Today we’re launching Delta, an AI engine that will serve all of our products.
As companies integrate centralised data repositories, larger amounts of information are available for analysis. What if we could allow our customers to leverage this knowledge, in a way that’s simple and safe? Delta is our answer for all data associated with our products.
Features
Delta is a natural language interface for our products: even non-technical users are empowered to ask questions freely, for as long as needed. By choosing to make it available exclusively for our products, we were able to tailor Delta to their needs - it includes, for example, inside knowledge on how their internals work. All of this efforts simply result in better performances: Delta can reason about relationships between data.
We also focused a lot on building security guardrails: Delta includes a policy-based system1 to make sure models can only access data they’ve been authorized to.
For the complete technical overview, see Delta’s dedicated page.
About the name
We like to waste a lot of time thinking about how to name stuff - if you’ve missed the previous episodes, see here and here.
Delta is a Greek letter universally used in science to indicate a difference between two quantities: we mainly chose it as a hint to one of the mechanisms involved in training a neural network - a continuous comparison between actual and predicted values, with the goal of minimizing their difference.
Why “mainly”? See here for a couple of additional reasons.
One more thing
As you may have guessed by now, Delta is just an enabler to make our products even better. This means there are a few announcements coming soon: make sure to subscribe to our newsletter, they’ll be worth the wait.
- Heavily borrowing from Connhex IAM↩