What about mitigating hotel chatbot failures? Thing is, you have no way of knowing if you’re delivering value to users on your hotel website. There is no feedback loop to verify that you’re doing a good job. You essentially throw content at your visitors, expecting them to like it. With an average hotel conversion rate of 2%, this technique probably needs some improvement.
Live Chats yield a higher conversion rate but most hotels do not have enough human resources to operate them properly. Chatbots are a smart option to help hotels engage online visitors and generate direct sales. However, chatbots are not magical. If you’re aware of robots’ limitations from the very beginning, you’ll find ways to mitigate them and achieve an average of 17% conversion rate. BUT you will have to put work into it: there is no free lunch.
Chatbots – smart tools or stupid machines?
Chatbots can have conversations with humans. However, the processes used to teach them speech is very different from the way we learn language.
Chatbots do NOT have common sense. They do NOT think. Every time you pose a question to a bot it runs a semantic and statistical analysis to determine if it is similar enough to a request it knows the answer to. The more you feed your chatbot database with different formulations for the same demand, the better it will be at finding the right answer.
Thanks to this training, robots become very good at managing frequent requests. However, if it doesn’t have the answer the customer is looking for, or if the question is asked in a very tricky way, your bot won’t be able to respond – and it is not the end of the world.
In all honesty, some failures are due to really stupid customers, but they have money too, so you should really setup a process to call them back and offer assistance.
Most hotel chatbot failures are not related to AI
Artificial intelligence is not perfect but it is already mature for hospitality. In fact, most mistakes on hotel chatbots are not related with AI but with knowledge management. They can be sorted by the hotel itself. You can prevent most mistakes from happening with good copywriting. Usually the chatbot will perfectly understand what the customer is asking for and deliver the corresponding predefined answer. If the predefined answer is not precise enough the customer will ask for details that the chatbot will not be able to give.
When you give answers to your chatbot make sure they cover the question from every angle. Be aware that you have a learning curve too and might need to rethink your predefined answers several times.
How to extract value from failed conversations?
The mission of every bot company is to make sure its chatbot gets it right as much as possible. This article focuses on mitigating hotel chatbot failures and the value you can extract from them.
When we launched Zoe, our chatbot for hotels, I thought that any conversation where the chatbot eventually got it wrong was an unacceptable failure. I would not sleep at night because I was terrified of losing all my hotel customers overnight. Then I realized that I was worrying for naught as hoteliers are ultimately the ones defining the success or failure of a conversation.
Zoe retrieves the customer’s name, email address, mobile number and his expression of need 100% of the time. It means that even when the conversation derails, she always succeeds in turning an incognito visitor into a qualified lead, making it extremely easy for hoteliers to re-engage by taking over the conversation, calling back or emailing. When hoteliers actually follow-up on sales requests that the chatbot couldn’t handle, the conversion rate is about 17%.
Put yourself in the shoes of a customer: you started a conversation with a chatbot, asked for information regarding an event you’re organizing, and at some point, you asked for the price of the meeting room + 10 rooms for 2 days. Zoe cannot answer that, but if you call back with just the right offer, the customer will be happy anyway and end up buying from you.
Regardless of how good or bad your chatbot is, its notification center is much less sexy but just as important as its AI component. The notification center is the key to staff adoption; do not overlook it.
Some staff members are always on the move, so a mobile app can prove just as handy.
What to do when the chatbot fails and there is nobody available at the front desk?
Most times, these conversations are just dismissed by hoteliers. However, that would be a mistake. Since you have all the customer details it is essential to always call back, email or send your answer via text message. If you do that, customers won’t even remember the chatbot and thank you for your service all the same. Yes, it works even if the bot interaction happened hours ago.
When a customer interacts with you it is almost certain that he is going to travel to your destination. The question is: will he go to your hotel or your competition and will he book direct or through an OTA? Take that into account in your reply message or call-back.
The structure could be:
- Answer the customer’s question
- Create an opportunity to nudge the customer to choose you. It can be a special offer, something free or, if you really have nothing else at hand. Tell the customer how much you look forward to welcoming him. Since you know what he wants, it should be easy to think about the right thing to say.
- Conclude with a clear call to action. It can be a link to your booking engine where you have preset the dates and occupancy. You can request a confirmation at the end of your email or SMS.
Chatbots are a fantastic opportunity to engage customers and generate bookings. They have very clear limitations, which you should be aware of in order to overcome them. It is all about mitigating hotel chatbot failures.
Bots that are good enough make customer interaction a viable option for hotels and increase direct bookings. Mobile and interaction are hot challenges for hotel marketing and chatbots are clearly the answer. The question is: are you ready to do what it takes to succeed with this new form of marketing?
If you’d like to discuss more about mitigating hotel chatbot failures or have questions about artificial intelligence for the hotel industry, please visit our website or contact Benjamin Devisme at bde@quicktext.im.
