If an organization doesn’t input customer and vendor data in a consistent way across the enterprise, a whole host of finance-related issues can arise—including the risk of inaccurately recognizing revenue or even making duplicate payments.
John Burns, senior director, financial systems and controls at Summit BHC, a Franklin, Tennessee-based company that operates a network of behavioral and mental health centers throughout the country, discusses the dangers and shares how he handles enterprise data at his organization.
You’ve argued that companies rigorously control physical spend but leave enterprise data largely ungoverned. Where do CFOs have a false sense of control today, and what risks are they missing?
Finance teams tend to have a clear handle on how money moves through the organization. There are defined approvals, visibility into spend and strong audit practices around transactions.
Where things get less consistent is with the data behind those transactions. In many environments, large volumes of data can be created or modified with limited oversight compared to the controls placed on financial approvals.
That gap shows up over time. Reporting requires more reconciliation, payment issues surface and teams spend cycles correcting data rather than using it. It is less about a lack of systems and more about how consistently data is governed as it enters and moves through the process.
Can you offer an example?
A simple example is vendor master data. A new vendor gets created quickly so a payment can go out, but key fields are inconsistent or incomplete. The legal name does not match tax records, payment terms are entered differently across business units, or duplicate vendors are created because there is no clean search or approval step. Over time, you end up paying the same vendor under multiple IDs or misapplying payments, which creates reconciliation issues for both AP and the vendor.
The same pattern can show up in customer data. Sales or operations teams may create new customer records to move quickly, but naming conventions, hierarchies and credit terms are not standardized. That affects revenue recognition, credit exposure and forecasting accuracy. Finance ends up spending time reconciling what should be a single customer across multiple records.
None of these changes are complex on their own, but the issue can be the volume and lack of control. These small inconsistencies compound quickly when they are repeated across thousands of records.
In my role, I lead how we put structure around how core data is created and changed, whether that’s vendor records, customer hierarchies or chart of accounts updates, so those changes are reviewed with the same discipline as financial transactions.
That includes defined ownership in the business, approval workflows for mass updates and controls that prevent duplicate or inconsistent entries from reaching production. When that structure is in place, the downstream impact includes cleaner reporting, faster close cycles and fewer surprises that require manual correction.
In finance systems, who should actually “own” core data like customers, vendors and chart of accounts and what breaks when that ownership is unclear?
In my experience, the most stable environments are the ones where ownership sits clearly with the business functions tied to the outcome. Customer data aligns with the revenue organization, vendor data with procurement, and the chart of accounts with finance.
IT plays an important role in maintaining the platform, but it should not be responsible for defining or validating the data itself. When ownership becomes unclear, data tends to evolve based on immediate needs rather than long-term structure.
Over time, that leads to duplication, inconsistencies and more effort required to produce reliable reporting. The impact is usually felt in longer close cycles and increased reliance on manual workarounds.
You’ve seen automation initiatives backfire. How do CFOs distinguish between processes that should be automated and processes that should be eliminated entirely?
Automation is most effective when applied to processes that are already well understood and tied to a clear outcome. In practice, some processes continue simply because they have always existed, not because they add value.
Taking a step back to evaluate whether a process is necessary often creates more impact than automating it. If a workflow does not support a decision, control or required output, it is worth reconsidering whether it should remain in place.
Once that clarity exists, automation becomes more targeted and tends to deliver more meaningful improvements in both efficiency and control.
As AI becomes embedded in finance, how should CFOs think about governance differently to ensure accuracy, auditability and accountability at scale?
AI increases the importance of consistency in both data and process. The outputs are only as reliable as the inputs and the structure behind them.
That puts more emphasis on how data is created, maintained and approved across the organization, along with clear ownership at each stage. These are familiar principles in finance, but they need to extend more broadly as AI becomes part of the workflow.
From a governance standpoint, the expectations remain the same. Outputs need to be explainable, auditable and tied to accountable owners. The difference is that those standards now need to hold up at greater scale and speed.
