The 4 key success factors to maximize your hotel chatbot performance
Hotel room booking chatbots are powerful tools but they are not magic genies. This article will give hoteliers the 4 factors that they need to consider to make sure that their hotel chatbot project becomes successful.
The first 2 success factors are related to choosing the right provider for hotel room booking chatbots. However, you also need to know that half the success depends on your own work. Hence, the 2 last success factors concern what you must do to make sure that your chatbot is doing a great job.
Success can be measured in terms of:
- ROI: A good hotel room booking chatbot should help your hotel generate +30K€/year in direct sales (average for 100 rooms)
- Successfully handled conversations: Once your chatbot is fully operational it should be able to handle 70 to 80% of customer interactions automatically.
1 1. The AI component of your hotel chatbot2 2. Your hotel room booking chatbot design
1. The AI component of your hotel chatbot
Your hotel chatbot’s ability to understand what your customers are asking for is essential. The problem is that hoteliers are not AI experts. Also, there is a terrible information gap between what potential providers tell you and what you can verify. So choosing an AI partner that you can trust is essential. Here are some important questions you need to ask potential providers. This will help you bridge that knowledge gap and make the right decision.
Are you a hospitality specialist?
The kind of AI Solutions that we use for chatbots is called Natural Language Understanding (NLU). Hotel chatbots do not think. They are only machines trained to recognize specific topics and match them with the right answer. Your chatbot’s ability to understand is directly correlated with the accumulated experience of your provider. Experience takes a lot of time and practice. It is a mistake to think that you can start from scratch and reinvent the wheel.
What are the tasks that your chatbot is already able to perform?
Of course, your chatbot will grow over time. However, there are 4 essential capabilities that you should look for from day 1:
- Answering frequently asked questions (FAQs).
- Ability to perform tasks thanks to 3rd parties connectivities (API). For example, a connection to your booking engine is essential for room reservation.
- What happens when the chatbot doesn’t have the answer?
- Chatbots are not only about giving information. They are also here to capture customer details that you can then use for sales purposes. So be sure to ask about lead generation.
How do hoteliers populate your chatbot answers?
If your potential partner gives you an Excel spreadsheet, run away! Excel means that you won’t be able to update your content easily You’ll constantly make changes and improve your answers. So, look for a partner that can give you something more practical than an Excel sheet.
If your chatbot provider is experienced in hospitality, he should already have prepared almost 100% of the questions that you want your hotel room booking chatbot to answer. If you see that the pre-established question list is too short and you’ll need to think about adding specific questions yourself, again, run away! There is one rule in development: If the exterior is ugly, the code behind is ugly too. You can’t check the code that sustains what you’re buying so as a rule of thumb: The more you can see that a potential partner has invested in ergonomy the more you can be reassured.
2. Your hotel room booking chatbot design
Artificial intelligence is great but it’s limited, and that’s okay. Once you accept that simple fact you can start using AI. Use it for what it’s worth and think about the other 3 pillars that sustain your chatbot.
Conversation design refers to your hotel chatbot’s ability to fall on its feet most of the time. This ability of your chatbot depends mostly on your chatbot provider.
There are 3 elements of design that are particularly important and require your attention:
Something that may seem obvious to you is sometimes not so easily understood by your customers. One example is the “book now” button. Customers would go through the room reservation dialog, see the rooms offered by our chatbot and the big “book now” button below the picture of each room. Some customers would not click on it because they are not frequent travelers. So, they do not know that “book now” only means “move 1 step further into the booking engine”. Instead, they would say “I want to see prices before I book”. The conversation would thus get derailed. The solution was not more AI but more user-friendliness. We added a second button “see prices” and the problem was solved without the need for an extra layer of AI.
Sometimes your chatbot will get confused and deliver the wrong answer. A typical example is “I want a room for my family consisting of 2 adults, 3 children, and a baby”. Here the AI component will hesitate between “book a room”, “family room”, child policy” and “baby cot”. At this stage there is nothing that AI can do to solve this. However, good design can. If you know the consistent mistakes, you can choose to add smart suggestions below your chatbot answers just in case it answered wrong. This way, customers will get their answer very fast. Furthermore, they won’t even notice that the chatbot was confused.
Fallbacks and notifications:
What happens when your hotel chatbot doesn’t have the answer? How good is your chatbot at getting customers to rephrase (fallbacks)? Also, when necessary, how easily can your staff take over (notification center)?
3. Your hotel chatbot’s answers
A great chatbot provider is necessary, but not sufficient to ensure the success of your hotel room booking chatbot. A strong AI will be very good at matching your customers’ questions to the right answer. Likewise, good design will keep a lot of conversations on track. Subsequently, the next big pillar is your own input.
At some point, you will need to tell your hotel chatbot what answer it should give to specific questions. The way your answer is formulated is critical and entirely up to you.
When you populate your hotel chatbot answers you must cover each topic at 360°. Leave no breach because customers are always going to ask precisely about what is missing. You probably won’t get it 100% right the first time. And you’ll need to improve your answers over a few months by reading conversation history and asking yourself, “Did my chatbot make this mistake because of its limited intelligence or because I gave it a poor answer?” Do not expect your chatbot to improve your answers magically. It is just not that kind of AI.
To help hoteliers save time populating their chatbot database we have created 2 types of answers:
- Templates where you just need to insert your data points.
- Custom answers, if your hotel has a very specific way of handling check-ins for example, and you prefer to answer your own way.
4. Follow up process
If you’ve made it this far, your hotel room booking chatbot is probably excellent from a technical standpoint. The last pillar of successful hotel chatbot implementation is the follow-up process that you need to set up.
Some conversations will require human intervention. This could either be because the chatbot didn’t have the answer or to follow up on a lead for a complex group reservation, or maybe it is just the fact that you the quote is for 5000 Euros and the person is just not going to spend that amount on an impulse.
Depending on how you are staffed, in case of need you can decide to answer your customer on the fly while they are still online. However, many hotels do not have the manpower to do this. Therefore, a much easier alternative would be for the hotel sales manager to sit down at a precise time every day, say 4:00 PM to deal with:
- Reservation requests: Have a look at the request. See if they converted and take action if they didn’t. These are still hot leads!
- Unanswered questions
People take time to make their decision. Consequently, even if your answer didn’t come immediately, as long as it comes before your potential customer has booked at another hotel, you’re fine. Of course, sooner is often better.
Viewing each conversation as a sale opportunity will go a long way in setting up the right processes. When a customer speaks with you, it is almost certain that he is going to travel. The only question that remains is, “Is he going to travel with you?”
Your hotel room booking chatbot should always give you the customer’s contact details. Then, you can re-engage by email, phone, SMS, WhatsApp and so on. What you are selling is expensive and some people need more assistance than others. Your chatbot takes care of the first level of support and your sales manager should be in charge of the second level.
For second level answers, here are the key elements of a follow-up message:
- Answer the question that your customer asked.
- Nudge your customer. You should create a reason for the customer to choose your hotel. It can be a special offer or something free. Then again, if you really have nothing else available, tell the customer how much you look forward to welcoming him.
Add a clear next step:
- A call to action: link to your booking engine where you have preset the dates and occupancy.
- (Optional) A date at which you’ll get back if you have no news.
If you’d like to discuss more about hotel room booking chatbots or have questions about artificial intelligence for the hotel industry, please visit our website or contact Benjamin Devisme at firstname.lastname@example.org.