Let’s talk about Chatbots
What are chatbots and how do they work?
Chatbots are programs designed to simulate a human conversation. We typically interact with chatbots through a written-chat interface, or through voice. The technology interprets and processes user’s words or phrases and gives near-instant answers. There are multiple types of chatbots, and their functioning varies based on how they answer and what they can do:
Rule-based chatbots: based on a pre-determined set of user entries, they provide logical answers. These are not conversational per say; they are typically to answer simple and specific questions or take appointments.
Intellectually independent chatbots: these use Machine Learning and are trained to understand specific keywords or phrases, and provide answers matching the theme they understood. If a query is outside of what they have already been trained on they might not provide accurate answers.
AI-powered chatbots: they combine the rule-based and intellectually independent features, layering artificial intelligence on top. These chatbots can process natural human language and provide smarter answers than their two counterparts. While not perfect, as human intent or mood are complex for computers to understand, they are increasingly accurate and can even anticipate needs.
Why are chatbots trending ?
While the technology has been around for many years now - Eliza, the first Chatbot was created in 1966 - it has significantly grown in adoption. Preparing this blog post, I remembered writing a post on the rise of chatbots back in 2017!
What makes them trend now is two-fold:
First, the underlying technologies (A.I and Machine Learning) have made significant progress, and the user experience is therefore much better than the early days robot-like and limited experiences. The progress in technology also widens the array of things it can do. The use cases are now virtually limitless, ranging from medical applications to Fintech, Retail to Edtech, and customer service to leads generation.
Second, the accessibility of the technology has progressed so much that it is now available to any company. The cost of building a chatbot isn’t prohibitive anymore, and companies can get started with a limited investment. Additionally, the technology doesn’t require a complete developers team and an extensive IT setup: multiple options have flourished offering to build a chatbot with almost no prior coding knowledge, allowing brands to test the waters with a first step.
Why build a chatbot?
Building a chatbot allows you to be present for your customers at any time, with an automated customer service available 24/7. Being able to “talk” to the brand, will increase your customer’s engagement. Having a chatbot on your website can help you catch leads when people are browsing your website or provide them with relevant and timely support if they are experiencing trouble with checkout or constantly returning to the same page for example.
They can also help you grow your audience by driving them to your website or app. Furthermore, they are a great place to boost your content’s visibility and engage dynamically. Last but not least, chatbots allow you to seamlessly gather data about what your customers need the most support for / wonder about, and therefore develop solutions.
Are there any risks associated with Chatbots?
Yes, as with any opportunity, risks are associated, but when anticipated they shouldn’t deter you from getting started.
The most common risk resides with data security: you want to make sure you know where your data is stored and safe-guard it to avoid any leakages, in particular if you collect personal data with the customers interacting with the chatbot.
Another risk is to offer an experience which is not up to par with your company’s image. You are better off not offering a chatbot at all rather than offer one which doesn’t hold up to your standards of quality. A customer engaging with your chatbot when encountering an issue will be exponentially frustrated with your brand if the chatbot doesn’t offer compelling answers. Furthermore, a new customer experiencing your brand through your chatbot should be wowed and not scared or disappointed by a poor experience.
Finally, as for humans, languages are key to the chatbot experience. While AI and NLP (Natural Language Processing) have made impressive progress in English, other languages have not yet benefited from the same attention and developments, and the experience in other languages might not be up to par yet. Also related to languages, the intent of the customer, his/her mood, and body-language and so on are - for now - out of reach for a computer to understand.
Which platforms can chatbots be built on?
Companies can choose to build their own chatbot directly in their apps and / or website, or they can choose to offer a chatbot where their customers are, such as on Facebook Messenger, Instagram, or Telegram.
Larger companies, with significant resources and an established IT ecosystem might favour a custom development to own the process and data end to end.
Smaller companies, or first timers, can dip their feet in the water and customise pre-built chatbots to interact with their customers on their websites, or engage with their audiences on Social Media.
For example, Mobile Monkey, provides a hassle-free chatbot for Instagram which can be set up rapidly and without coding knowledge. Live in the matter of minutes, it provides a great option for first timers.
How do I get started with chatbots?
Are you now ready to get started building a chatbot? GREAT!
The first thing to do is define what you want this chatbot to do; reflect on the specific use cases the chatbot it is going to support and identify the most compelling ones.
Do not build a chatbot because it is trendy, build it to serve a company or a customer need. Improve the process through the chatbot. This way, your chatbot will be relevant and useful.
Second, pick your platform, unless you want to develop it from scratch for your own app or website, make sure you pick the most compelling platform to build your chatbot on, based on your use cases.
Third, make sure your company is set up for success by training the relevant teams: if the chatbot comes as a complement to customer service, make sure that team knows the chatbot inside out, if it is geared towards generating sales leads, your sales team better get ready to have a lot of new leads to follow up on.
Last but not least, remember to keep a hand on the technology, and test it often to make sure it works as expected, and make it evolve where it needs to.
Sounds like a lot? Get in touch today, we’ll help you figure out your approach and build your first chatbot!
Learn more about chatbots
- Complete guide to Chatbots for Marketing