Graphlet Ai, Llc

Autonomous agents feed on knowledge graphs. To build agents to automate business processes you need more than fuzzy vector RAG or even GraphRAG to avoid costly mistakes. Agents need facts: a concrete model of your business in the form of a knowledge graph it can reason over using text-to-query to find clear answers to complex questions. A graph representation of your business domain built from your structured (databases) and unstructured data (documents) with PySpark, GraphFrames, BAML and Gemini enables agents to reason about real-world problems. We developed semantic entity resolution to use text embeddings to more accurately block data into groups for comparison and Large Language Models (LLMs) with BAML to schema align, match and merge duplicate records in entire blocks of entities, from different sources, in a single operation. We build knowledge graphs ranging from thousands to billions of nodes. We speak both small and big data, ensuring that today’s pilot project can become tomorrow’s enterprise knowledge graph without a costly rebuild. We proudly use and contribute to Spark GraphFrames to precompute answers whenever possible, minimizing your graph database bill. Principal Consultant Russell Jurney has more than 20 years of AI, knowledge graph and big data experience. Contact rjurney@graphlet.ai for a proof-of-concept proposal.