RSMUS.comMost healthcare finance teams have already had the AI conversation. Someone may have already demoed a tool or asked how it would hold up in an audit, but then the project quietly moved to next quarter. That pause is the gap between AI interest and AI adoption, and it usually has less to do with technology than with the data underneath it and the governance around it.
The pressure to move isn’t going away. Gartner reports that 59% of finance organizations now use AI, and IDC forecasts that by 2030, 45% of organizations will orchestrate AI agents across core business functions. For healthcare organizations juggling multiple entities, service lines and payers, the question is no longer whether AI belongs in finance. It’s whether you can adopt it in a way your auditors, compliance team and board will accept.
The gap isn’t ambition. It’s trust.
A Sage Intacct study released in July 2026 put a number on it: 71% of finance leaders said they would reject an AI tool that was 99% accurate if it could not explain its answers. More than half said they would pay more for AI that shows how an output was generated. In finance, almost right has always been wrong, and healthcare raises the stakes further. The same figures flow into cost reports, payer negotiations and board packets.
The cost of skipping explainability shows up as rework. The same study found finance professionals spend nearly 13 hours a week reconstructing, validating and defending AI outputs, with 49% of U.S. respondents spending 15 or more. Ungoverned AI doesn’t remove work from a finance team. It moves the work downstream into verification and hands it to your most senior people.
Data readiness comes before ai readiness
In most healthcare organizations, the finance data story is fragmented by design. Volumes and revenue detail live in the electronic medical record (EMRs). Labor lives in payroll and scheduling. The general ledger, often a legacy system, receives summarized journal entries and loses the dimensional detail that made them meaningful.
AI can’t explain a variance it can’t see. If important details such as entity, location, department, service line and payer aren’t captured consistently at the transaction level, no assistance, AI or otherwise, can tell you why supply cost per case moved at one clinic and not the other. This is why modernization to an AI-ready ERP is a prerequisite rather than an upgrade: a multidimensional general ledger gives AI a structure to reason over and gives your team reporting they can defend.
The practical move is not to wait for perfect data. It’s to fix the dimensional foundation for the workflows you intend to automate first.
Governance, ethics and cybersecurity: what to require from the platform
A healthcare finance system sits close to protected health information, so any AI evaluation has to answer four questions plainly: What data does the model see? Who is allowed to prompt it? Where do outputs go? Is the whole trail auditable?
This is where platform choice matters more than model capability. Sage Intacct, for example, labels the AI across its products so users know when they’re interacting with it and builds to global privacy and AI regulation including GDPR, CCPA and the EU AI Act. Its security approach draws on the NIST AI Risk Management Framework, the UK NCSC AI Cyber Security Code of Practice and OWASP AI/LLM security best practices, with encryption, anonymization, access management and continuous monitoring built in rather than added later.
When evaluating an ERP for your organization, ask your vendors for the same specificity. Make sure it is clear that security and data protection are at the core of the process and not an afterthought.
Snapshot into how ERPs, such as Sage Intacct, approach governance
The reason AI-native ERPs like Sage Intacct works well for governance-conscious finance teams is that its AI is scoped to assist and identify issues to be escalated, not to make decisions on their own. Some features within Sage Intacct that display these include:
- Finance intelligence agent: Ask questions in plain language across your financial data and see the reasoning behind each answer. It’s enabled by an administrator and controlled by user permissions, so access follows your existing security model.
- Outlier detection: Flags unusual journal entries in real time by comparing new entries against historical patterns and materiality thresholds within the closed system, giving the approver the power to make the final call.
- AP automation: Creates draft bills from emailed or uploaded documents, matches them to purchase orders, codes accounts and dimensions, and detects duplicates. Your team reviews and approves; the model improves from that feedback.
- Close automation: Tracks close tasks across entities, flags issues and shows where the close is actually stalling, which is useful when you’re consolidating a dozen entities on different rhythms.
In every case, the human checkpoint is preserved. As Sage CTO Aaron Harris framed it, AI in finance must be accurate, auditable and reliable in real workflows, not just impressive in a demo.
The change management piece most organizations underestimate
Finding the right tool is only the first step. The current obstacle for most organizations has become the overall shift in the finance role between preparing the numbers to owning control around it. Sage’s research found U.S. finance leaders now rank risk, governance and decision judgment as the most important skill for a finance hire, cited nearly twice as often as deep technical accounting.
That shift becomes crucial and should be explicitly outlined in the organization’s governance in any AI practices. Define the governance around what AI should or should not be a part of, who oversees reviewing flagged items and the expected time line for reviewing. Teams resist AI when it feels like a verdict; they adopt it when it feels like a reviewer that saves them an hour and never gets tired.
A practical exercise for building AI readiness in the next 90 days
|
The takeaway
Closing the AI readiness gap in healthcare finance isn’t a technology purchase. It’s a data foundation, a governance model and a team that understands its new job. Organizations find the most success when these foundational elements are mastered and presented effectively to their organization.
If your organization is struggling to turn AI interest into AI adoption, speak with an RSM professional to better understand your AI and data readiness and explore opportunities to move forward with confidence.
Connect with an RSM advisor to evaluate whether your finance systems, data structure and controls are ready for AI adoption.
Mark Smallwood
Mark is a Senior Associate in the Technology Consulting practice and a Certified Sage Intacct Implementation Consultant. With a background in accounting and data analytics, he works with not-for-profit, financial services, and healthcare organizations to streamline financial operations and support efficient system implementations from planning through go-live.
