Artificial Intelligence in use today within various parts of the aviation industry
By Anthony Kioussis
President, Asset Insight
As most of my colleagues at Asset Insight will tell you, I’m not the brightest bulb on the company’s information technology chandelier. That being the case, I need things explained to me in layperson terms, offering me the ability (I hope) to explain how AI can enhance – and is already enhancing – planning and decision-making for business aviation managers.
Bear with me through the techno-speak that is about to follow and you’ll understand its importance shortly (and it might even allow you to communicate with a millennial).
AI utilizes programming that is able to combine vast amounts of data with high-speed computer processing and advanced algorithms (a set of instructions that are repeated in a sequence a specified number of times or until a condition is met) that are adjusted by the AI, allowing the software to learn automatically from patterns in the data and move closer to “reality” over time.
How has this helped you on a day-to-day basis? When you use Google, you’ll notice that your computer is offering you suggestions as you type. These suggestions are not random, but rather based on information that you, and others who have demonstrated similar interests, have found useful in the past.
Google is actually utilizing a search engine algorithm that learns what people want as they use it, and then attempts to “guess” what the users are seeking.Historically, the process used to establish, and allow for, data transference between computers has been accomplished through “REST APIs.” API stands for “Application Programming Interface.”
It is a set of rules that allow programs to talk to each other. The capability is created on a computer server to permit communication with other computers. REST, which stands for “Representational State Transfer,” is a set of rules that developers follow when they create their API.
One of these rules states that you should be able to obtain a specific piece of data when you link to a specific web address.
Aid in searching
What does all this mean in layperson terms? Let’s say you’re trying to find videos about Hawaii on YouTube. You open up YouTube, type “Hawaii” into the search area, hit enter, and you see a list of videos about Hawaii. A REST API works in a similar way.
You search for something, and you get a list of results back from the service you are communicating with. REST APIs have worked well. However, as people’s reliance on technology has increased, the load on computing power has as well – in some cases exponentially, hence the reason certain websites seem to “crawl” (if not crash) when too many queries attempt to utilize the REST API simultaneously.
Other issues are that the form and content of information returned to a user’s query is predefined by the responder, and any changes to the API structure by that responder (the data path to the information the user is seeking to follow) can lead to applications not working.
To provide a place where any application can go to obtain whatever data it needs (read: standardization), and to deliver what any requestor is seeking, as opposed to fitting a requestor’s query into an existing bucket, a data query language called GraphQL was developed that has been used by Facebook since 2012 and was released to the public in 2015.
GraphQL replaces REST and is a powerful new language for APIs. How does that translate to planning and decision-making advances for business aviation? Say that you’re managing a maintenance facility and the completion of a maintenance event is running ahead of schedule.
You could have your sales group search various databases for an aircraft requiring maintenance that could fi t within the potential new slot. Alternatively, through the use of AI, your company’s program management system could have already communicated the potential capacity increase to Asset Insight.
Rather than waiting for the facility’s personnel to ask for the information, Asset Insight’s AI capability, and implementation of GraphQL and Asset Insight’s proprietary technology would automatically investigate which aircraft had upcoming maintenance requirements that could fit the available timeframe (based on the facility’s capabilities), and recommend the facility consider pursuing specific aircraft serial numbers to create additional revenue.
Rather than researching the correct aircraft operators to contact and allowing time to narrow the decision-making window for the facility and the potential maintenance client, the facility’s sales team now has targeted prospects to pursue and more time is available for both sides to plan and decide, improving efficiency and (potentially) increasing revenue.
Planning production or utilization
Let’s follow this thought process a bit further. Suppose your company manufactures aircraft components utilized in various maintenance events, and the need for such parts is based on aircraft flight hours and/or cycles. You could plan your production on general industry utilization figures.
You could contact individual shops to learn what consumption levels they foresee for the next year. Or you could harness the power provided by AI’s ability to recognize the pattern of component use and allow it to recommend production level revisions based on changes in anticipated demand and your company’s inventory and existing production schedule.
Utilizing REST API technology, thousands of connections between the API and data sources would need to be actively managed to ensure the communication stability necessary to provide an accurate and consistent response.
However, by virtue of GraphQL’s stability, the new API language, in conjunction with Asset Insight’s proprietary technology, allows for data monitoring directly from the database itself, then permits the data source to communicate directly with the user. Updates to the database do not inhibit the connection and accurate information flow between the user and data source is assured.
Think this is futuristic? OEMs today have the ability to monitor the engines and systems on many newer production aircraft while they are airborne, allowing for parts to be prepositioned based on actual or anticipated component failure.
Airlines are logistically able to take advantage of such AI capabilities, and it won’t be long before business aviation finds ways to do so as well. What if you’re a frequent user of charter flights. You can use many capable systems to search and book a charter flight today.
But how much could your personal efficiency increase, and your travel expenses decrease, if your computer’s AI, knowing your schedule, continuously monitored charter deals available through countless sources and recommended bookings to you based on your needs.
Suppose the AI also told you that a better alternative existed, such as rail transportation – information you never requested and an option you never even considered.
AI in aircraft transactions
When it comes to aircraft transactions, entities taking advantage of this technology will benefit substantially at the expense of non-users.
For example, a large number of factors are employed by Asset Insight’s AI platform to derive current trends and long-term Residual Values ranging from the economic outlook and market momentum to the projected future cost for specific metals.
The AI expands and refines these factors and proactively models (quite literally) millions of calculations involving parameters that aircraft owners, buyers and sellers have never considered. The ability to think in such complex terms is only some of the value created by AI.
GraphQL assists the AI process by acting as something akin to a combination traffic flow monitor, data compiler and information desk. Rather than forwarding a user’s request to a source that can only respond with the information it has available, API systems utilize GraphQL to access data from countless sources to match the information request/flow patterns they detect, and they respond to a user’s query with answers that are not constrained by the information available through any single data bucket.
Also, as was mentioned earlier, changes made at the data source do not break the API connection between the query originator and the responder. Although it will take time for GraphQL to demonstrate its full value to the business aviation industry, there are many reasons why entities will gravitate to it.
Asset Insight is already applying GraphQL technology in the API available to our partners, and we see great potential in how it can help our clients and our partners’ customers in the years to come.
Anthony Kioussis is President of Asset Insight, which offers aircraft valuation and aviation consulting services. His 40+ years of experience in aviation includes GE Capital Corporate Aircraft Finance, Jet Aviation, and JSSI.