What Is Actually Changing
The phrase "AI in finance" encompasses a wide range of tools with very different levels of maturity and impact. To plan your team structure correctly, it helps to be specific about which tools are now genuinely production-ready and what they actually do to workloads.
Microsoft Copilot integrated into Excel and Dynamics 365 is now in active use at a material number of UK growth companies. Its most significant practical impact is in the preparation of management account commentaries, variance analysis, and the routine summarisation of large datasets. Tasks that once took a finance analyst two to three hours per month close can now be completed in minutes, with human review taking perhaps another thirty minutes to validate and refine the output.
Automated reconciliation tools, including those embedded in Xero, NetSuite, and specialist reconciliation platforms, have effectively eliminated the bulk of manual bank reconciliation work for companies with clean transaction data. For a company processing 5,000 transactions per month, a reconciliation that once required half a day each month is now largely automatic, with the finance team reviewing exceptions rather than processing matches.
AI-assisted close processes are increasingly compressing the monthly close timeline. A company that previously ran a five-day close can, with appropriate tooling and data discipline, achieve a two-to-three day close. This creates capacity: not capacity to reduce headcount immediately, but capacity that should be reallocated to higher-value activities if it is managed deliberately.
What is not changing quickly is the work that requires contextual judgement, relationship management, and genuine financial expertise. Business partnering, investor relations, regulatory engagement, M&A, treasury management, and the synthesis of financial data into board-level narrative remain firmly human activities. The tools can assist; they cannot replace the judgement that underpins these functions.
Roles Most Affected by Automation
The Gartner Finance Automation Survey identified four role categories where AI and automation are having the most material impact on time allocation in 2025-2026:
- Accounts payable and receivable processing: invoice matching, payment runs, cash allocation, and credit control chasing are all substantially automatable with current tools. The residual human role in AP/AR is exceptions management, supplier relationship management for complex or disputed invoices, and credit risk assessment.
- Management accounts preparation: the mechanical assembly of management accounts from source data, including inter-company eliminations and standard journals, is increasingly handled by automated processes. The human role shifts to review, interpretation, and communication.
- Basic FP&A work: routine variance analysis, budget vs actuals commentary, and standard reporting pack preparation are all being compressed by AI tools. The residual human role is scenario modelling, forecast challenge, and commercial interpretation.
- Payroll processing: for companies using integrated payroll systems, the processing element of payroll is now largely automated. The human role is governance, compliance checking, and exception resolution.
Roles Least Affected (and Most in Demand)
The flip side of automation is that it dramatically increases the value of roles that cannot be automated. CIMA's Future of Finance Report identifies business partnering as the most under-resourced capability in UK finance functions for 2026: the ability to sit alongside commercial teams, challenge assumptions, translate financial data into decisions, and provide the financial backbone for strategic conversations is in acute shortage.
The roles that are least affected by current AI tools — and where demand is increasing — include:
- Senior FP&A and commercial finance: building integrated business models, scenario planning, and capital allocation analysis require financial judgement that AI tools cannot replicate.
- Regulatory and compliance finance: the complexity and pace of change in UK financial regulation means that regulatory finance expertise is increasingly valuable and difficult to automate.
- Treasury and risk management: cash management strategy, FX risk, counterparty risk, and investment policy all require contextual judgement and relationship management.
- M&A and transaction support: financial due diligence, structuring, and post-deal integration are high-complexity activities where AI is an aid to research rather than a replacement for expertise.
- CFO-level business partnering: the ability to sit on an executive team, challenge commercial decisions with financial data, and represent the finance function at board level is inherently human.
Before and After: Finance Team Structure Comparison
The following comparison illustrates how a finance team at a 50-person growth company might evolve as AI tooling matures. This is not a prescription — company-specific factors including regulatory obligations, revenue complexity, and fundraising cadence all affect the right structure. But the directional shift is broadly applicable.
The headline headcount change is modest. The composition change is significant. Fewer hours spent on processing and preparation; materially more hours on analysis, partnering, and regulatory work. The net cost impact depends heavily on seniority mix: a finance team of three senior analysts and a CFO is more expensive than a team of four processing-focused staff, even if the headcount is lower.
"The CFO's dilemma in 2026 is not whether to reduce headcount — it is whether the people currently in transactional roles can make the transition to analytical roles, and how to manage the human reality of that conversation."
The CFO's Dilemma: Reduction vs Redeployment
When AI tools create capacity in the finance function, CFOs face a choice: harvest that capacity as headcount reduction, or redeploy it into higher-value activities. Both approaches can be right, depending on the organisation's stage and circumstances.
For a pre-profit company where cost discipline is critical, capturing AI-driven savings as a reduction in headcount costs can be the right decision, particularly where the team does not have the appetite or the skills to shift into analytical roles. In this case, the CFO's responsibility is to manage the transition fairly and with appropriate notice, and to ensure that the capability that is redeployed elsewhere in the business (or released through natural attrition) is replaced with the right skills when the next hire is made.
For a company investing in growth, redeployment is often the better answer. A transactional finance person who genuinely wants to develop analytical skills, and who has the intellectual capability to do so, can be a more cost-effective path to building FP&A capability than hiring a new senior analyst from outside. The redeployment path requires investment: training in modelling tools, data analysis, and the commercial context of the business. It also requires honest assessment of whether the individual has the aptitude for the transition, which is a conversation that many managers avoid.
Managing the Transition
The practical management of a finance team transition toward an AI-augmented structure has several components that CFOs frequently underestimate. The technical change, implementing the tools and reconfiguring processes, is typically the easiest part. The human change management is harder.
Finance professionals whose roles are being automated often experience the transition with anxiety, even when they are not at risk of redundancy. Clear communication about what is changing, why, and what the team will look like after the transition is essential. A CFO who introduces automation tools without explaining the implications for the team's work will create uncertainty and attrition among exactly the people they most want to retain.
Training investment is non-negotiable. The finance team that will thrive in the AI-augmented environment needs to be comfortable working with AI-generated outputs: knowing when to trust them, when to challenge them, and how to iterate on them to improve quality. This is a skill that can be taught, but it requires deliberate investment, not passive exposure to tools.
What the Finance Function Looks Like at Different Company Sizes
The right finance team structure in an AI-augmented environment varies materially by company size, regulatory complexity, and growth stage. The following provides a reference framework:
- Seed to Series A (10-30 people): Finance function is typically a fractional CFO plus one or two bookkeepers or finance assistants using automated tools. The CFO's time is concentrated on fundraising support, financial modelling, and basic controls. AP/AR automation is table stakes at this stage.
- Series A to Series B (30-100 people): A full-time CFO or Finance Director is warranted. The team extends to include a senior FP&A analyst and a finance manager overseeing the automated transactional processes. Business partnering starts to emerge as a distinct function. Regulatory compliance becomes a dedicated part-time role.
- Series B and beyond (100+ people): The finance function differentiates into clear sub-functions: financial control (with automated transaction processing as the backbone), FP&A (model-driven, analytically focused), treasury, and regulatory or compliance finance. The headcount in financial control shrinks relative to company size; the headcount in FP&A and compliance grows.
Key Takeaways
- AI tools are materially reducing time spent on AP/AR processing, management accounts preparation, bank reconciliation, and routine FP&A work; plan for 60-80% time savings in these areas with appropriate tooling.
- Business partnering, regulatory finance, treasury, and M&A support are least affected and most in demand; rebalance team composition toward these functions.
- The directional shift at a 50-person company is from 4 FTE of mostly transactional staff to 3-3.5 FTE of mostly analytical staff; the cost impact depends heavily on seniority mix.
- Redeployment before reduction is the right default for growing companies; it requires honest capability assessment and a structured 6-month development plan per person.
- Managing the human dimension of the transition is harder than the technical implementation; communicate clearly, invest in training, and avoid creating uncertainty that drives attrition.
- The highest-value finance hire in 2026 combines deep technical financial expertise with strong commercial acumen and effective stakeholder communication skills.
- Data governance and system integration are prerequisites for AI ROI; do not expect AI tools to create capacity in a finance function built on disconnected spreadsheets.
