(Upbeat music)
- Scandinavian Airlines is the
oldest airline in the Nordics.
We roughly had 800 departures per day
flying roughly 30 million
passengers per year.
We saw an opportunity to use AI
and machine learning across the company
to drive efficiency and
improve customer experience.
When you run a loyalty
program within an airline
there is a lot of fraud
because our loyalty points,
is seen as valuable to people.
- So we need an AI that
can respond in real time
to block these accounts,
or block the transactions
and see the pattern.
- Azure Machine Learning
and its capability
to interpret data helped us
to increase the transparency.
In our models, we saw that
we could remove a lot of them
due to them not really being relevant.
In one of the models
we were basically close
to predict 100% certainty if a transaction
or activity was fraud and/or suspicious.
- For us, I would say
responsible AI means a lot.
We need to trust AI and if
we were to find ourselves
in a situation where the wrong
member is flagged and accused
that would be devastating.
- [Daniel] Responsible ML capabilities
with Azure Machine Learning
helps us understand the models
and make sure the models are
built to be accurate, fair, and secure.
- I think the explainability
features available here
certainly help us to see why a
member transaction, etc.,
is being flagged
and that creates that needed trust
and bond between our staff and AI.
- Our developers have a lot of freedom
within Azure Machine Learning.
It's very easy for data teams
to use the no-code drag-and-drop designer
before moving the projects
to other environments.
- Azure Machine Learning and MLOps
has helped us improve efficiencies
through automatic
retraining of our models.
It has given us the ability
to do controlled roll-outs.
Continuous integration,
continuous developments.
- We use Azure Machine Learning
to solve real business problems
without worrying about scale.
We can trust our data and models
and focus directly on gaining
value from the technology.
(Upbeat music)