The 2019 A to Z of Chatbots

The 2019 A to Z of Chatbots

Chatbots might be a new technology to many, but they are already a key tool for huge numbers of businesses, providing increasing value to their customers with useful information. Rather than take the grand tour of product brochure pages, here are the key terms to consider when looking for a chatbot suite or tool to build your creations.

Artificial Intelligence is a major buzzword. Chatbots don’t need AI to succeed, but boy does it help. AI helps bots better understand the questions put to them by users through natural language processing analysis of the text to discern key phrases. AI can also provide smart searches of wider information sets to help provide answers to more open-ended questions that a simpler bot might not understand.

Business use cases for bots are growing, from the established basics like customer service and taking bookings to a growing range of sales and upselling from the bot. Advanced use cases include deep conversations for medical, legal or other purposes. Whatever bot plans your company has, they will only succeed if it fits in with the business goals, improving performance, efficiency and creating new opportunities.

Chat quality is the same as when we judge another person by how they speak. If your chatbot comes across as dumb and lifeless, people won’t want to engage with it. Work hard on making your chatbot text, lively, pithy and to the point.

Data life is how long the content of a chat remains on your systems and what other services can access it. In some cases, advanced bots find it useful to access previous conversations to shorten new queries or retain previous answers. However, data protection laws and guidelines might come in to play, so ensure your business has all the legal and ethical angles covered when retaining chatbot data.

Enjoyment of a bot might not be a high priority for businesses such as a car part ordering or similar service. Yet, other bots like those of charities, schools, fun brands and entertainment services need to be actively engaging, so users enjoy talking to the bot. Make the bot fun first and informative a close second to capture the interest and hearts of visitors who will appreciate the love and interest from a bot ahead of the core information.

Faceless chatbots? Don’t worry about it, your bot doesn’t need an image to go with it. But, if your business has the inclination, then bots can have digital faces, bodies, whole AI personalities crafted for them to make them more accessible and engaging for customers. Companies like Expressive AI can give you bot a face from a cartoon

Games, quizzes, exams and other set-response tests, either for fun or education are a useful way that chatbot platforms can deliver tests. Bots can rapidly deliver personal scores and overall results, with back-ends providing team scores or access to competition prizes. With job interviews and other wider conversations being held by bot, there are plenty of ways to exploit them for engagement.

Happiness is something that is rarely found in technology. iPhones, smart home devices and games machines might deliver a positive experience. But can technology really make someone happy? Chatbots are in the frontline of customer services, where the core aim is to make each person happy. Businesses need to remember that when creating their bot, not just delivering a positive outcome, but crafting it in a way that makes the customer happy.

Intent is what a user is trying to achieve with your chatbot as part of their input. If they say “When is the next flight to Madrid?” the intent is to find out flight times that are immediately available. The bot should be able to discern the timeframe from such a query and provide actionable information based on the keywords in the text.

Julie does it best, as you would expect from the massive infrastructure and needs of US rail giant Amtrak. When it comes to chatbot success stories, there are many to look up to, but Julie is regularly cited as delivering the goods and passengers, with the right information, helping streamline the customer service needs of hundreds of thousands of daily commuters. When planning your business chatbots, look at how Amtrak delivers Julie or how any of the smaller success story companies have transformed or boosted their business through the use of bots.

Keep it simple. For a small or large business, sinking time and effort into a major bot project can lead to overcomplication. Create the first bot and get it out there when it is capable of delivering an overall high-quality service, then aim for perfection and the extra tweaks that will make it stand out from the growing number of rivals in your market.

Language support is a growing feature of bots, enabling them to deal with customers from around the globe and help a business expand without adding bloat to the organisation. Delivering messages and translating responses can be handled by a growing number of bots, helping the company do business in new territories and handle queries from any time zone..

Money is at the root of most business transactions, and bots can join in the fun, offering sales, upgrades, accessories or additional services to a customer based on their previous purchases and needs. Most bot packages can offer PayPal and other payment services within the bot to bring revenue into the business without the need to talk to a customer service agent or go through a typical sales channel.

Natural Language Processing (NLP) is a key part of chatbot technology, allowing the bot to inspect user inputs and break down sentences into words, establish what the important elements are and respond appropriately. More bots feature AI elements like NLP, natural language understanding and growing deep learning capabilities to extract meaning and understanding to improve the quality of conversations. While there remains a lot of hype around AI technologies, the core benefits help you build a better bot, and you shouldn’t be put off by the jargon or hype – adding simple NLP to a bot is not a complicated task.

Open-mindedness is a growing part of a bot’s success. While your business might have a male or female dominated clientele, the bot needs to appeal to all audiences. Health bots need to understand the differences in age group, while mental health bots must be mature and sensitive enough to be able to appreciate the variables in gender, emotional state and other aspects without causing unintended offense. All of these require extensive testing with live users to totally master these nuances.   

Personality is something that every chatbot should have. Many businesses are institutionally afraid of giving a bot a sense of humour, or making their virtual representative seem overly chatty. But, people expect and need a degree of personality in a bot to encourage them to talk to it. Brands are more likely to give a bot personality, based on their products or the type of conversational output their social media teams use. Any business should be open to giving their bot a bit of character and the basics of some too-and-fro in the chat.

Quality is at the heart of a chatbot’s reason for existing. They need to provide high-quality information rapidly to satisfy the customer. To create that quality, companies need to study best practices, master brevity to deliver concise information and learn from both successful and unsuccessful outcomes what works and what doesn’t.

Rich content means a bot doesn’t have to be just words. Bots can be used to show pictures of products, run a video to help explain a support issue, or use audio and music to add a more engaging tone to the conversation and provide other benefits. In short, don’t think a chatbot has to be solely about the word content, find areas where you can add colour and visual flair to make it more useful and impressive. .

Supervised learning is how chatbots can be trained through machine learning to deliver better results. Plain machine learning is how the bot trains itself, but supervised learning lets the bot make suggestions for a better response, with a human trainer selecting the best answer or suggesting a more refined version. Live bots can also ask a human supervisor for confirmation that they are taking the right approach with a customer.

Thank you and please are some of the most powerful words in the English language. Remember to include them when your bot is welcoming people to the service, providing information and soliciting responses. Apologies for delays or having to hand the conversation can be done in a positive manner, but remember the basics of politeness.

Unified approaches to chatbot launches should see the bot’s launch clearly highlighted and advertised, with customer awareness sessions or information to teach people what it is and what it does. Launching on social media platforms, messaging tools, collaboration services, apps and other avenues show the business is serious about its bot efforts, and less likely to consider it a sideshow or experiment they don’t mind failing.

Virtual Assistants like Siri and Alexa are rapidly becoming synonymous with chatbots as a technology, despite the two starting at very different ends of the IT spectrum. Chatbots started out as a simple text-based help service for websites or apps, while the VAs provided cloud-driven voice assistance on smartphones and smart home devices where voice interaction made for simpler interaction. Both technologies are borrowing from each other and merging to provide best-of-breed help, wherever the user is, and the trend will continue until VAs and Bots are largely one and the same.

Wizards are just one of the tools that any business can use to create a feature or product. The major vendors, like to sell their chatbot products as magical with all the AI bells and whistles. But any company can build a chatbot using a What You See… tool or wizard that makes the building process a block-by-block approach.

eXtend your bot as customers get used to the original features and treat it as a regular part of the business. The more you can add to the bot, in the form of information services, wider sales offering or greater knowledge bases, that provide more reasons to use it will cement the bot in the minds of customers as a truly useful tool.

Yesterday is the day you should have launched your bot, or last year. But since there’s no time like the present, focus your IT team or marketing department on building a bot that can deliver the required level of service to customers as quickly as possible. Your rivals will already have products on the market, or their own plans, and the longer you wait, the further you will be behind in chatbot development experience, customer satisfaction and other areas, just as market acceptance becomes universal.

Zealotry is something that chatbots should avoid. While they can recommend your products or brand, having an overkeen chatbot monastically spouting the company slogan or every single little feature of a product is not what customers need from a bot service, they need help, advice, options and a reasonable level of engagement, just leave the PR bullshit behind.

Chatbots need not be complicated, but they exist as part of a complex world of IT, digital services and an increasingly confused customer landscape. Among all the jargon and buzzwords, your customer should only see a helpful tool that makes their life easier and delivers valuable information or resources. How complicated can that be?