Clever Columns
Last updated
Last updated
Description: A way to do research across all accounts, crossing both public and private data, all through natural language.
Demo
How are people using it today: 70% of beta rox users are using CC for ICP scans & Research and 30% for pige gen prioritization.
Users are excited about the ability to ask questions across all their account data for more efficient research and pipeline generation.
The most innovative companies in other verticals are already moving towards agent-powered research using AI tools like ChatGPT and Harvey to conduct research with greater speed and visibility at scale
Clever Columns empowers users to discover more about their accounts to identify untapped opportunities, optimize targeting strategies, and drive better business outcomes.
Feature desiderata
Fast: Clever Columns returns results in seconds, even for complex queries across massive datasets
Transparent: Provides clear attribution of where each insight is coming from for full traceability
Data-type agnostic: Seamlessly combines structured and unstructured data from any source
Technical details
Unified data layer: indexes both structured and unstructured data across all sources
Query engine: uses LLMs to parse and semantically understand user queries. Employs novel query decomposition algorithms to break down complex queries into a series of atomic search and retrieval operations that are executed across Rox's system of record (SOR) and external web resources. Utilizes intelligent caching, indexing, and query optimization techniques to ensure rapid response times even for highly complex, multi-part queries.
Distributed computation layer: parallelizes queries for rapid results even on large customer lists.
What’s next?
Suggested Clever Columns: useful for users looking to do discover over their accounts
Multimodal Clever Columns: expanding beyond text, integer and boolean research ouputs.
