Dominik Asam discussed the company’s enterprise AI strategy, explaining why businesses are becoming more focused on AI cost:
There is now a recognition that AI tokens are not free. Every enterprise we talk to is grappling with the fact that token spending is going through the roof.
There is now a recognition that throwing tokens, in a probabilistic way, at every problem might not be the most efficient approach.
There is also a fear of vendor lock-in, as certain models are starting to increase their prices.
That is why SAP has chosen a completely different architecture. I checked this morning, and we now have 58 large language models integrated into our AI platform.
This allows us to take advantage of competition across models, both in terms of performance and cost. We can also route each task to the most appropriate model.
Sometimes you do not need the most expensive model.
Even more importantly, sometimes you do not need a model at all. You can solve the problem in a deterministic, algorithmic, and easily auditable way by simply crunching the numbers.
At SAP, we have the ability to leverage all of these approaches.
https://vm.tiktok.com/ZN8JPBNvs/