Systems Of Record, Systems Of Action And The Parts Of The Enterprise That Still Don't Have A System

Direct Source Verification: This story is aggregated from Forbes (forbes.com). Full reporting rights and copyright belong to the primary publisher.
Few systems fully achieve a single source of truth, but the ones that come closest become the foundation the rest of the business is built on.

Kunal Chopra is the CEO of Certivo, an AI-powered compliance system of record for global manufacturers.

getty​I’ve spent most of my career building enterprise software, and the pattern I keep returning to is simple: A business function matures when it gets a system. Before that, it runs on people, spreadsheets and inboxes. After that, it can be measured, scaled and audited. AI is now changing what a system can do, which makes it worth asking which functions still don’t have one.

A system brings that knowledge into one place everyone can rely on. It defines what data counts, who owns it and what happens next, so a company can run the same process a thousand times without depending on the one person who remembers how it’s done. Few systems fully achieve a single source of truth, but the ones that come closest become the foundation the rest of the business is built on.

Geoffrey Moore popularized the term “systems of record” in a 2011 white paper for AIIM, describing the transactional systems companies had relied on for decades. A system of record is the authoritative source of truth for a domain. Other teams trust it because it agrees with itself.​

CRM holds customers, contacts and opportunities, while ERP holds orders, inventory and financials. SCM manages suppliers, purchase orders and logistics, and PLM tracks parts, bills of materials and engineering changes. Each one made its function legible to the rest of the company, which is why a CFO can close the books and a sales leader can forecast with confidence.

Systems of record store state, and people have done the work around them: reading documents, reconciling records, deciding what to do and entering the result. AI is moving that work into the system itself. A system of action uses the system of record as its foundation, reads incoming information, interprets it against rules and context and takes the next step: updating the record, requesting a missing document or routing an exception to the right person. Customer support teams already use agents that resolve routine tickets end to end, and finance teams use them for invoice matching and reconciliation.​

The system of record becomes more valuable in this model, because every action depends on the quality of the data and rules underneath it. Leaders evaluating a system of action should ask which actions it takes on its own and which go to a person for approval, how each action traces back to its source and logic and how it handles cases where its confidence is low. In regulated work, those answers need to be reproducible and auditable.

​Compliance is the clearest example I know. A manufacturer has to show that its products meet requirements such as REACH, RoHS and PFAS reporting rules. The evidence for that lives everywhere. Supplier declarations arrive as email attachments. Safety data sheets sit in shared drives. A question about a substance gets answered in a Teams thread, and a supplier’s change notice shows up in Slack. Part data lives in PLM, supplier records in ERP and SCM and customer requests in CRM.​

When a customer asks whether a product contains a restricted substance, someone pulls the answer together by hand from all of those places. PwC’s 2025 Global Compliance Survey found that 85% of respondents say compliance requirements have grown more complex over the past three years, so that manual assembly is getting harder every year.​

Compliance went without a system for three reasons. It cuts across every other function, so no single system was a natural home. Teams treated it as a periodic project, which made spreadsheets seem good enough. Its data was also too unstructured for traditional software to handle. In an earlier article, I wrote about how AI-native software can work with data in whatever format it arrives. That removes the third barrier, and it means compliance can build its system of record and its system of action at the same time.

Compliance isn’t the only function in this position. Ashu Garg of Foundation Capital has argued that systems of record capture outcomes but not the reasoning behind them. A VP approves a discount in a Slack message, and the CRM records the final price without the approval or the rationale. The same is true of a waiver for a nonconforming part, an extension on supplier payment terms or an exception granted to a key customer.​

Those decisions shape how a company operates, and the precedent they set is often more valuable than the transaction itself. Today that context lives in chat threads, meeting notes and people’s memories. As AI agents take on more work, the absence of a record of why decisions were made becomes a practical constraint, because an agent can’t follow a precedent it can’t see.

The functions that lack a system tend to share three traits: the work crosses departments, the data lives in conversations and documents and the process depends on judgment. Those were the conditions that kept traditional software out. AI is making them workable.​

Where else in your enterprise is important work still happening without a system?​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

Original Source
https://www.forbes.com/councils/forbestechcouncil/2026/11/20/systems-of-record-systems-of-action-and-the-parts-of-the-enterprise-that-still-dont-have-a-system/
Visit Forbes ↗
SHARE STORY:
𝕏 f in

Related Coverage in Culture