Looking Back: Best Chatbots of 2017
With 2018 just beginning, we thought it would be great to take a look back at some of the best chatbots that changed the way we had conversations online last year.
— — -
What were the best chatbots of 2017? It was the year of the chatbot, every week an anthology of chatbots were born but which stand out from the lot and are worth being used?
Being a conversational UX designer, I put a point of attention to the conversational side of a chatbot. It must solve a user problem: the conversational must be the answer to this problem and not just a simple tool. Moreover, the adoption of the project by its target can be a revealing element of the relevance of the chatbot.
Let’s not forget that the technical aspect is important, the use of AI as the understanding of natural language, the image recognition, the analysis of feelings, are to emphasize as they bring value to the experience.
Now, let’s move on to what you all expect, the list! 👇
Best chatbots of 2017: Zo (Conversational experience)
After the failure of the Tay experience. Microsoft came back with a conversational chatbot, Zo.
Microsoft intentionally wanted the bot to avoid sensitive topics to avoid bad-buzz. And it turns out that Zo offers a truly conversational experience.
Zo can do else than chit-chat, and this is a technological feat when you consider the state of the art of AI.
This bot shows that it is possible to offer a conversational agent that is based on deep learning while having a relative control, over what the bot says. And this is good news as « hard-coded » bots — the majority of chatbots today — are limited and will not be viable in the long run.
If there is a bot to try for his technical demonstration, this is it!
What to remember:
- A truly conversational experience based on machine learning.
Best chatbots of 2017: Meekan (Utility)
When we think chatbot, we often think of user-oriented uses in a context of daily personal service or entertainment. However, the transactional automation offered by AI can offer a high value-added service to employees within a company.
This is the case of Meekan, integrated with Slack or Teams, it will allow you to manage for you the planning of meetings. We all know it’s a well-identified pain-point when you’re in the office, it can quickly become a nightmare for the organizer!
The artificial intelligence is in the anticipation: No slot available? Meekan will set up a doodle. Remote employees? The bot will propose to create a conference call.
Moreover, the company was bought by Doodle in the middle of the year, a sign that the organization increased by the AI has good days ahead!
What to remember:
- A very interesting value proposition, so much time saving can be important.
- Use of natural language which gives the impression of having an assistant.
Best chatbots of 2017: Trivia Blast (Entertainment)
Why so serious? ;)
And if chatbot mass adoption went through entertainment? This is what offers Trivia Blast, a trivial pursuit chatbot. Here we won’t have a conversational experience but a new way to have fun in a natively messaging application.
Widely used, more than 50,000 fans on their Facebook page, Trivia Blast will offer many formats in the chatbot to answer questions, but also a Facebook Messenger chat extension to play with your friends in a group chat. Even if the bot is not very talkative, the use of the NLP is very well integrated with some situation, as for choosing the subject of the questions.
What to remember:
- A complete experience: all formats are used optimally.
- The NLP as a mere feature — Less risk of disappointing experience in case of misunderstanding.
Best chatbots of 2017: Hipmunk (Travel)
Available on Messenger and Slack, the Hipmunk chatbot will be your next travel assistant. It simplifies travel planning by providing a single point of entry for research, memorizing and personalizing the flight recommendation.
The bot embeds a push function to be notified when a good deal is on the flight of his choice.
The features are limited but the personality of the bot is really well done it makes the experience very intuitive and human-like, contextualization also helps to enhance the conversational and especially the smart part of the bot.
Think about it on your next trip! 🚀
What to remember:
- The bot’s personality
- The push strategy
Best chatbots of 2017: Poncho (fur factor 😹)
Do I really need to introduce Poncho, the weather bot?
It dominates all the rankings of chatbots since its launch in early 2016 — it was then part of the first bot line-up Messenger when opening the Messenger platform to chatbots developers.
Simple functionality but five-star experience: Poncho is not trying to do too much but what he does, he does it well.
We have fun testing his small-talks as his personality is cool, the messages that accompany the daily weather report are always accompanied by witty messages.
The real question is, when will Poncho have a voice? A conversational voice experience seems to be a perfect match for this type of use case.
What to remember:
- Poncho offers a simple feature but delivers more than you would expect from him.
- Having a bot using NLP early allows a good maturity on the understanding of natural language.
That was it for 2017!
There has been great progress in all areas but the road to mass adoption is still long. Brands are positioned on this medium which has a lot of advantage if they are well identified and used.
Instead of foolishly speaking about the “hype cycle” that would like to put the bots in a phase of delusion (hi Gartner!), I think we have too much talk and share sensational articles on the background of science fiction and other fears of readers (no, AI is not going to replace your job anytime soon).
We have everything to prove to users and their expectations.
Let’s focus on what we expect from us: to advance the experience by relying on the available technical capabilities and solving existing problems by taking into account the characteristics of this new medium.
Read more over at Botsociety’s blog centered around UX design and conversational interfaces.
