In our last EXLRT Insight we introduced the concept of how AI and Natural Language Generation can and will evolve into the new content creators. And this doesn't mean just any content; we're talking hyper-personalized, created instantly and within the context of each individual customer clicking their way through the travel booking process, content. As more websites harness the power of personalization, AI and NLG, what does this mean for traditional content generation methods?
We've shared our knowledge and views on AI, personalization, and machine translation throughout the year. In our last Insight we aimed to exemplify how all of these will come together to become the new means of instant, translatable and relatable content for every single customer. While this is an exciting prospect for any traveller getting tired of the same generic and irrelevant content, it also means that content generation organizations will have to switch gears and develop ways to master the new technology.
A content generation organization that doesn't actually "create" content? Yes, it is possible! Content generation organizations of the future will have to:
Train engines to deliver a good piece of content.
Define the input that is used to generate the content.
Gather data touch points that will differentiate the input that is used as the basis for the content.
Training Content Engines
Even AI technology needs training from its human counterparts to make sure it's producing the best content. At EXLRT we already have a trained NLG engine available that is generating hotel descriptions based on hotel amenities that we know are relevant to that specific traveller. This is done by feeding the engine tens of thousands of hotel descriptions from which it learns how to generate one on its own. Engines can also be trained to generate descriptions of rental cars or cabins on a cruise-liner.
Defining the Input
Content Generation Organizations will still have to be on top of what input the engines are picking up on. This means making sure to collect and utilize as many data touch points available. With hotel descriptions, this means collecting the data touch points that identify what kind of hotel amenities are relevant to the traveler on that specific trip. Is he or she travelling for business or pleasure, together with spouse and/or children? How do flight times coordinate with the hotel check in/check out? These are some data touch points that can be used to create hyper-personalized hotel descriptions with higher conversion rates.
External Data Touch Points
Useful information about a traveller comes from many different points of communication. Airlines will utilize internal booking and loyalty program information. But external data from Facebook, LinkedIn, Apps like Uber or Skyscanner, Booking.com, AirBNB, or Yelp can all be used to create a more contextually relevant experience for the online travel booker.
What about copyrights?
Now that the newest AI, NLG and machine translation software are really starting to turn traditional methods of content generation on its head, there are many things to think about. Who will own content that has been generated by a machine? Will “content” creation and copyrights cease to exist all together when it comes to online content production? This topic will be interesting to see develop as these technologies become the new content generators.
Content Generation: Gone But Not Forgotten
The world and its billions of online shoppers and travellers will never not need content. And though AI, NLG and MT engines will make the jobs of content creators easier when it comes to tedious details and constant planning, researching and producing for countless markets and buyer personas…it is understandable to feel a sort of nostalgia for what could be the loss of human powered content generation. Though the job descriptions of human content generators may change, the role will never go away completely. The future is exciting!