Artificial intelligence and finance: 7 challenges and risks

AI in finance: a state of play 2026-2027

The adoption of AI remains contrasted across sectors and countries. According to a study relayed by Squid Impact, only 10% of French companies with more than ten employees were using at least one AI technology in 2024, compared to 6% a year earlier, a level still lower than the European average of 13%. Finance, however, stands out as a priority investment hub, supported by global momentum: the global AI market reached $638 billion in 2024 and is expected to exceed $3.68 trillion by 2034, with a CAGR of 19.2%.

In this transformation, the finance function — like management control or operational departments — is seeing its processes profoundly transformed. The subject is part of a broader sectoral dynamic, which is also addressed in our analysis dedicated to artificial intelligence for banking.

8 strategic challenges of AI for finance

1. Optimization and automation of financial processes

AI makes it possible to automate the reconciliation of accounts, the management of payments, the production of reports and the processing of supplier invoices. Beyond operational efficiency, these gains free up time for higher value-added analyses and fuel a real operational efficiency approach.

2. Predictive Analytics and Algorithmic Trading

Predictive models leverage massive volumes of market data, news feeds, and macroeconomic indicators. Algorithmic trading has become the norm: nearly 92% of transactions in the foreign exchange (Forex) market are now driven by algorithms. The integration of large language models (LLMs) now also makes it possible to exploit textual data — dispatches, reports, corporate communications.

3. Financial Risk Management

AI refines the assessment of credit, market, liquidity and operational risks. Learning models can be used to simulate stress scenarios, anticipate defects, and continuously monitor exposures. This increase in power is part of a structured risk management approach.

4. Personalization of services and KYC

The detailed analysis of customer data feeds into more accurate customer knowledge (Know Your Customer), enables personalized recommendations and supports loyalty. Conversational chatbots and virtual assistants extend this experience 24 hours a day, leveraging machine learning and LLMs.

5. Fraud detection and prevention

AI engines detect atypical behavior, suspicious transactions, and anti-money laundering (AML) patterns in real time. The results are tangible in the public sphere: €628 million in fraud was detected and stopped by the Health Insurance in 2024, an increase of 35% compared to 2023, thanks in particular to the deployment of new technological tools.

6. Portfolio optimization and robo-advisors

Robo-advisors and algorithmic allocation engines adjust portfolios according to the risk profile, objectives and investment horizon. Asset management is becoming more granular and responsive, while remaining governed by strict prudential rules.

7. Cost reduction and competitiveness

Automating high-volume processes makes it possible to reduce unit costs, streamline operations and allocate human resources to higher value-added missions. This logic feeds the financial transformation of the company.

8. A battle for AI talent

Skills in AI, data science and MLOps applied to finance have become strategic. Finance departments are structuring hybrid teams combining business, data and compliance profiles, and rely on specialized interim managers to steer these transformations.

The European AI Act: a new structuring framework for finance

The gradual entry into force of the European regulation on artificial intelligence (AI Act) marks a turning point for financial players. This text establishes a classification of AI systems by level of risk — unacceptable, high, limited, minimal — and imposes reinforced obligations on systems considered to be high-risk, which include credit scoring and certain insurance uses.

For banks and insurers, this translates into requirements for technical documentation, data governance, model transparency, human oversight and risk management. The gradual implementation of the Basel III agreements began on 1 January 2025, the DORA regulations also apply from mid-January 2025, and the entry into force of the AI Act is announced for 2026. The financial sector must therefore deal with a veritable regulatory layer, where anticipation is becoming a key factor for success.

Wayden’s key point: finance and compliance departments must now map their AI use cases, qualify the associated AI Act risk level and formalize documented governance. The ACPR is the competent authority to supervise these obligations in the French banking and insurance sector.

7 Risks of AI for Finance

1. Data Security and Cybersecurity

AI is fueled by sensitive data: customer data, transactions, scoring. The attack surface is expanding with the proliferation of models, APIs, and third-party vendors. Operational resilience, now supervised by DORA, becomes a subject board.

2. Algorithmic biases and fairness

Models trained on historical data can reproduce and amplify discriminatory biases (gender, origin, geography). A biased credit scoring can lead to undue refusals, litigation and reputational risks. The AI Act imposes explicit dataset quality requirements for high-risk systems.

3. Technological dependence and systemic risk

Over-reliance on algorithms — especially in trading — can amplify volatility and create pack effects. With nearly 92% of Forex trades driven by algorithms, human control and circuit breakers remain indispensable to preserve market stability.

4. Regulatory Compliance

GDPR, AI Act, DORA, Basel III, MIFID II, ACPR requirements: finance is under dual supervision, prudential and digital. Every AI model must be able to be documented, audited, explained and traced. This complexity requires robust internal systems, sometimes reinforced by a legal transition manager.

5. Explainability and governance of models

Complex models — deep learning, LLMs — often work like black boxes. However, regulators and customers demand explainable decisions, in particular for credit refusals or insurance risk assessments. MLOps governance is becoming a cross-functional project for CIOs, compliance and business lines.

6. Impact on employment and retraining

Automation is transforming the nature of jobs: some tasks are disappearing, others are emerging. Financial institutions must invest in training, upskilling and support change management to keep teams engaged.

7. Implementation costs and uncertain ROI

Deploying industrial AI requires significant investments: cloud infrastructure, data governance, recruitment, security, compliance. Without a defined use case and clear performance indicators, the ROI can be disappointing.

Things to doMap use cases, qualify AI Act risk, document models, maintain human oversight of sensitive decisions. What to avoidDeploy generative AI in production without governance, traceability or bias assessment, or continuity plan in case of drift.

Real-world examples of AI deployment in finance

• Dynamic credit scoring : retail banks integrating alternative variables (transactional behavior, digital history) to refine the granting of loans, under AI Act control.
• Real-time fraud detection : Machine learning engines analyzing every card transaction and transfer, with instant risk scoring.
• Algorithmic trading augmented by LLM : Leverage news reports and financial publications to automatically adjust positions.
• Wealth robo-advisors : personalised asset allocation and automated arbitrage for individual savers.
• Compliance chatbots : Internal tools to help KYC/AML/CFT analysts process complex cases faster.
• Automated financial reporting : Accelerated production of financial statements, consolidation bundles and executive dashboards.

How Wayden supports finance departments in their AI trajectory

Wayden mobilizes experienced interim managers to lead financial institutions’ AI projects: strategic scoping, AI Act and DORA compliance, data governance, operational deployment and change management. Our missions cover both the IS and the finance departments, with a cross-functional approach integrating business, technological and regulatory issues.

Are you structuring your AI finance roadmap?

Our interim managers help you frame, secure and accelerate your projects, in compliance with the AI Act, DORA and ACPR requirements.

Talk to a Wayden Expert

FAQs: AI and finance

What are the main use cases of AI in finance in 2025-2026?

Fraud detection, credit scoring, algorithmic trading, robo-advisors, reporting automation, KYC/AML, conversational assistants and predictive market analysis are among the most mature uses.

What is the AI Act and what is its impact on finance?

The AI Act is the European regulation on artificial intelligence. It classifies systems by risk level and imposes reinforced obligations on high-risk uses (credit scoring, certain insurance uses). It will come into force until 2026, under the supervision of the ACPR in France.

What are the major risks of AI for a financial institution?

Cybersecurity, algorithmic biases, technological dependence, lack of explainability, regulatory non-compliance, impact on employment and uncertain ROI are the seven structuring risks.

How to comply with the AI Act in the financial sector?

Map AI systems, qualify the level of risk, document models, guarantee data quality, formalize governance, ensure human supervision and prepare controls by the competent authority (ACPR for banking and insurance).

Will AI replace finance jobs?

AI transforms jobs more than it replaces them. Repetitive tasks are becoming automated, but analytical, model management, compliance and customer relationship skills are gaining in value.