Need to create a chatbot? No worries, we analysed and compared main platforms’ features for you!
Before diving into the platforms available to create a chatbot, it’s useful to clarify a key aspect: chatbots are not directly connected to its visual interface nor can be directly related to the messages exchanged within the platform. Virtual assistants are the result of an extensive developing process and only during its final operational phase they are connected to a visual interface for messages exchange purposes.
The first adaptability experiments between conversational chatbots and third-party platforms begun during the 90s but we have to wait till 2013 to experience the first international commercialisation of the technology thanks to Asian leader WeChat.
Two years later, Telegram followed their footsteps offering end-to-end conversational encryption to their users, a compelling and competitive advantage that lasted until WhatsApp adopted the same technology into their platform one year later.
Viber, Skype and Slack – communication tool for companies – are the other main messages platforms that also allow virtual assistant integrations in them.
Messages exchange aside, the use of chatbots for marketing, customer care, and sales purposes became common practices only after Mark Zuckerberg launched such option on Facebook Messenger in 2016. It was a sudden success since the platform was already comprised of million of business pages eager to test new, automated communication processes with their users.
Now that we briefly introduced the history of the technology and its key players, let’s discuss the two possible ways of choosing the best messaging platform: create a chatbot from the ground up or start within a pre-built platform.
Chatbot creation: what option to choose?
Let’s be clear: there isn’t a universal answer, it mainly depends on clients’ policy and their needs.
Often, building a chatbot from the ground up isn’t the best option since the amount of time and effort behind creating the core machine functionalities (primarily data sync and messages management) can be easily replaced with pre-build and free services available on the platforms mentioned before.
Sometimes however, custom platforms are required when companies demand chatbot implementation within their website or internal chat, or when they need a chat without user authentication. Companies operating in restricted sectors such as the energy trading, need to share infos publicly without requiring user sign-ins.
[bctt tweet=”What’s the best platform for your chatbot?” username=”ndg_ai”]
Lastly, programming complexities aside, custom platforms give more autonomy and ease of solving issues when they arise.
READ ALSO: NLP and Machine Learning: the thinking side of chatbots.
Tools for chatbot creation
Each platform has a set of elements available for the users. This allows to enrich the dialogue flow with the bot, otherwise only limited to text.
All the platforms allow to add buttons, in different shape and locations, within their interface in order to let users process direct tasks and intents to the bot.
On Facebook, they are real buttons while on Telegram they are displayed as hyperlinks. They are different design approaches based on the desired perception to be given to the users: on Telegram and Slack, buttons are perceived as real commands to the bot while on Facebook they are intended more as an option within the whole UX.
Professionals working with NLP (Natural Language Processing) see a chatbot build on buttons as an interface without text that is solely based on a tree-system of pre-selected options, a pretty unconventional approach for what is supposed to be a conversational tool.
Once again, it will be the scope of the bot that defines the right balance between buttons and text-based dialogue. Customer care and bot mainly created for engagement purposes will have an interface primarily based on buttons, while solutions conceived to navigate the users among different services and products will be build with conversational interactions in mind.
In this case, platforms such as Facebook Messenger offer text-based suggestions – called Quick Actions – that guide the users toward the final objective of the whole experience.
They are sliders with the function of suggesting actions (purchases, contacts to add, etc) and update users on group activities and following page news.
They pop up as classic webpages so that’s why native mobile platform such as Telegram and WhatsApp don’t support such feature.
Carousel on Facebook are really popular and are shown as a mix of images and titles comprised of up to ten slides.
Conceived as b2c audiences, later on they had been adopted by famous b2b platforms such as Skype, though with reduced functionalities.
Web Views and Web Extensions
Web View is a system component that allows platforms to show webpages content on mobile. A real browser that pops up within the tool, exactly how happens on Android with components that use Chrome technology.
Web Extension is a sub-category of web view, it’s proprietary of Facebook Messenger and it allows developers to create small web apps within the platform in order to track and retrieve user IDs so that they can geo-locate page viewers. A really valuable piece of information for better and more customisable offers.
Differently than carousel, it’s possible to add Web Extensions within platforms that not natively support such integration by creating a dedicated web app to later add into the chat. It’s important however to keep an eye on privacy and security constraints since data have to be manually encrypted. In fact, without native extensions such as the Facebook Messenger one, IDs can be retrieved by adding custom parameters into the related page url.
The possibility of tracking user location has plenty of great advantages. For instance, it’s possible to suggest the right type of info or even language setting based on user’s current location.
Telegram does that automatically on background while Facebook start tracking such parameters after obtaining users authorisation. Two different approaches that are most likely due to different policies applied on the countries where servers are located and the amount of data the two companies manage (Facebook most certainly gets access to way more user info than Telegram).
The last messaging aspect analysed on this post is the possibility to add bots into group chats, a really common option for bots that manage work activities as well as bot dedicated to user free time, such as musical playlist management.
On this aspect, countries such as Cina are way more developed than western countries where such service is not yet fully developed.
The most used interface in Asia is WeChat, a platform that not only allows implementations with third-party payment methods but that also gives the possibility to add a broad option of services such as Uber, food delivery providers, and QR code integrations for store payments.
In the western part of the world, payment via chatbots were originally allowed only via Facebook Messenger in the US (the first service ever to adopt the technology was flower delivery company 1-800 Flowers) but nowadays other platforms such as Telegram begun adopting such integration for their users.
Ultimately, the right platform choice for your chatbot is strictly related to your needs. Probably, Slack and Skype would be the best option for work-related purposes while Facebook Messenger might better fit b2c needs.
How can users engage with the right chatbot at the right time? There are dedicated sources and channels for these type of searches but also specific approaches like the one offered by Facebook, where people can connect to the chatbot via QR code. All platforms implements a chatbot search engine within their UX – like the Discovery section on Facebook Messenger – so that people can always search and find the right service based on their needs.