From our last article on travel data, we discussed where travel data has evolved from and where we are today. And so, what is next? Like other industries, there is a lot of interest and hype around machine learning and artificial intelligence possibilities. Those technologies will be very helpful in the decades to come. We are only scratching the surface on what those technologies will accomplish in the future. But how do we set ourselves up for success to leverage new platforms using artificial intelligence (AI) and machine learning (ML)?
Machine learning uses data to become smarter, to build statistical models to provide predictions based on the information provided. In travel, the data can be inaccurate and sometimes misleading. A hotel property for instance can have many different names even though it is 1 location, 1 brand, 1 hotel. Sometimes it is due to the acquisitions or rebranding, but often it is a result of different booking channels having different names for the same property. If we feed bad data into machine learning platforms, how likely are we to get helpful results or real insights? Is it possible that the output will be incorrect?
It would not be the machine’s fault for providing output that does not make sense. It is simply making an analysis based on information it has been provided. So, if an output of business travel information shows various fares and classes of service and insights on behavior based on machine learning, how do we know if it is correct? If a data scientist discussing the learnings indicates that according to the data, the majority of business travelers are taking one-way trips, do we believe it?
There is still a lot of work that needs to be done in order for the tools of the future to be reliable, and our understanding of the data to be set up for success.
When it comes to data, generally speaking, we want to learn something from it: to understand it better, gain insights on patterns that are identified, or to uncover a new discovery. It is super interesting to learn something new of your business that you may not have realized before. It feels like a win. But does it actually make a difference to the business or is it just something you didn’t know before?
Regardless of the industry whether it’s travel, healthcare, technology, manufacturing, etc. we tend to have an innate default to want more information. It doesn’t mean that it’s helpful or actually changes the business. It could be just curiosity. But that does have a cost on time and distracts us from what can be important.
For data, if the question is would you like to see this information, this report, a particular dashboard, then almost always the answer will be ‘yes’ even though we may not know why. We believe that more is better. But what really happens? If there is a material change in the data, what is the action taken? If there is none, what are we doing with that data? And sometimes it is taking information or data and manually transposing it into another system. We may do this because it’s the way it’s always been done. But it doesn’t have to be. There are other options.
And if we are building beautiful intelligence UIs, dashboards, widgets, etc. are we simply taking what we have done before and made it more appealing? Would it be more effective to have those data points take an action or feed into other systems?
Referring back to previous comments around setting ourselves up for success and also to the previous article about queues and the delays of travel data; we can do better. Newer technologies now exist which can more quickly react to event-driven data. Most of us may have heard of APIs (application programming interface) which enable programmers to build integrations between various disconnected systems. Those have evolved over time and we are improving the integration capability of many platforms.
Webhooks are one of the newer tools which allow data to become real time. Instead of polling and more or less continually asking an API “is there more data?”, “is there more data?”, webhooks actually push data to a recipient when an event happens like a payment being applied or a booking being made. It’s much more efficient and truly in real time.
Robotic Process Automation (RPA) is a rules-based technology which allows companies to automate repetitive tasks. These bots are automation wizards. They can take real time data and make meaningful actions instantly. This does more than just reduce costs. It creates greatly reduced processing time, removing human error, and actually, better morale. People are happier not manually moving data from one system to another. And having data automatically flow in real time from one system to another with a rules engine can remove the inaction that dashboards and reports unfortunately can have.
Real time reconciliation is an example. Rather than wait until the end of the month for information to hit a credit card, wouldn’t it be amazing to know charges and even authorizations as they happen in real time? This information could decrement budgets or flow into other components of the financial system to trigger other workflow items. Rules could be put in place to identify any red flags so that fraud and anomalies could be prevented.
With this real time automation, we are setting ourselves up for success for the future.