Agriculture is at a crossroads.
On one hand, you have innovation,
sustainability, progress.
On the other,
you have sticking to what we know...
...age-old tradition,
and risk being left behind.
At SEGES we made that choice
years ago...
...and we’ve devoted ourselves to it
ever since.
We want to make a difference
in agriculture.
As the leading agricultural knowledge and innovation centre
in Denmark...
...we want to help farmers
be more sustainable and innovative.
Combining data and AI,
we’ve found a new way to do it.
To help farmers
take the innovation road...
...and never look back.
SEGES Innovation specializes
in helping farmers and food manufacturers...
...to run their business
in a more sustainable way.
To do that, we use data,
huge amounts of data.
We’ve spent decades
on building massive data sets...
...on cows, dairy, beef, crops,
and a lot more.
To make sense of all this data,
we’ve put together a data science team.
They’re tasked
with understanding the data...
...and turning it into products and models
that we can use to support our partners.
For years,
SEGES had an on-premise legacy solution...
...that was hard to maintain
and hard to work with.
Using Azure Machine Learning...
...we built a MLOps platform that helps them
train, test and productionalize machine learning models...
...in the easiest possible way.
The MLOps platform simplifies the process
of using data...
...to develop and test
new machine learning models...
...and makes it easier to build them
into software products.
It also connects
to Microsoft Intelligent Data Platform...
...which allows us to connect
to Azure Data Lake Storage Gen 2...
...and services like Azure Synapse Analytics
and Azure Databricks.
Another really cool thing we did
as part of the collaboration...
...was to build a repeatable MLOps framework
that allows SEGES to get up and running...
...really quickly with new use cases
without having to reinvent the wheel each time.
This platform is allowing us to significantly optimize
the machine learning lifecycle.
Maintenance costs are down by more than 95%,
and the direct cost of running a model...
...have been reduced by more than 80%.
Before, all machine learning models
were retrained and deployed manually.
Now, we have completely automated
these processes...
...allowing us to automatically retrain, test
and deploy models as often as we like.
As a result, the average time between retraining
has gone from six months down to a single day...
...which is a massive benefit
for our users.
Most importantly, we're even better at helping our farmers
to be more sustainable and efficient.
We can now predict high-risk cows
and detect diseases early...
...up to 90% of the time.
Ultimately, this helps the farmers,
the food producers...
...and the organizations to transform our sector
and make it better...
...day by day.