2014 witnessed the turning point when chat applications became more widely used than social media platforms. Indeed, building interactive chatbots have become increasingly important for companies that are always on the lookout for new growth opportunities. In marketing language, this means that chatbots are playing an increasing part of conversion rate optimisation.
- What are chatbots?
- How chatbots work
- Building your own chatbot
- How chatbots benefit businesses
- Are chatbots replacing human jobs?
So, what are chatbots?
Put simply; they are computer programs that allow the automation of specific tasks through having a conversation with people through an interface such as Facebook Messenger. The technology that is supporting chatbots is evolving at a fast pace and is driven by AI (artificial intelligence). AI allows more advanced chatbots to use user requests for continually improving their knowledge base, leading to increasingly more intelligent and human-like responses.
Bots have been repeatedly described as the future since they have come to lead the way in some areas of digital marketing, and swiftly making their way to being highly sought after digital assistants for businesses of all sizes — not just companies. Many of us already have our own personal digital assistant – ever had a conversation with Siri or Alexa?
Why have chatbots become so popular?
One prerequisite for the rise in the popularity of chatbots is the high level of sophistication reached by AI, whereby bots can engage in increasingly human-like conversations. Back in 2011, Gartner predicted that over 85% of all B2C interactions would be handled with no human-to-human contact by 2020. More recent industry research has supported this estimation, forecasting that by 2022, chatbots will be able to deal with 90% of all customer inquiries, cutting costs for businesses by circa $8 billion.
Undeniably, there are plenty of reasons why businesses and advertisers should adapt to the use of chatbots, perhaps the biggest one being that messaging apps such as WhatsApp, Facebook Messenger, Slack and many others continue experiencing a significant leap in the number of users. With this growth comes an increasing demand for services that can respond to questions and solve user problems through simple, automated conversations. According to Statista, the number of active advertisers on Facebook passed the 7 million mark in Q1 2019 but so far, only slightly more than 300,000 Messenger bots in use, suggesting that considerable opportunities to stand out are still untapped by businesses.
How do chatbots work?
Chatbots are based on artificial intelligence, and they communicate with users through distribution channels such as email clients, SMS services and messaging apps, using both speech and text. They are typically built by third-party companies that allow businesses to position their chatbot across a given platform.
Currently, we can divide chatbots roughly into two distinctive categories:
- Less sophisticated chatbots that are based on either a simple A+B=C model that is based on if/then logic, or a straightforward offering of selections from which the user can choose. These are mainly rules-based bots that can assist users with simple, pre-programmed actions, such as password resets or presenting users with current deals and offers.
- NLP (Natural Language Processing) bots: bots which sound more human-like and are not strictly script based in their responses. Such bots can contextualise user requests. These are more complex chatbots that use artificial intelligence to imitate human conversation – they “learn by doing,” in other words, they can improve in their similarity to a human agent thanks to machine learning.
But what exactly is Natural Language Processing?
Natural Language Processing (NLP) gives bots the ability to analyse and comprehend even highly informal language, meaning that at times, we may have engaged with a bot on Facebook without even realising it. Moreover, it represents steps in the process during which text or speech is transformed into a structured set of data. Based on which appropriate response is given to the person in the course of their dialogue with the chatbot. These steps include:
- Sentiment analysis – estimating whether the user has a positive experience or not and whether the conversation should be directed to a human representative at a particular point in the conversation.
- Tokenization – the process where a line of words is divided into linguistically symbolic tokens, or where the words are in some other way beneficial for the application
- Named Entity Recognition – the program model distinguishes categories of words such as product name, users contact details or other data that may be required.
- Dependency Parsing – searching for users’ text for common phrases, verbs and nouns, to find related messages which the users may want to convey.
To perform its function, Natural Language Processing requires another function – Natural Language Understanding (NLU), that entails three notions: entities, intents and context.
- Entities represent the essence of what is offered by the chatbot. For instance, this may be a downloadable guide, a booking system for a beauty salon.
- Intents are the actions that the bot should take when a person says a particular thing – so the same command is triggered whether the user is wording their request as “I want to book a beautician appointment”, or “Do you have any available beautician appointments?”.
- Context is where the chatbot obtains the answer it should give to the person. The NLU algorithm performs an analysis of the user’s sentence and flags specific phrases or parameters based on which the answer will be provided. These can be flags like “Booking a hair appointment” or parameters like “Hairdressers: Toni & Guy”.
Understandably, terms such as AI, machine learning, NLP and NLU can make any less tech-savvy persons head spin. However, this does not mean that you should not take advantage of the benefits that chatbots could bring to your business. The next section will provide a step-by-step overview of how even those with minimal or no IT-skills can start with building their chatbot.
How to begin building your chatbot
If coding skills are not your forte – do not worry. Various companies specialise in making bot-building easy and do-able for everyone.
Here is a list of some of the most popular platforms:
1) ManyChat – arguably one of the best-known on the market and, ManyChat is intended for those with really minimal understanding of coding. Similarly, to Facebook, ManyChat also provides a whole library of materials with all the necessary guidelines to help you with the process of getting everything set up quickly and in the right way. For newbies who are just starting to explore the features of chatbots, there is a free version offered; however, the paid version by far wins with the features available for users.
2) Chatfuel – with high calibre clients such as Lego and TechCrunch on board as their clients, this platform also offers a free option. Though, should you decide to opt for the free version, Chatfuel logo will appear on bot landing and payment pages. Chatfuel is a user-friendly platform that allows people without developer skills to build AI chatbots for Facebook Messenger.
3) MobileMonkey – this bot-building platform is praised for its impressive support with all the necessary training offered for those with minimal technical skills. Just like ManyChat and Chatfuel, you can also opt for a free subscription on MobileMonkey, though the latter comes without significant features such as messaging automation, analytics and integrations.
4) HubSpot Conversations – free edition includes tools to organise and track conversations with existing and new customers. Upgraded and paid versions will give you a more extensive variety of tools, such as custom landing pages, analytics. Automation, more in-depth insights into prospects, automation and other benefits.
5) Drift – this platform offers a comprehensive selection of different packages for individuals, businesses and large enterprises. Individuals can use a free basic subscription; however, some advanced monthly paid plans for business can cost thousands of euros.
There are many other bot builders in available, like Pandorabots, AgentBot and Intercom to name just a few. I would suggest that you first evaluate what features are the most important to your business and then spend some time comparing the various packages offered by service providers to find the most suitable for your needs. Also, make sure to check out customer reviews and don’t be afraid to reach out to the sales teams and ask questions before you decide to purchase.
The next step after you have found the best bot builder for your needs is to write your chatbot scripts. Some of the recommended programs for the writing, editing and sharing of your chatbot scripts are:
- Microsoft Excel
- Google Sheets
- Google Drawings
Alternatively, you wish to consider some other programs including:
Having picked your bot builder, selected a great program were to write your scripts, you are ready to start sketching down script(s) for your new agent bot.
However, what should it say to your potential customers? How should it interact and represent your business? My suggestion would be to adopt an initial approach like you would approach training a new client-facing member of staff:
- What function is the bot serving for your business? – Is the bot there to answer general queries about your products and services? Or maybe to promote a special offer? Design the dialogue with the goal in mind.
- Do your research – who is your target audience, and where/how do they want to reach out to you? How can you best reach out to them?
- Consistency is critical – has your business already created its unique brand? Think about this and make sure that the communication style of your bot would reflect your brand’s overall image. Once you have decided on those aspects, maintain a consistent tone and style throughout the script. Try and make your bot sound as natural as possible, and slightly pace the bot’s responses, creating a small 0.5-second delay between responses. Do not, however, mislead people to think that they are speaking with a real person. It will not reflect well on your brand.
- Respond succinctly – refrain from writing responses that are longer than three lines in the mobile text or separate your messages into several answers. People are no longer accustomed to reading long paragraphs of text, and you may lose valuable customers should you not pay attention to providing short, concise answers. To further improve user experience, prepare several differently worded answers to a single question.
- Be polite and finish with thanking the users and wishing them well!
As you can see, building a chatbot has been made easily accessible to everyone. As with anything, there is tons of additional information and guidelines freely available on the Internet, so investing a little time and resource can pay off. How? Let’s look at some key benefits chatbots can offer businesses in 2019.
How can businesses benefit from chatbots?
No doubt that the level of service provided to customers can make or break a business. Call centres, and more recently, social media channels like Facebook and Twitter have created a new shift, whereby people can contact businesses with their issues or queries at any time of the day.
Most people experience fairly long waiting times to get their questions answered, and here is where chatbots can add value – by offering access to assistance 24 hours a day, 365 days a year. No doubt, a bot will not be able to provide real human interaction in all cases, however, it can serve as a useful middle-step that can prepare customer service representatives to provide higher quality service due to improved understanding of customers’ issues, having gained prior information from the details provided in the conversation with the bot.
Messages are used as the medium for performing countless tasks, such as helping consumers complete their purchases online, scheduling meetings and webinars, and so forth. According to Business Insider, an increasing number of major brands, including eBay, Starbucks and LinkedIn are already taking advantage of the fast-developing bot technology to support various functions within their businesses. Equally, major players such as Microsoft, Google and Facebook have been, and continue investing heavily in the advancement of chatbot technologies.
What kind of businesses can benefit from a chatbot?
Chatbots are not just there for big businesses with an already established customer base. Companies of all sizes accustom chatbots to perform a wide variety of tasks. The question you should ask, is whether your potential customers have questions and inquiries that tend to crop up repeatedly? If this is the case, then creating a chatbot that could solve less-complex tasks and respond to frequently asked questions could be a great time-saver.
Equally, Klientboost suggests that chatbots can be an excellent solution for businesses that offer high-cost products or services that require some pre-purchase deliberation. Bots could add value by responding to crucial questions customers have when researching which product to buy and provide a free downloadable guide at some point in the conversation. Once the customer has received a positive experience through their communication with the bot, it is easier for the salesperson to continue building on that first interaction and close the sale.
Since Facebook Messenger is one of the most used messaging apps worldwide, people regularly message businesses on their Facebook page to learn more about their offers and services. Moreover, Facebook Messenger bots can also be successfully used for building brand awareness and providing your audience with more information about the business. All can be done, for instance, by creating a custom welcome message, and including some essential information about your business.
Whilst it makes sense to optimise the bot to provide a near human-like conversation experience to improve customer satisfaction, it is equally important to remain transparent to the users that they are interacting with a bot – vital from the perspective of building your brand’s trustworthiness, but also, to manage expectations.
Thus, chatbots in 2019 have been developed to a degree of sophistication, where they can genuinely add much value to both sides – businesses and customers. Here are just some of the benefits for users:
- They are available almost 24/7 – apart from any downtime due to security or maintenance issues.
- Instant and consistent responses – though sometimes this may not replace the level of depth that a human responder may be able to go to in their interaction, for the most part, the efficiency of service offered by bots contributes to positive customer experience.
- Conversations and responses are recorded – the customer will have the ability to refer back to their written communication with the bot, and should there be a need, challenge the information provided to them. Mostly not the case when talking to a customer service representative over the phone.
- Instant transactions – Bots can perform specific actions almost instantaneously, for example, obtaining information from a database, or changing customer records.
- Bots don’t run out of patience with the customers.
- Programmability – they enable the automation of the most mundane tasks, such as scheduling bookings and meetings.
Undeniably, bots also offer excellent additional value to businesses, including:
- Reduced costs – as companies grow, so does their need for more customer service representatives. Here, bots offer a great way to manage increasingly complex customer queries.
- A higher level of customer engagement and sales – chatbots represent an additional way to increase engagement between the business and customers.
- Reaching new potential customers – chatbots on major platforms such as Facebook Messenger provide a “non-committal” way for potential customers to reach out to companies, without having to contact them through email or phone directly.
- Source of in-depth knowledge about customers and their needs – records of customer interactions with bots can provide companies with an advantage to better understand customer needs, expectations and problems to improve and develop their products and services in a more intelligent way.
Are machines slowly but surely taking over our jobs?
People have become immune to traditional ad formats such as static images, email marketing and cold calls. Just like the world around us has become more instantaneous and interactive, so has marketing and customer service. Customers want businesses to be at close reach to them when they want their questions answered, but they do not want to receive constant intrusive communication from businesses.
It is not uncommon to hear claims like “machines are taking the jobs from real people,” really this is not the case – at least in the context of chatbots. Bots are here to simplify the job for humans, and where possible, to make their jobs easier, enable a higher volume of customer conversations and more seamless and personalised interaction with the customer.
The experience of conversing with a chatbot should not feel like you are talking to a lifeless robot. It is possible to program bots, so that they would interact in a more human-like manner, for instance, including light humour in their responses. Small touches like that can add a lot to the conversation, improving customer retention and creating an overall positive image of your brand or business.