Enterprise Knowledge Search · Risk Management
LexisNexis needed an AI retrieval architecture that could preserve relationships and time context across fast-changing financial and commodities data, because standard RAG could not produce reliable answers for high-stakes decisions.
Reported impact
LexisNexis eliminated synchronization tax between disconnected AI stores, improved answer accuracy for time-sensitive market intelligence, and created a scalable platform that can reason over large, changing financial datasets.
LexisNexis Risk Solutions used MongoDB Atlas as the single store of truth for a time-aware reasoning platform built by its Maverick team. Instead of relying on standard RAG pipelines that treat facts as isolated text chunks, the team modeled nodes, relationships, vectors, and ontology as native MongoDB collections so the system could scale beyond the limits of out-of-the-box graph tooling. Every fact carries explicit start and end dates, allowing the database to pre-filter results into chronologically correct timelines before anything reaches the model. That architecture preserves ownership changes, temporal relationships, and other structural context that ordinary semantic search misses. Atlas therefore serves not only as persistent storage for the graph-shaped knowledge base, but also as the mechanism that grounds answers in history-aware evidence. By eliminating synchronization between separate graph, vector, and application layers, LexisNexis created a more accurate AI system for financial and commodities intelligence.
Articles, commodities facts, nodes, relationships, vectors, and ontology data are stored as MongoDB collections with time bounds; queries filter those collections into chronologically valid timelines before grounded results are handed to downstream reasoning and answer-generation steps.
Rust ran 14 times faster than Python in their performance tests.
“Rather than relying on the LLM to perform maths across conflicting dates, the database delivers a pre-filtered, chronologically accurate timeline.”
“Typical RAG implementation, at the end of the day, it's really just semantics. And the problem with semantics is, it doesn't understand the relationship of data.”
“We set out to say: can I make the nodes a collection? Can I make the relationships a collection? Can I make the vectors a collection? Can I have the ontology as a collection?”
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
OpenAI Sample May 6, 2026 Chatham Financial Chatham needed to reduce the routine execution work around capital-markets advisory, trade validation, and internal software delivery while keeping accuracy, auditability, and human judgment central to the workflow. Employees build applications with GPT‑5.6 Terra by default and GPT‑5.6 Sol as an option ChatFIN · Chatham Onyx · Chatham Vibes 
Microsoft Sample Jul 18, 2024 Lionbridge Realizing the need for faster and more agile localization services Founded in 1996, Lionbridge combines the best of human and machine intelligence to support large companies around the world with content creation,… "We already have more than 500 customers using AI to help with content optimization, and we're getting that… Azure AI Foundry · Azure AI Translator · Azure OpenAI Service 
MongoDB Sample Jul 8, 2026 Mercor Mercor needed to replace a join-heavy recruiting architecture that made profiles slow to load and delayed newly interviewed candidates from appearing in search results for hours. Mercor reached a 100% fill rate on time-sensitive client requests as a direct result of this change. Apache Kafka · Fireworks AI · MongoDB Atlas 