I've had the privilege of leading a world class design team across research, product and experience design to create the first modern AI-powered pricing platform. We seamlessly integrated recent strategic acquisitions and new products, culminating in a cutting-edge pricing platform that serves users better than anything before.
Creating the first interactive AI-powered pricing platform for Airlines
AI Powered Pricing
Building the interface for airlines to maximise revenue
The Airline industry mostly runs on legacy software from the 70’s. FLYR developed an AI to power millions of base line pricing decisions that have to be made every day at an airline.
With AI taking care of 80% of all pricing decisions, humans can be way more granular in their analysis and decision making. As part of a team of six designers I was able to lead many impactful initiatives to improve the business of our customers and in the process I was able to add millions in additional revenue for our company and customers.
«Because […] I'm able to be more targeted to where these customers want to book, the proportion sold in the [most expensive] 4 booking classes doubled week-on-week […] and that trend is continuing
[This is happening even while] We've been struggling domestically with load factors in the last few months, which was hurting us in terms of corporate traffic - because we had to reduce fares across the board»
– Revenue Analyst commenting on my work
Design that drives results
250x more powerful
Engagement +3h/week
Interactive Graph
50% fewer clicks
New Price Strategies
Easier to manage, 250x more powerful
To enhance one of our most used features, Market Constraints, I worked closely with a small group of users over a few months. My research discovered a lack of flexibility and major inefficiencies in the current workflow as the main pain points.
Addressing these points, it turns out, would require only minor changes in back-end logic but a complete rewrite of the front-end. What we achieved by working closely together was nothing short of remarkable: A new design framework that enabled complex and/or statements to be created intuitively.
250x more flexible flight segmentation made our customer millions of dollars
What used to take 250 individually configured constraints could now be achieved with a single strategy that was easier to set up, read and manage. This increased flexibility allowed analysts to segment more granularly and immediately unlock additional revenue – even in a historical market slouch.
The new framework I designed to power complex conditional and/or statements has been so successful with users that it has been adopted across different company verticals and now powers 10+ strategy and segmentation products across the company.
Wait, there is even more efficiency to unlock!
With every additional strategy type we implemented we were able to invest more time and gain more experience about user behaviour to keep increasing the speed for users. Analysing hundreds of strategies that our users created certain patterns emerged and a bunch of different initiatives are currently under way to increase flexibility and efficiency by another expected 3-4x.
Seat Availability
Making legacy processes finally feel polished
We can help our customers make money, we can help them save money or we can do both at the same time. Airlines use many different tools to visualise various aspects of their network.
One of these external tools visualises sales and availability across different booking classes. Integrating this natively into our product sped up their work significantly, lowered their costs and helped them make better decisions by reducing the friction of having to use multiple tools.
Engagement increased by 3h per user per week
The development of this simple looking heat-map table required many intense customer interviews and wide exploration to come to life. I tested many different visualisation techniques in user interviews to see how quick they were to spot anomalies.
The reason we ended up going with a heat-map is that this testing revealed a crucial insight: They need to compare availability across three dimensions.
Class availability within one flight
Between classes across the network
Entire flights to the the rows above and below
Since this information is crucial to making decisions I also optimised the visualisation of the impact of the price changes they intend to make. There is an inherent link between current price available classes that gets intuitively visualised with opening and closing of classes.
Instead of visualising the new totals in red or green tones we chose to focus on the severity of change instead to help users identify areas easier where the price change might create odd results or jumps.
Fare Ladder Optimisation
Compare your current fare ladder (black) with our AI suggestion (blue) and the surrounding price points to make smart decisions (purple) that scale
Spot opportunities and correct them immediately with the interactive graph
This tool was built to support legacy operations that don’t yet use continuous pricing but rely on “fare ladders” that consist of 26 distinct price points split among the available classes.
These price points are extremely important and hard to plan and thus get changed rarely. Thanks for our AI we can display immediate feedback around which price points are popular and well positioned.
If an analyst spots new opportunities he also gets an immediate way to change the fare structure with the help of an interactive graph. Both typing and drag and drop are supported.
The compact view of the large graph above
Visualise demand compared to the optimal price structure to help your decisions
–– Work in progress section ––
AI Correction
All decisions are recorded through the proxy of price through a process we call “influencing”. Observing real users make these influences showed me that the process is not optimised for their actual reality which includes a lot of previewing of the changes and further adjusting to get it just right. Making previously mandatory steps optional and increasing the flexibility by removing some barriers made the process a lot more efficient.
At the same time I also improved the preview function since the current solution was not scaling well with the newly introduced features. these changes are not ground breaking but important maintenance for to enable easier feature additions in the future and remove one of the painpoints that became especially egregious on small laptop screens for one of our customers.