AI is taking the world by storm and bringing a whole new world of vulnerabilities, but the bulk of breaches are still around classic security problems and under-protected data. Our most recent work is around protecting AI data, but we also continue to improve and push forward on our broader platform to deliver on our vision of a more secure cloud.
And on the topic of evaluating security, we've dusted off our own SaaS Security Maturity checklist and are using it to evaluate the security of vector databases. We've done two so far, Pinecone and Qdrant, and the scores and some of the misleading security claims were eye opening.
I'll be at RSA in May and speaking at RMISC in June. Let me know if you'll be at either of those and we can connect up.
Patrick Walsh CEO, IronCore
5 Things SaaS Companies Get Wrong with BYOK
BYOK or Bring Your Own Encryption gets interpreted in many different ways and delivered to customers many more ways yet. There are five ways that software providers often get it wrong when delivering BYOK to their customers.
Claiming end-to-end encryption is becoming commonplace; fortunately testing claims can be easy. In this post, we show you how you can validate or disprove such claims and we use our own SaaS Shield as an example
We've added a number of integration examples showing how to use Cloaked AI with popular vector database options including in hybrid search scenarios. We now have examples for Pinecone, Elasticsearch, OpenSearch, Milvus, Marqo, pgvector, Qdrant, and Weaviate.
