Chatbots are one of the hottest areas in Internet marketing right now. These are software components that know how to keep a seemingly intelligent conversation with your leads and customers, for example by offering 24/7 autonomous customer service.
Machine leaning, natural language processing (NLP), natural language understanding (NLU), and natural language generation (NLG) are used to understand and drive the correspondence forward, sometimes in combination with sentiment analysis to understand the emotional mood of the user.
A chatbot is usually found in a chat window on your website, in social media platforms like Facebook or Twitter, or in messaging platforms like Apple iMessage or Slack. Chatbots are often used for customer service and can help by answering basic questions automatically.
Depending on the industry, a chatbot may be able to answer 50-80% of incoming customer service enquiries automatically. What’s more, they are on duty around the clock and never need a coffee break.
The chatbot can either be statically programmed to know predetermined answers to common questions, like “when does the restaurant close on Thursdays?” or “do you have free parking?”, or have dynamic database lookup capabilities, for example checking and responding with the price of a car insurance for a vehicle with a specific license plate number, or the price of a hotel room on a particular date.
Some chatbots can hand over the correspondence to a human agent, should it detect it is unable to cope with the situation. This may be the best of both worlds, thus providing automation of simple customer service queries at scale, combined with human care in case of difficult matters.
Chatbots don’t just have to be customer service agents, though.
Growthbot is an interesting example. This marketing and sales agent integrates into Slack, Facebook Messenger, and Twitter. It was created by Dharmesh Shah, the legendary co-founder and CTO of HubSpot, and it can give you quick, easy access to information and services related to growing a business, sales, marketing, and more.
Another great example is eBay’s chatbot (or shopbot, as they call it) inside Facebook Messenger. They position it as a virtual personal shopper, and you can have natural conversation with it in plain English.
You can say things like “show me red jogging shoes under $100” and it will ask about shoe size and whether it is for a man or woman, and other relevant questions. Once the shopbot has filtered down the query sufficiently, it presents a number of product recommendations, each with a photo, product description, price, and perhaps other information.
If you like one of them, only a touch is required to add it to the shopping cart and purchase it. If you don’t, you can say things like “show more models,” “similar but in blue,” “the same but in size 42” and more—all within the familiar environment of Facebook Messenger.
It also remembers your shopping habits, like shoe size and gender, and improves its product recommendations over time based on your preferences. This gives it an edge on simply searching an online store using keywords by automating the interaction and learning over time how to better help you.
One of the coolest features of eBay’s shopping bot is the ability to upload a photo containing something you would like to buy. Say you pass a shop and see something you like, or want to get the same T-shirt as a friend.
Using the eBay shopbot, you can take a photo of the desired object, upload it, and the shopbot will detect it using image recognition technology. It then presents products that look the same and can be bought on eBay.
Similarly, Google now has Lens, which can analyze a picture taken with your phone camera and present information on the item, like the price and where to purchase it.
This technology is incredible, and many predict chatbots or shopbots integrated into Facebook Messenger, Apple iMessage, and other platforms will replace traditional web shops as the preferred mechanism for online shopping and e-commerce in the next few years.
This paradigm shift may be boosted by the emergence of natural voice control with full conversational capabilities. If you want to make purchases online using voice, you will need a chatbot to do so.
In fact, voice-based assistants like Apple Siri, Amazon Alexa, Microsoft Cortana, and Google Assistant are becoming increasingly popular, propelled in part by smart speakers like Apple HomePod, Amazon Echo, and Google Home.
These are continuously listening to your requests in your home. It isn’t hard to imagine the equivalent of an eBay shopbot integrated into these voice assistants. I predict these voice-based assistants will be integrated into many other products outside of smart speakers in the future, like refrigerators, TVs, cars, and more.
You can then simply buy new detergent—or virtually any other product—using a short voice conversation.
With Google Duplex, this is taken one step further.
Or rather, the issue is reversed.
Instead of having a voice-based chatbot listening to requests for services, Google Duplex is a voice assistant that can initiate phone calls and have automated, natural voice conversations with people on the other end.
When presenting Duplex at Google I/O in May 2018, a demo showed how the software made phone calls to hair salons and restaurants, negotiating an appointment for a haircut and a restaurant reservation, respectively.
With voice-based chatbots initiating outbound phone calls, there may soon be no need for call centers. A voice-based chatbot can easily dial and manage thousands of phone conversations in parallel, removing vast numbers of call center jobs in the process.
It also allows for improved service and cost reductions if the chatbot speaks many languages, thus removing the need for translators or multi-lingual staff. The possibilities are endless.
For example, automatic calls informing customers about changed delivery schedules or flight delays, calling parents informing that the teacher is sick and the kids have to stay at home from school, or unmanned phone surveys—all with the nuance and personalization of a real human.
In fact, voice interfaces may disrupt the online industry by making traditional web pages far less important.
Voice interfaces may disrupt the online industry by making traditional web pages far less important.
If people interact with internet services verbally, they will be less likely to turn to traditional websites to search for information.
If you’re sitting on your couch and want to find the closest pizza restaurant, it’s far easier to speak the command to your voice-activated home assistant device than to bring up a search on your phone or laptop.
That would disrupt the search engine optimization (SEO) industry, to start with.
If voice assistants rather than web pages become the main interface to the internet in the future, then ad-financed companies like Google and Facebook may be in for a massive reduction in revenue.
As an example, how would Google monetize search ads (Google Adwords) from a voice interface? I don’t think the users would accept ten ads being read aloud before the answer to a search is presented.
The future truly is here. In fact, you can already add speech recognition to your own products. Amazon pioneered the option to inject your own services into Amazon Alexa. Using a feature called Skills, any company can add voice commands to Amazon’s system. The Google counterparts are called Actions.
An interesting example of this is the Swedish company LogTrade, which develops cloud-based software for digital logistics (ordering and management of freight services).
LogTrade developed their own Alexa Skills and Google Actions, meaning that their customers can now order shipments or query for freight status using simple voice commands.
This provides unprecedented convenience in many situations, particularly if you can’t use your hands to type on a keyboard for whatever reason. You might be driving a fork lift, for example, or wearing protective gear.
In June 2017, Apple announced a similar “hook-in” capability for its iMessage text service called “Business chat” or “Chatbots for iMessage”.
Chatbots built using this technology can integrate with the rest of iOS, and schedule meetings in the calendar or initiate payments using Apple Pay. Using this tool, airlines could allow ticket purchase and seat selection, and hotels could allow you to book rooms, all from within Apple iMessage.
How do you develop a chatbot or shopbot for your business and have it integrated into popular platforms like Facebook Messenger, Twitter, Slack, Alexa, or Siri?
The companies offering chatbot integration with their products have opened up this capability using APIs (application programming interfaces), enabling software developers to hook in their own functionality.
The easiest solution is to use a chatbot development platform, which is effectively a high-level chatbot authoring tool that can be used to develop these systems with a minimum of programming expertise.
It took me just a few hours to develop my first chatbot. I used Motion.AI (now acquired by HubSpot) and designed the conversational flow by building a state-chart using graphical drag and drop tools.
Most other chatbot development tools are similarly easy to use. They then transparently integrate with the chat platforms (Facebook Messenger, Apple iMessage, Slack, etc.), completely removing the need for you to understand how to do this manually.
Many of these chatbot authoring tools offer a “design-once, deploy on many platforms” approach, further increasing their appeal. This means your chatbot only has to be developed one time, but can be deployed more or less automatically in many different chat platforms.
Examples of such solutions include Chatfuel, Flow XO, ChattyPeople, ManyChat, Octane.ai, and many more. IBM Watson, Microsoft Bot Framework, and Microsoft Language Understanding Intelligent Service (LUIS) are other solutions from major players in the industry. Google DialogFlow is another example.
In fact, there are now so many options available in this space, that I can’t list them all. A Google search on “chatbot development tools” should give you a place to start.
One of the reasons there are so many available is the importance many think chatbots will have in the years to come. You may want to monitor this, or perhaps get your own chatbot developed right away.
I have to say it was immensely rewarding to talk with a virtual copy of myself the first time I developed my own chatbot.