7 min read
Use Case: How to Convince Customers to Use Chatbot Instead of Calls
Enhancing CX and cost-cutting in a booking agency – it wasn’t easy, but it was worth it. Originally published at Chatbots Life.
In 2018 the travel and tourism industry added a record $8.8 trillion to the world’s combined GDP (more than all other economic sectors). It was one of the first sectors to embrace online, and automation becomes a regular feature in this industry — baggage weigh-in, registration, notifications, etc. People don’t even have to contact travel staff anymore.
But despite the innumerable apps, web scraping tools, and aggregators, travelers still have to get through an exhaustive search to find a perfect shopping option. To meet the needs of the new-age tech-savvy travelers, travel companies continue to seek new ways to improve the customer journey and make travel more convenient. Chatbots are increasingly being utilized by the travel sector to recommend bookings, assist travelers through the trip, suggest places to visit, and collect feedback. Chatbots become go-to assistants and they are always ready to help — that is essential, especially if you’re alone in a foreign land.
If you are on the fence about whether you should invest in a chatbot, you should definitely read through this case study where a major booking service has used it to improve its customer experience. That was a hit or miss process and many lessons would be drawn from it.
Booking with a travel bot
A major booking service that sells railway and plane tickets had to reduce the number of calls, automate FAQ and keep costs down. Summer is a high season for this business and last year the company felt like telephony comes at a cost — service had a toll-free phone and a fair number of contact center agents.
In July ’18, Contact Rate (CR) was at 36% which means every third customer was trying to reach them on the phone. The reasons were quite conventional — people have been calling to hang out for ticket prices and to make a reservation. To offload agents and to bring omnichannel experience, it was decided to implement an intelligent chatbot, so that users could choose the most convenient way of using services — text or voice.
We did that and for some time we’ve been testing our chats and fixing bugs, but it was obvious that CR was not actually falling:
To lower CR we also took a toll-free phone number off the website’s homepage. It used to be quite noticeable and we thought this might be the reason our customers call us instead of trying out our chatbot. Right after that, the About Company page became one of the most frequently used. That means people who really needed to reach us on the phone, found the number eventually. No slumping sales, no breakdown in brand trust.
Still, later it became necessary to cut costs, and we just had to look for a real comprehensive solution. First of all, we implemented a new IVR line and put there all the callers asking to look and book a ticket for them. We told that we do not book up via phone anymore, said that there’s a handy website where they could book up without assistance. Unfortunately, we did not get the hoped-for impact. Out of 5% of callers who got to the new IVR line, 3% tried to reach an agent. Again.
We’ve taken further steps. Regardless of the reason, people were trying to reach us, first of all, they spent 2 minutes on hold where a tender voice convinced them to use our chat first. This method dropped off 26% of callers — some of them actually did use our chatbot, but some of them left for good, becoming a closed lost. That was a disaster.
Phone calls, February ’19:
To fix this we rearranged IVR and fully integrated CRM. Now we define the caller: for instance, if this is our client who’s just about to take a flight — they might have an urgent matter — and we pick up immediately. The same goes for corporate customers. But if there’s no such urgent occasion or in case it is an unknown number, the caller is redirected to chat.
To measure the quality of service we use CSAT (Customer Satisfaction Score) and SL (Service Level) methods. By renouncing online bookings we’ve cut processing time from 4,5 to 3 minutes. Plus, we react very quickly, when we get an urgent call. This adds to the 90% Service Level metric pretty well. That makes sense, for people are using chats to refer to less complex matters, and most commonly they are pleased with their experience.
CSAT metric, which shows how fast we solve our customers’ problems, is not as good though. But it has to do with telephony itself — a customer calls when they have a serious issue. And we are doing our best, but we can’t affect some things like fares, weather, and politics that can ruin our customers’ plans. We are setting their minds at ease and trying to come up with the solution, but in some cases, there’s just nothing we can do.
Chatbots cost 3–5 times less than telephony, and one operator can handle 3 to 16 (peak time) chats
All that trial and error thing has demonstrated that chatbots are really cool and they do work. They allow to scale up in the easiest and least costly way.
Today 22% of all queries are solved with the bot+agent alliance.
5% are solved with bot only — those are the simple queries — FAQ and available tickets to top destinations. Booking agency claims, it is crucial for them that agents’ stress level has dropped — they don’t have to pinball between urgent questions and simple requests anymore and feel pretty happy about their jobs. Besides, chatbot implementation got 25% of all their agents off — they were promoted and redeployed to other divisions.
Not all the queries are being sent to the chatbot presently. This channel is open mostly for unauthorized users, whose cases may be easily and automatically solved — setting a password, booking a ticket, etc. Things are a bit more complicated when it comes to airlines’ regulations and fares — the agency works with over 500 airlines and each of them has its own issues. It’s not that simple and sometimes only agents are able to help.
It was decided not to use voice automation for now, because there’s a wide range of questions callers may ask, and in this case, it’s just too complicated and expensive. But the agency is planning to automate cold calling and lead warming-up. They believe this is a good strategy for now, and they will take a gradual approach to elaborate their customer satisfaction strategy.
To perform strongly, a chatbot has to be well-thought-out
Chatbots offer multi-faceted advantages for any business. 24/7 availability, quick response, saved user history, personalization, etc. Chatbots provide a means of showing your customers that you care about them, and that makes them stick to your brand.
Sure, it takes time and passion to create a beneficial and handy chatbot. Both — business’ and customer’s values should be taken into account when designing it. Also, a few very important aspects — like understanding of the context, smart personality, and proactivity — should be kept in mind all the time when building a dialog. These experiences should be instantaneous, personal, relatable, adaptable, and multi-modal. We at Just AI know that pretty well by means of tests, pitfalls, and successful solutions.