Question

A financial services company wants to build a customer-facing gen AI assistant that answers questions strictly from their internal policy PDFs and product documentation. Their small engineering team wants a managed way to connect a foundation model to these documents so responses are grounded in the company's own content, without building and maintaining their own retrieval and chunking infrastructure. Which Vertex AI capability best fits this need?

Accepted Answer

Use Vertex AI's retrieval-augmented generation (grounding with your own data) so the model retrieves relevant passages from the company's documents at query time

Answer

Fine-tune a foundation model on the policy PDFs so the document text is memorized in the model weights

Answer

Increase the model's temperature setting so it produces more accurate, document-specific answers

Answer

Deploy a larger model on TPUs to give it enough capacity to hold all the documents in memory