AI has moved from boardroom discussion to daily reality for most finance teams. The question now is no longer whether to use it, but how to use it well – with the right data foundations, the right controls and inside the systems where the work actually happens.
To get a clearer picture of where finance teams sit on that journey, we surveyed more than 250 finance professionals during our recent AI in Finance webinar. The findings paint a picture of widespread experimentation, growing appetite for action and a sizeable gap between where AI is being used and where it could deliver the most value.
TL;DR
- 81% of finance teams use AI alongside their systems rather than inside them
- 70% move data out of their core systems to use AI tools like ChatGPT
- Only 16% feel very confident their data is ready for AI
- 54% want AI with human approval, while 35% are open to autonomous action
- 47% say reporting and insights is where AI goes next
- The teams pulling ahead are those embedding AI inside the platforms where transactions, controls and reporting already live
Here’s what the data tells us and how it sits alongside what’s happening across the wider industry right now.
AI is mostly sitting outside core finance systems
When we asked how AI is being used today, the answer was striking. 81% of finance teams said AI is used alongside their systems – through tools like ChatGPT and Copilot – rather than embedded inside them. Only 11% report partial workflow integration and just 1% describe full integration across workflows.
That picture is reinforced by another finding: 70% of finance teams have moved data out of their systems to use AI, with 42% doing so regularly. Just 8% have AI embedded directly within their workflows.
This is consistent with what’s happening globally. The CFO Connect State of AI in Finance 2026 report found that 45% of finance teams remain in limited pilot mode, with only 17% actively using AI in core workflows. ChatGPT remains the dominant entry point, used by 35% of finance teams worldwide.
For finance leaders, the implication is significant – most AI activity in their organisation is currently happening outside any governed, auditable framework. Data is leaving the system of record, being processed by external tools and returning as insight – often with no formal policy guiding which tools are approved, what data can be shared or what happens to that data afterwards.
AI usage is widespread, workflow adoption is still limited
63% of finance teams are already using AI in some form, yet fewer than 5% are applying it effectively within operational workflows. Half are using tools like ChatGPT alongside their systems, 17% are exploring how to connect AI into their systems and 13% are relying on AI features built into their existing platforms.
This mirrors a broader pattern across the profession. KPMG’s 2026 Global AI in Finance: The Decision Advantage survey, released in May 2026, found that active use of AI in finance has more than doubled since 2024 – rising from 30% to 75%. And 71% of finance leaders say AI is meeting or exceeding their ROI expectations, with the strongest returns appearing in decision-making quality, forecast accuracy and responsiveness.
The opportunity here is significant. The finance teams moving ahead fastest are those bringing AI inside the platforms where transactions, controls and reporting already live – turning it into part of how the function operates day to day.
Data confidence is still a work in progress
Only 16% of finance teams feel very confident their data is ready for AI-driven tools. 52% are somewhat confident and 32% are not confident at all.
This aligns with the wider industry view. KPMG found that 36% of organisations cite data quality as both their biggest barrier and their biggest opportunity, while Grant Thornton’s 2026 AI Impact Survey revealed that 78% of business executives lack strong confidence they could pass an independent AI governance audit within 90 days.
The interesting tension in our data is that adoption is happening regardless. Teams are pressing ahead while the data groundwork is still being laid – which makes choosing the right foundation even more important.
Finance teams want AI with guardrails – and the appetite for action is stronger than expected
When asked what level of control they’d be comfortable giving AI in finance workflows, 54% favoured a recommendation-and-approval model, mirroring how finance controls already work. Another 35% are already comfortable with AI taking some form of autonomous action, either for low-risk tasks or within defined controls. Only 11% want AI limited to insights with no actions at all.
It’s clear the appetite for governed AI in finance is real and it’s running ahead of where most governance frameworks currently sit. Cambridge’s 2026 Global AI in Financial Services Report found 52% of industry respondents are now actively adopting agentic AI, with 81% expecting it to be meaningfully achieved by 2030.
Reporting is the clearest next step
Asked where they’re most likely to apply AI next, 47% of finance teams pointed to reporting and insights, followed by 26% for forecasting and planning, 16% for accounts payable and receivable and 11% for close and reconciliation.
That priority order reflects where AI delivers the fastest, lowest-risk wins today – turning raw data into narrative, surfacing variances and accelerating the path from numbers to decisions.
What this all points to
The picture from our poll, combined with what’s emerging globally, is clear. Finance teams have moved past the question of whether AI belongs in their function. They’re now wrestling with where to put it, how to govern it and how to bring it inside the systems where the work happens.
Cloud ERP platforms like NetSuite are racing to meet that demand. Through 2026, NetSuite is rolling out embedded AI capabilities including Autonomous Close, Ask Oracle (a natural language assistant across the entire dataset), AI-powered bank transaction matching, AI-generated narrative summaries in 2026.1 and agentic workflows that operate within NetSuite’s existing governance framework. The shift is from AI as a companion tool to AI as a core part of the operating model.
The teams that benefit most will be the ones who pair this technology with strong data foundations, clear policies on how AI-generated outputs are used and a finance-owned view of materiality and escalation.
How Annexa helps with AI finance strategies
At Annexa, we work with finance teams every day to bring AI inside their NetSuite environment – where data, controls and audit trails already live. Whether you’re refining your data foundations, switching on NetSuite’s embedded AI capabilities or building out a governance framework that lets your team move with confidence, we can help you turn AI from a parallel activity into part of how finance actually runs.
If you’d like to talk through where AI could deliver the most value in your finance function – and how to do it safely – get in touch with our team.
Frequently asked questions
What’s the difference between AI alongside finance systems and AI inside them? AI alongside your systems means tools like ChatGPT or Copilot operate separately from your ERP – you copy data out, get an answer, and bring insights back manually. AI inside your systems means intelligence is embedded directly into your ERP, working with your existing data, roles and permissions. The second approach keeps audit trails intact, applies your existing controls automatically, and removes the data movement that creates governance risk.
What does “agentic AI” mean for finance teams? Agentic AI describes systems that take multi-step actions on their own – completing reconciliations, proposing payments, or running close tasks within defined boundaries. For finance teams, the practical question is which tasks suit autonomous action and which still need human approval. The answer usually starts with low-risk, high-volume work and expands as confidence in the controls grows.
What does an AI governance framework for finance need to cover? A practical framework covers four areas: approved tools (which AI systems can touch finance data and which can’t), data handling (what can leave the system of record and under what conditions), human approval points (which decisions need sign-off and which can run autonomously), and audit trails (how AI-generated outputs are logged, explained and reviewed). Getting these four pillars in place early makes scaling AI much smoother later on.
How long does it take to embed AI into NetSuite workflows? Timelines vary with scope. Switching on NetSuite’s native AI features – such as AI-powered bank transaction matching, narrative report summaries, or invoice capture – can happen within weeks of a release upgrade. Building governance around those features, training teams, and rolling out agentic workflows is a longer effort that benefits from a phased approach: start with one or two high-impact use cases, prove the controls work, then extend.
What should finance leaders ask before turning on AI features in their ERP? Five questions worth asking:
- Which controls shift when AI does the work?
- Who reviews and approves AI outputs?
- How are exceptions flagged and escalated?
- What’s the audit trail for AI-generated decisions?
- And how will materiality thresholds change when speed increases?
Treating AI as a control redesign – rather than a feature toggle – sets teams up for sustainable adoption.
How can smaller finance teams compete with larger ones on AI adoption? Cloud ERP platforms like NetSuite are levelling the playing field. Smaller finance teams now have access to the same embedded AI capabilities as enterprise finance functions, without needing in-house data science teams. The advantage often goes to teams who move first on focused use cases – automating accounts payable, accelerating close, or strengthening forecasting – rather than trying to transform everything at once.