How AI ‘digital minds’ startup Delphi stopped drowning in user data and scaled up with Pinecone
1 min read
Summary
San Francisco-based start-up Delphi, named after the Ancient Greek oracle, has hit a data bottleneck with its digitised chatbot ‘Digital Minds’.
Each chatbot is created using an end-user’s writings, recordings and other media and is designed to interact and respond like the real person would.
However, every time more content is added, the system becomes more complex and difficult to manage in real time, posing a major issue for the firm.
To combat the problem, Delphi is using a managed vector database from Pinecone, which has provided a more secure, scalable and flexible solution.
The Digital Mind now enjoys its own namespace within the Pinecone database, ensuring privacy and narrow search parameters to improve performance.
The database is SOC 2 compliant, provides encryption and built-in namespace isolation and can delete a creator’s data with just an API call.
Pinecone’s database dynamically loads vectors when needed, and offloads them when not, making the service ideal for bursty, multi-tenant workloads like Delphi’s.