Key takeaways
- A system of record is chosen per type of data, not for the whole company.
- The CRM is usually the system of record for contacts, companies and deals.
- When two tools both own a field, nobody knows which value to trust.
- Assistants that update records should write to the system of record, not start a parallel list.
What does system of record mean
System of record means the place where a piece of data is created, corrected and owned. Everyone else treats it as the answer. The abbreviation SOR turns up mostly in operations and engineering conversations.
Companies usually have several, each owning its own type of data. The CRM owns contacts, companies and deals; the billing system owns invoices and subscriptions; the HR system owns employees. Trouble starts at the edges, where a field such as contract value could plausibly live in two of them.
Is the CRM the source of truth for customer data
Usually yes for who the customer is and where each deal stands, but not for everything about them. What a customer actually paid lives in billing, and how much they use the product lives in the product's own database.
People use source of truth and system of record as if they were the same, and in daily talk that is fine. Strictly, a system of record is where data is entered and owned, while a source of truth can be a view assembled from several systems, such as a reporting warehouse that combines CRM and billing data.
Why should one tool own each record
One tool should own each record because two owners mean two answers. When a seller updates a job title in the CRM and an old spreadsheet still has the previous one, the next campaign can go out with the wrong title and nobody can say which was right.
- Duplicates multiply when each tool creates its own copy of a contact.
- Reports disagree, and meetings turn into arguments about whose number is correct.
- Corrections get lost when they are made in a copy instead of the original.
- Access rules break down, because sensitive fields leak into tools that were never meant to hold them.
How do AI assistants respect the system of record
An assistant respects the system of record by reading from it and writing back to it, instead of keeping its own version of the customer. If a call note changes a deal stage, the change belongs on the deal in the CRM, not only in a chat history.
Good practice is to have the assistant look up the existing record before creating anything, so it updates the contact that is already there rather than adding a second one. A person should be able to see and approve what changes, especially while the team is still learning to trust the assistant.
How do you choose a CRM system of record
Choose the tool where your team already does the work that creates the data. If sellers live in the CRM, make it the system of record for contacts and deals and point every other tool at it.
Write the decision down per type of data and share it. A one-page table of data type, owning tool and owning person settles most arguments before they start.
Where Pebbles Ai fits
- FoundationIntegrationsIf Attio or Pipedrive is your system of record, the assistants work on it through MCP: they read records and create or update them from chat, each write an approval action under the execution mode. There is no automatic or two-way sync.
- CapabilityProspecting & customer recordsTeams without a CRM can use the Pebbles Ai CRM (in Beta) for contacts, segments and deals, where leads from Find fresh leads arrive as contacts.
Common questions
What is an SOR?
SOR is short for system of record: the tool a company treats as the authoritative home of one kind of data, such as the CRM for customer records.
Can a company have more than one system of record?
Yes, one for each type of data. The rule is only that no single type of data, or single field, has two owners.
Is a data warehouse a system of record?
Usually not. A warehouse copies data from the systems of record for reporting, so corrections should still be made at the source, or the next copy overwrites them.
About this page. Written by the Pebbles Ai team for people who run sales, marketing and growth. Last updated .