The numbers don't add up. 88% of finance professionals believe AI will be the most transformative technology in accounting and finance over the next 12–24 months. But when the AICPA and CIMA surveyed over 1,400 senior finance leaders, only 8% said their organisation is "very well prepared" to manage that transformation.
Finance teams are facing the same problem as the rest of the enterprise (tools arriving faster than skills) but with higher stakes. When a marketing team experiments with AI and gets something wrong, they publish a bad social post. When a finance team gets it wrong, they risk misstated financials, compliance violations, or audit failures.
The Finance AI Readiness Gap
AI adoption in finance functions has grown rapidly, from 37% in 2023 to 59% in 2025, according to Gartner. But the skills to use these tools effectively haven't kept pace. The gap is showing up everywhere: in survey data, in the job market, and in the C-suite.
| Finding | Source |
|---|---|
| 88% say AI will be most transformative tech in finance (12–24 months) | AICPA/CIMA (1,400+ finance leaders, Aug–Sep 2025) |
| Only 8% say their org is "very well prepared" | AICPA/CIMA |
| 56% cite generative AI as most prominent skills gap | AICPA/CIMA |
| 50% say lack of skills/talent is top barrier to AI adoption | AICPA/CIMA |
| 31% of finance job postings now reference AI/ML skills (up from 25%) | Datarails (5,000 job listings analysed) |
| Building AI talent is CFOs' #1 near-term challenge | Gartner (100 CFOs, Jan–Feb 2026) |
The job market tells the story most clearly. Datarails analysed over 5,000 US finance job postings between January 2025 and January 2026 and found that nearly one in three now explicitly reference AI or machine learning capabilities. For FP&A roles, it's even higher: 43%, up from 33% a year earlier. AI mentions in accountant postings surged 67%.
The biggest barrier to AI in finance isn't technology maturity or budget. It's people. 50% of finance leaders say a lack of skills and talent is their top obstacle, ahead of security concerns (47%) and doubts about technology readiness (42%).
What Goldman Sachs, JPMorgan, and EY Are Doing
The finance organisations getting real value from AI share a common pattern: they're investing in their people's ability to use AI tools, not just deploying the tools themselves.
Goldman Sachs: AI Assistant Goes Firmwide
Goldman Sachs rolled out its GS AI Assistant firmwide in mid-2025 after piloting with roughly 10,000 employees. The tool handles document summarisation, content drafting, data analysis, and research translation. In investment banking, it reportedly reduced deck preparation time by 50%, translating into thousands of reclaimed analyst hours and faster client turnarounds.
But Goldman didn't just hand everyone an AI tool and walk away. The rollout included structured onboarding that taught employees how to use the assistant for their specific workflow, with research analysts learning different capabilities than wealth advisers or compliance teams.
JPMorgan: Coach AI for Client-Facing Teams
JPMorgan expanded AI beyond back-office automation with Coach AI, an internal tool for private client advisers. During the April 2025 market volatility, when clients flooded advisers with questions, Coach AI helped advisers find relevant information up to 95% faster, shifting time from searching to client engagement.
Combined with their LLM Suite (200,000 users onboarded in eight months) and $18 billion technology investment, JPMorgan's approach treats AI training as inseparable from AI deployment. Every tool launch includes a structured learning component.
EY: $1 Billion Bet on AI-Powered Audits
EY committed US$1 billion over four years to integrating AI into its global assurance platform, supporting over 160,000 audit engagements and equipping 140,000 assurance professionals with new tools. By 2025, the programme had launched over 100 AI capabilities including GenAI-powered knowledge search, intelligent checklists, and automated financial statement tie-outs. But EY's investment isn't purely technological. Every tool deployment includes structured training for auditors who need to interpret AI-flagged anomalies, assess model confidence, and make judgment calls that no algorithm can automate.
Goldman Sachs, JPMorgan, and EY all paired AI deployment with structured training. The tools only delivered value when people understood how to use them in their specific finance context: deck preparation, client advisory, or audit review. Generic AI training wouldn't have produced these results.
The Skills Finance Professionals Actually Need
AI fluency in finance isn't about understanding transformer architecture or writing Python scripts. It's about being able to work confidently with AI tools in a regulated, controls-heavy environment where errors have material consequences.
Prompt engineering for financial workflows. Writing effective prompts for finance isn't the same as general prompt engineering. Finance prompts need to encode decision trees, escalation thresholds, approval hierarchies, and compliance constraints. A prompt that generates decent marketing copy fails when applied to revenue recognition or lease classification. kju covers this in the prompt engineering track with scenarios grounded in finance workflows.
Data interpretation and model evaluation. When AI recommends a credit decision, flags a suspicious transaction, or generates a rolling forecast, someone needs to understand the model's confidence level, its training data limitations, and where it might be wrong. This is especially critical in FP&A, where the AICPA/CIMA survey shows the fastest adoption: 47% of organisations now use AI to recommend personalised development pathways, and similar tools are emerging for financial planning.
Controls and auditability. Every AI-assisted decision in finance needs an audit trail. Segregation of duties doesn't disappear because a model made the recommendation. Understanding how to preserve controls while adopting AI is non-negotiable, and it's covered in depth in the AI governance track.
Regulatory compliance. The EU AI Act classifies AI systems used in creditworthiness assessments as high-risk, requiring documented training, human oversight, and bias monitoring. Finance teams handling credit, insurance, or investment decisions need to understand these requirements before the August 2026 compliance deadline.
Workflow redesign. L.E.K. Consulting found that only 11% of CFOs currently use AI within their finance functions, despite 60% believing it will be highly impactful. The gap isn't budget; it's that teams are trying to bolt AI onto existing processes rather than redesigning workflows around it.
The core skills gap in finance isn't technical; it's contextual. Finance professionals need to understand AI in the context of financial controls, regulatory requirements, and fiduciary responsibility. Generic AI training misses this entirely. The AICPA/CIMA survey confirms it: 61% of finance leaders rank on-the-job training as the most effective format, and 62% favour internal training programmes over external alternatives.
Why Finance Can't Afford the Workshop Approach
Finance teams have a specific version of a universal training problem: the gap between knowing about AI and knowing how to use it in a regulated environment.
56% of finance professionals identify generative AI as their most prominent skills gap. Self-directed experimentation (trying ChatGPT on your own, sharing prompts in Slack) doesn't address this. In finance, unstructured AI usage creates specific risks: inconsistent outputs across teams, unauditable decision processes, and compliance exposure that individual experimentation can't detect.
BCG's AI at Work 2025 data quantifies what works instead. Employees who receive at least five hours of structured AI training become regular AI users at a rate of 79%, compared with 67% for those who get less. And critically, only 25% of frontline workers say their leaders provide enough guidance on AI, meaning even well-intentioned CFOs aren't translating their AI ambitions into team-level capability.
The alternative is structured daily practice. Six minutes a day, focused on finance-specific AI skills, builds lasting competency without competing with close cycles, audit schedules, or quarter-end deadlines. That's how kju works: short daily sessions designed for working professionals who can't afford multi-day workshops but can't afford to fall behind on AI either.
When an entire finance team builds AI fluency together, they develop shared standards: consistent prompt approaches, agreed evaluation criteria for AI outputs, common controls for AI-assisted decisions. Individual learning produces individual habits. Team learning produces the kind of organisational capability that Gartner's 100 surveyed CFOs identified as their most challenging priority.
Frequently Asked Questions
- What AI skills do finance professionals need?
- Finance professionals need prompt engineering for financial SOPs and reporting, data interpretation skills for evaluating AI-generated forecasts and anomaly flags, model risk management to assess AI reliability in credit and compliance decisions, controls and auditability knowledge to preserve segregation of duties, and AI governance skills for regulatory compliance. These are finance-context skills, not deep technical ML knowledge.
- Will AI replace accountants and finance professionals?
- No, but it's reshaping what they do. AI mentions in accountant job postings surged 67% year-over-year according to Datarails, and one in three finance roles now require AI skills. The demand is for finance professionals who can work alongside AI, not compete with it. Goldman Sachs and JPMorgan are upskilling existing teams, not replacing them.
- How is AI currently used in corporate finance?
- The most common use cases are recruitment support (51% of organisations), learning and development (39%), and HR technology (42%), according to SHRM. In finance specifically, AI powers fraud detection, accounts payable automation, FP&A forecasting, and close-cycle acceleration. Goldman Sachs reports its AI assistant cut deck preparation time by 50%.
- What ROI can finance teams expect from AI training?
- Gartner found that 67% of CFOs already using AI in finance are more optimistic than a year ago, and adoption jumped from 37% in 2023 to 59% in 2025. But the AICPA/CIMA survey shows only 8% feel well-prepared. Structured training closes this gap. BCG data shows employees with 5+ hours of structured AI training become regular users at 79% versus 67% for those with less.