User-governed AI will help CFOs take the lead
AI is already changing the shape of finance – from streamlining reconciliation to surfacing instant insights. But for many CFOs, the question isn't always about what AI can do, but how much control they'll have. After all, the work they do demands precision, auditability and clear lines of accountability.
In fact, for today's finance leaders, AI shouldn't replace human oversight – it should enhance it. And that's where a new chapter in enterprise tech is emerging - user-governed AI.
The shift toward user-governed AI in finance
Traditional finance automation often meant rigid workflows or opaque algorithms. Today, that's evolving. A new approach is emerging – one where AI is embedded directly into finance systems, but designed to be shaped and governed by the people who use it.
The focus is shifting from automation for its own sake to AI that can be tailored, trusted and held to the same standards finance teams already apply - accuracy, auditability and control.
Vendors like Oracle NetSuite are leaning into this model. Their latest features are built to sit inside everyday finance workflows and use existing system data – not as separate tools, but as embedded support. And critically, each one is being developed with user oversight and explainability in mind.
Below, we look at three of those capabilities – Prompt Studio, the SuiteScript GenAI API and Financial Exception Management – and how Oracle's infrastructure is helping finance leaders adopt AI with confidence.
Tailoring AI outputs with NetSuite Prompt Studio
One of the challenges with AI is ensuring its outputs meet your organisation's standards for tone, accuracy and context. NetSuite's Prompt Studio directly addresses this by allowing administrators to customise the behaviour of generative AI that produces text and recommendations in the system.
Essentially, Prompt Studio is a toolkit for prompt engineering within NetSuite. It lets system admins configure the format, tone and creativity level of AI-generated content to suit their business needs.
For example, a CFO might prefer that automated customer emails maintain a formal tone and include certain disclaimers, or that explanatory text in financial reports follows a specific style. With Prompt Studio, those preferences can be built into the AI's prompts and templates.
By refining AI prompts, finance teams ensure the outputs are consistent, compliant and on-brand. NetSuite's AI text generation feature (called Text Enhance) already helps users draft emails, collection letters, and report narratives based on company data. But Prompt Studio takes it a step further by giving human overseers the dial to adjust how those drafts are written.
In practice, an administrator could, for example, reduce the "creative" level of AI suggestions to make them more straightforward and factual for financial documents, or enforce that certain terminology is always used for compliance reasons. This kind of admin control over AI ensures that while generative AI speeds up mundane writing tasks, the finance team maintains control over the voice and correctness of the content. It's a safeguard against AI going "off script," effectively embedding corporate policy into the AI's outputs.
Building custom AI use cases with the SuiteScript GenAI API
Out-of-the-box features are great, but what if a company wants to go further and infuse AI into a unique process or build a bespoke assistant? For that task, NetSuite now offers a SuiteScript Generative AI API (GenAI API) that lets developers embed generative AI into custom workflows and applications. Rather than a closed system, NetSuite provides this API as a bridge between its platform and large language models (LLMs), all within a governed framework.
Using the GenAI API, customers can call AI services from scripts to do things like generate text, summarise data or even answer questions based on NetSuite records. This means an operations manager could, for example, create a script that scans open accounts receivable and drafts personalised reminder notes for overdue accounts. Or a CFO's team could build a custom "chatbot" that answers finance queries by pulling the latest ERP data. You could even build an AI agent that both flags an overdue invoice and checks if there's an open customer support ticket, then alerts the account manager with a combined insight – this kind of cross-functional AI workflow can be developed in-house using these APIs.
Crucially, the GenAI API is designed with governance and security in mind. NetSuite's generative AI functions are backed by the Oracle Cloud Infrastructure (OCI) Generative AI service. When a developer sends a prompt via the API, it's processed by an LLM running in Oracle's cloud, not a public service. Oracle has made clear that data passed to the OCI AI service "never leaves Oracle, nor is it used by third parties for model training."
In other words, your financial data and prompts are kept within a ring-fenced environment, protected by Oracle's enterprise-grade security. Finance teams can leverage powerful AI models like Cohere's Command or Oracle's own, but with the peace of mind that sensitive data - like financial figures or customer info in a prompt - isn't being exposed to outside systems.
All this occurs under the governance umbrella of the platform – with access permissions, data governance policies, and Oracle's security controls still firmly in place.
For CFOs, this means their team isn't locked into only the vendor's imagination. They can extend AI in directions that matter most to them. It's AI on the CFO's terms, deployed in a sandbox they control.
Proactive oversight with AI-Driven Financial Exception Management
NetSuite's Financial Exception Management exemplifies "human-in-the-loop" AI – combining machine efficiency with human judgment to strengthen oversight. The AI scans high volumes of transactions (journal entries, invoices, and orders), flagging anomalies that deviate from historical patterns.
For example, if an expense entry is unusually high, the AI might suggest reviewing it for a possible data entry error. Or, if a sales order includes an excessive discount, it could be flagged for approval before processing. By catching these outliers early, finance teams reduce the risk of financial misstatements and move closer to achieving a continuous close.
Critically, the AI doesn't override records or make corrections on its own – it acts as a vigilant assistant, leaving the final decision to finance professionals. Each flagged exception prompts a review, ensuring that context is considered and corrections are properly documented.
Finance teams in the AI driver's seat
The emergence of AI in finance is often portrayed as a choice between automation and control. But as we're seeing, it doesn't have to be a trade-off.
With the right design, AI can empower CFOs and finance managers rather than sideline them. The key is a focus on user governance - giving business users the tools to shape AI to their needs, and the transparency to always see what's happening under the hood.
For ANZ-based finance decision-makers evaluating cloud ERP and EPM solutions, this shift is especially pertinent. Many organisations in the region are accelerating digital finance transformation, but they face strict regulatory environments and high expectations for governance.
Embracing AI doesn't mean handing over the keys to a robot bookkeeper. It means equipping your finance team with smarter tools that they control. The payoff is substantial too - faster closes, more insightful analysis, and liberated staff time, all achieved in a way that upholds the integrity and security of your financial data.
Ready to explore this topic further?
Watch the on-demand webinar: AI outlook for finance - what every CFO needs to know
Annexa's Matthew Owens and Ananda Iyer join Oracle NetSuite's Matthew Vorpasso to explore what's possible with AI, what's next and how to prepare your systems, data and people for AI-driven finance.
Watch the webinar replay now: https://info.annexa.com.
