When it comes to digital communication, hardly anything has been discussed as often in recent months as interacting with bots. They are also becoming increasingly interesting for museums and cultural institutions — after all, there is often a shortage of personnel here. Can the dream of a fully automated visitor dialogue on social media be fulfilled? Can a ChatBot even end up supporting a community manager in their work? It’s worth a test. Here you can find tips, tricks and tools on how to create a ChatBot yourself — from conception to implementation, without programming.
See also : How do you create a WhatsApp chatbot?
Plan de l'article
The triumph of bots
Anna is likely to be one of the most famous chatbot personalities of the past — IKEA’s virtual assistant. The chatbots.org site currently lists 1,223 similar chatbots for 50 countries. For Germany alone, 56 relatives of Swedish Anna are named, mostly by the way, sisters with names like Julia, Clara, Sophie or Nina — virtual maids and robo-helpers.
Could these digital assistants also take over the community management of museums, theaters, libraries or concert halls? Will the support of a ChatBot make it possible to offer continuous customer service 24 hours, 7 days a week? Can the social media manager of an institution be relieved of his work?
See also : How do you set up a Facebook chatbot?
Technology is still in its infancy. But answering FAQs — frequently asked questions — is already quite straightforward. This allows users to be automatically directed to information or find answers to their most pressing problems — without taking up an employee’s time. And the best thing: Setting up a chatbot is possible for everyone — even without programming knowledge.
(screenshot: Facebook messenger/novi chatbot) News from Resi and Novi
Messages are now being sent by “personal news assistant” Resi, “exactly as a friend would do via WhatsApp” — wink wink, crying emoji.
Since the beginning of the year, the public service offer funk has also been offering news via ChatBot since the beginning of the year, developed by tagesschau and NDR. Everyone can test how mature the bot is on Facebook Messenger (@getnovibot) themselves. But be careful, the virtual message delivery agent is a pre-planned sensitive device, not artificial intelligence. If you write why you should be interested in the messages sent to you, he reacts with many sad emojis and a story about Beyonce.
(screenshot: Facebook messenger/novi chatbot) The conception of a chatbot
After an extensive test of “competition”, you can start building a dedicate your own chatBots. The structure of the virtual dialogue can first be developed and tested using the Botsociety page. Chat histories between bot and user (for example Facebook Messenger) can be simulated here. Various content elements that the ChatBot offers to the user are also available, such as text, gallery elements, images or selection buttons. As soon as a selection is scheduled, an additional field opens in the right-hand side of the screen, which maps the different storylines. With this overview, every chat history can be simulated and planned out.
The following aspects should be considered when designing the chat history: (Screenshot: Botsociety, Musermeku-bot mockup — Bot Dialog content elements)
- The user should be made clear to begin with that they are communicating with a ChatBot, not a real people.
- The ChatBot should clarify what it can do — and what can’t.
- The bot’s messages should not be too long, but short and easy to understand.
- storyline should be developed for all predictable requests and interactions of users.
- The individual histories of a storyline should work independently.
- A relapse level should be scheduled if a user behaves unexpectedly. You should either refer to contact information for a person or you can specify navigation options for the chat.
(Screenshot: Botsociety, Musermeku-bot mockup — history of storylines) Implementing a ChatBot
If the concept stands for the ChatBot and for all storylines you want to communicate, it’s up to implementation. For a ChatBot that works via Facebook Messenger and is linked to a Facebook page, the Chatfuel page is recommended. All content prepared via botsociety can be transferred here.
It is also possible to specify bot responses via AI function for chatfuel in response to free text input from users. This includes, for example, reactions to frequently asked questions that you do not want to offer as selection buttons to make the dialogue between bot and user more dynamic. Classics would be the expected questions about opening hours or contact details, which the bot can answer sober factual or playfully supplemented with emojis depending on the concept. In this way, other dialog elements can be scheduled “out of series” in addition to the storyline.
(screenshot: Chatfuel response definition for free text input by users) Once you have transferred the concept of your chatbot, tested and connected to your desired Facebook page, you can by the way, use the offer from every Facebook user via messenger. You can also promote the chat function via your page via template and view statistics on how to interact with the bot via chatfuel.
By the way, further messenger services are in preparation — bots for WhatsApp, Telegram and the like can soon be created via Chatfuel.
Header image: Angelika Schoder — Hello, Robot — Vitra Design Museum, 2017
Angelika Schoder @musermeku
Angelika Schoder writes
about topics related to digitalization, museums and exhibitions, and travel and culture tips about the author
- ChatBots as interactive guides in
museum The Art & Culture Newsletter
In our free newsletter, we inform you about the latest news from the arts and culture sector once a month.
→ Subscribe to our newsletter