🏠︎ | Past Sessions | The AI-Ready CFO From Optimism to Real Impact
- Event: Finance Forum 25
- Date: 7 October 2025
- Speaker: Pauline Babel, CFO, Spendesk
- Estimated read time: 7-8 minutes
Quick summary
Finance leaders are confident that AI will transform the function, but most have yet to act.
Pauline Babel explored why this gap exists and why finance, despite seeing some of the strongest potential returns from AI, is lagging other functions in adoption. The issue is not lack of belief, but uncertainty around ROI, skills, trust and governance.
Drawing on real finance workflows, Pauline showed where AI delivers immediate value, particularly in month end close, reporting, forecasting and compliance. The session made a clear case for starting small, focusing on high volume repetitive work, and putting governance in place early so trust is built rather than eroded.
The underlying message was pragmatic rather than hype driven. AI will not replace finance teams, but it will fundamentally change how they work. The CFOs who move first will create time for insight and decision support, while those who wait risk falling behind.
The belief action gap in finance is widening
Pauline opened with a reality most finance leaders recognise. Teams are under pressure to deliver more, faster, and with fewer resources. In that context, AI is often discussed as a solution, yet adoption in finance remains slow.
Across UK industries, investment in AI tools increased dramatically between 2022 and 2024, with overall spend rising sixfold. Optimism among CFOs mirrors this trend. Around 80 to 85 percent believe AI will deliver real efficiency gains. Yet 61 percent have not implemented AI in finance in any meaningful way.
This disconnect matters because other functions such as sales, marketing and HR are already experimenting and learning. Finance risks being left behind, despite having some of the most automatable workflows in the organisation.
The challenge is not scepticism, it is execution.
Why ROI is clearer in finance than elsewhere
One of the most common reasons CFOs delay AI adoption is uncertainty around return on investment. Pauline acknowledged that this hesitation is understandable. Most public AI use cases are aimed at sales, marketing or talent teams, not finance.
However, when finance workflows are examined closely, the ROI case is often stronger than in any other function.
Month end close is a clear example. Finance teams regularly spend days pulling data from payroll, CRM, billing systems and other sources, reconciling figures and preparing reports under intense time pressure. These activities consume thousands of hours each year and are largely repetitive and rule based.
Even a modest reduction in manual effort can release significant capacity. Pauline noted that cutting 30 percent of this work can generate substantial savings, either in cost reduction or in redeploying time toward analysis and decision making.
The same logic applies to tax reporting, consolidation and forecasting. These are not speculative benefits. Case studies shared through the CFO Connect community show double digit reductions in effort when AI is applied to highly automatable finance tasks.
AI skills will become the new Excel baseline
A second barrier is capability. Many finance teams feel underprepared for AI, assuming it requires advanced technical skills or programming expertise.
Pauline challenged this perception by drawing a parallel with the arrival of Excel in the 1980s. At the time, finance professionals worried they would not be able to adapt. Today, spreadsheet proficiency is a basic expectation.
She argued that basic AI fluency will follow the same path. Skills such as prompting, interpreting outputs and light automation will become standard, not specialist.
One practical example came from a CFO who asked her team a simple question, what is the most boring task you do every month. The answers were predictable, expense reconciliation, invoice checking and report formatting.
By applying AI to these tasks, the team eliminated more than half of the manual work without learning advanced programming. The result was not job displacement, but higher engagement and more time for value adding analysis.
Trust, governance and data security come first
For finance leaders, trust is non negotiable. Pauline was clear that concerns around hallucinations and data security are valid and must be addressed upfront.
Finance cannot afford numbers that appear precise but are wrong. Nor can it risk sensitive data being used to train public AI models. Uploading payroll or contract data into free tools is simply not acceptable.
The answer is not rejection, but governance. Enterprise grade configurations, clear data access permissions and paid tiers that protect inputs are essential. Human approval must remain in place, particularly for external reporting and regulatory submissions.
Pauline emphasised that finance professionals are already familiar with these controls. Applying the same discipline to AI ensures trust is protected rather than undermined.
Where AI delivers immediate value in finance
The most obvious starting point for AI in finance is automating high volume, rules based work.
Invoice processing, expenses, accounts payable and receivable are all areas where AI can reduce manual effort significantly. Technology can already read receipts, extract data and categorise spend with high accuracy.
Pauline shared examples such as splitting hotel receipts into accommodation and food, a task that previously consumed hours of accounting time. At Spendesk, AI is used to detect duplicate invoices before they reach the ledger, flagging anomalies early and reducing downstream errors.
Reporting and forecasting are another high impact area. Today, many CFOs report that less than half of their reporting is fully digitised. AI can stitch together data from multiple sources, generate draft reports and even produce commentary.
Pauline described using AI agents to support monthly narrative reporting. While outputs still require review and refinement, the time saved is meaningful.
She also referenced a case where a CFO moved from complex Excel based forecasting to an open source AI model, achieving over 95 percent accuracy for a tested product line. This represents a shift from reactive, backward looking reporting to more forward focused insight.
Expanding analysis beyond what was previously possible
Beyond efficiency, AI unlocks types of analysis that were previously impractical.
Contract analysis is one example. Finance teams often manage hundreds of supplier agreements with inconsistent payment terms, renewal clauses and cash flow implications. Reviewing these manually can take weeks.
With AI, contracts can be uploaded and summarised in minutes, producing a consistent view of key terms across the portfolio.
Similarly, AI reduces bottlenecks in data access. Instead of relying on BI teams to run SQL queries, finance leaders can ask questions in plain language and receive structured outputs. This does not replace data teams, but it makes finance more self sufficient and responsive.
In compliance and audit, AI enables continuous testing rather than sampling. Instead of reviewing 1 to 10 percent of transactions, teams can move toward full population testing, improving accuracy and reducing risk.
Pauline described AI as a tireless risk officer, consistent, always on, and capable of monitoring transactions in real time when configured correctly.
Choosing the right tools requires discipline
The AI finance tool landscape is expanding rapidly. Pauline referenced research reviewing over 300 finance specific tools, narrowed down to around 100 that are truly relevant.
Rather than recommending vendors, she proposed a simple evaluation framework built around three questions.
Does the tool integrate with the existing finance stack, including ERP, payroll, CRM and billing systems.
Can the ROI be measured clearly, in terms of time saved, errors reduced or days cut from close.
Is the tool simple enough for teams to adopt and trust.
If the answer to any of these questions is no, the advice was clear, walk away. Paying for tools that are never used is the worst possible investment.
How to move from interest to action
Pauline outlined a practical roadmap for CFOs ready to move beyond experimentation.
Start small by selecting one repetitive, high volume workflow. Clean the data to a usable standard, recognising that perfection is not required. Choose tools with clear outcomes and run pilots over six to eight weeks to maintain momentum.
Governance should be established from day one. Clear policies on data use, permissions and approval thresholds protect trust and enable faster adoption later.
Above all, focus on adoption. Technology only delivers value when teams use it consistently and confidently.
What good looks like, practical actions for finance leaders
This session translated into clear, practical guidance for CFOs assessing their AI readiness.
Questions to ask your team- Which finance tasks consume the most time with the least strategic value
- Where are errors or delays most common in month end close
- Which reports are still heavily manual and why
- What data access bottlenecks slow decision making
- Where would additional capacity create the most insight
- Whether finance is spending less time reconciling and more time analysing
- Whether reporting cycles are shortening without sacrificing accuracy
- Whether teams trust AI outputs enough to use them regularly
- Whether governance is enabling experimentation rather than blocking it
- Waiting for a perfect business case before starting
- Assuming AI requires advanced technical skills to be useful
- Using consumer grade tools with sensitive financial data
- Rolling out too many tools without clear ownership or outcomes
- Treating AI as a side project rather than a finance capability
What good looks like in practice
An AI ready finance function uses technology to remove low value work, not to replace professional judgement. It applies governance early, builds trust through accuracy, and reinvests time saved into forecasting, scenario planning and strategic decision support.
Conclusion, AI readiness is a leadership choice
Pauline closed with a clear message. Adoption in finance is still early, but optimism is already high. The gap between belief and action is narrowing, and those who delay will find it harder to catch up.
The barriers are real, but solvable. Finance has navigated major technology shifts before and AI is no different. This moment represents an opportunity to reclaim time, strengthen insight and reposition finance as a true strategic partner.
Key takeaways
- Finance shows high optimism about AI but low levels of real adoption
- ROI is often clearer in finance than in other functions due to repetitive workflows
- AI skills will become a baseline capability, similar to Excel
- Governance and trust are essential, not optional
- Starting small and focusing on adoption delivers faster impact
Speaker
CFO, Spendesk, previously VP Finance at Aviv, with experience scaling finance operations, driving profitability and leading transformation across high growth organisations.