Delivering successful AI
AI can be a game-changer for organizations, but its success relies on careful implementation and, often, trial and error. Because it’s been talked about for decades, it’s easy to forget that the technology is still new. Getting it right can bring great rewards, getting it wrong can be costly and time-consuming.
At Satair, we have worked to develop Lilly, an automated quote handling platform for our e-commerce site. This has been successful in several ways: improving the customer journey through fast turnaround of quotes, releasing our own talent from low-value work and reducing costs.
As margins for the aviation industry continue to be squeezed, successful innovation such as this will be more and more of a necessity for survival. But where do you start?
Solving problems
At Satair, our call volume from customers was greatly increased by the pandemic. This was largely because stock inventory changed as travel decreased; moving to ordering more frequent, lower volume stock supplies. This meant we had a huge increase in the number of calls coming through to customer service.
We started with this problem and went directly to the source – our customer service representatives – to gather information and understand the impact on the ground. We made these employees active members of our development team. They supported the project through their knowledge and insights initially and later through trial-and-error phases.
This contributed hugely to the success of the project. The better the understanding of the problem and quality of the data you put into an AI project, the better it will work.
Technology-agnostic approach
Starting with the problem also means removing any pre-conceived ideas you have about the technology you will use.
If you build the technology stack from the basis of the problem using an agile methodology, then you’re able to identify issues as they appear, iterate, resolve them and move forwards. This also means you don’t get locked into a costly bespoke technology solution that cannot be managed and maintained in-house.
Fail fast
Short development cycles and daring to go live with Minimum Viable Products means there’s no wasted time and you find out what’s most useful quickly. If your team on the ground – in our case, our customer services representatives – are on board, they can test products for you and start giving you further feedback quickly.
This means there are no late discoveries of issues, onboarding new solutions is less fraught with potential failure and delivery is swift.
Keep proving your worth
Starting small and proving you can solve a problem means that asking for further investment to expand has more chance for success. We built our initial solution in six weeks and had results we could demonstrate before we thought about expanding.
At Satair we are developing an AI operations platform and moving away from a project-based development. Once innovation is established as the foundation to evolve the business, there will be further capacity and funding through freeing up time for revolutionary projects to take place.
Why is innovation vital
The aviation industry was dramatically impacted by the recent pandemic, but this is only a precursor to a far greater change that the industry must navigate around the impact of climate change and the changing face of air travel.
In the future, aviation must lead the way in early adoption of tech to pave the way to a better future – for the industry, our customers and the planet – learning from innovative projects such as this one will help us deliver large-scale success for re-building the foundation of our business.