To identify duplicates, Kernel performs a comprehensive analysis that reviews an account's complete identity as it exists within your CRM. This includes analyzing the account name, associated notes, and opportunity data. This internal context is combined with an analysis of the company's website content and a visual comparison against all other potential duplicate candidates in your CRM. The goal is to understand the relationship between accounts as a sales representative would.

Kernel categorizes duplicates into different types to provide you with the flexibility to define what a "duplicate" means for your business. This allows for a more nuanced approach than a simple "match" or "no match" system. For example, you can configure rules to automatically merge Exact duplicates while excluding Regional duplicates, allowing you to keep international offices as separate accounts if desired.

Kernel automatically maps complex corporate family trees using a structured classification system that defines each account's relationship within the hierarchy. The system categorizes accounts as either PARENT entities (not owned by others) or CHILD entities (owned by another company), with important subtypes: OPERATING vs. HOLDING companies for parents, and STANDALONE, ABSORBED, or REGIONAL subsidiaries for children.

This intelligent classification captures real business relationships—distinguishing between Berkshire Hathaway (holding company parent) and Kraft-Heinz (operating parent), or between Jell-O (standalone subsidiary) and Heinz (absorbed subsidiary).

Kernel builds the complete corporate hierarchy externally first, providing a flat file for review before making any CRM changes. Upon approval, the system adds missing parent companies. Kernel implements the hierarchy by populating Salesforce's standard Parent Account field with permanent 18-digit Account IDs. Hierarchy mapping creates clean, reliable corporate structures that reflect how businesses actually operate and make purchasing decisions.