar paradigms like Object Oriented Programming to understand main Deep Learning concepts.
  • Explore and implement 12 neural network architectures.
  • Solve various real-world problems with neural networks.
  • Learn how to generate images with neural networks.
GO!