Why AI Agents Need Better Retrieval (Not Just Bigger Models) | Voyage AI | .local NYC 2026
Subscribe to the MongoDB for Developers YouTube Channel: https://www.youtube.com/@MongoDBDevelopers?sub_confirmation=1 Learn more about Voyage AI: https://mdb.link/_og6Y919lnE-voyage Learn more about Voyage AI models: https://mdb.link/_og6Y919lnE-models Better retrieval isn't just about finding results — it's about finding the right ones, in the right order, the first time. MongoDB's Voyage AI embedding models rank #1 on the independent Retrieval Embedding Benchmark, ahead of Google, Cohere, and NVIDIA, and native reranking can boost retrieval quality by up to 30% out of the box. For teams already building with Voyage and those using it independently for the first time, this session covers what's new — including the latest reranking models with real gains for code and long-context workloads. Chapters 00:00 Introduction: retrieval and production AI 00:51 Enterprise AI adoption vs. business value 02:06 Why proprietary data is the moat 03:06 Giving agents the right context 03:34 Search, vector, hybrid, and graph retrieval 04:14 Agent memory is a retrieval problem 05:02 Account-manager memory example 05:57 Embeddings and semantic recall 07:04 Reranking memories for the task 08:08 Why retrieval errors compound in agents 09:32 Voyage AI’s retrieval model approach 09:52 Embeddings: recall vs. precision 11:25 Voyage embedding and multimodal models 12:53 Rerank 3: quality, latency, and instructions 14:40 Earnings-call retrieval demo 15:19 Comparing three retrieval strategies 16:22 Benchmark results with reranking 17:47 Three analyst questions, two retrieval paths 20:43 When agentic search adds cost without improving results 21:02 Coding-agent retrieval demo setup 21:40 Troubleshooting the live demo 22:44 Intended coding-agent comparison 25:16 Main takeaway: make proprietary context useful 26:31 MongoDB + Voyage AI retrieval stack 27:09 Why agent state fits the document model 28:09 Closing Visit Mongodb.com → https://mdb.link/MongoDB Read the MongoDB Blog → https://mdb.link/Blog MongoDB for Developers YouTube Channel → https://www.youtube.com/@MongoDBDevelopers