Arthur Mensch

And I'd say our strength has been to also work with European companies that are a bit lagging behind and that wants to adopt the technology to leap forward.

And I'd say our strength has been to also work with European companies that are a bit lagging behind and that wants to adopt the technology to leap forward.

And we've been able to do that through a forward deployment engineering engagement for our Forge product, for our Studio product that allows to deploy agents that do end-to-end automation.

And we've been able to do that through a forward deployment engineering engagement for our Forge product, for our Studio product that allows to deploy agents that do end-to-end automation.

But on top of that, the thing that we have announced today, like Forge, is something that is actually being used today with customers in the US because they come to us with needs for post-training, for making models specifically good at financial services.

But on top of that, the thing that we have announced today, like Forge, is something that is actually being used today with customers in the US because they come to us with needs for post-training, for making models specifically good at financial services.

And what's happening is that we have this product and we can bring the models to specialize them as well.

And what's happening is that we have this product and we can bring the models to specialize them as well.

Well, you need general purpose models to do the orchestration parts, etc.

Well, you need general purpose models to do the orchestration parts, etc.

But at some point, enterprises sit on a lot of intellectual property, on a lot of signals coming from physical systems, from factories, from tools.

But at some point, enterprises sit on a lot of intellectual property, on a lot of signals coming from physical systems, from factories, from tools.

And it's actually not trivial to connect those systems, to connect those data to models that are closed source.

And it's actually not trivial to connect those systems, to connect those data to models that are closed source.

If you have open models, you can actually add new parameters you can make.

If you have open models, you can actually add new parameters you can make.

a lot of deeper things that you cannot do with closed models.

a lot of deeper things that you cannot do with closed models.

We can also, and that's something that we do, not only do we work at the model side, but also at the orchestration side, we sit with subject matter experts to understand their needs and we build business applications that are fully bespoke to their needs by modifying the models, but also modifying the harness on top, etc.

We can also, and that's something that we do, not only do we work at the model side, but also at the orchestration side, we sit with subject matter experts to understand their needs and we build business applications that are fully bespoke to their needs by modifying the models, but also modifying the harness on top, etc.