AI in Financial Services Forum Reflections

Overview

The AI in Financial Services Roundtable, hosted by HCR Law in partnership with CyNam as part of the Everyday AI series, brought together professionals from across financial services, technology and advisory roles. The discussion focused on how AI is being adopted across the sector, the barriers to progress, and what responsible use looks like in practice.

Defining AI

A recurring point was the lack of clarity around what AI is. Participants noted that generally, people still associate AI with generative tools like ChatGPT and Gemini, rather than a much broader selection of tools and technologies. In practice, all AI currently in use is narrow AI, which is defined by systems that are designed to perform specific tasks. More advanced concepts, known as artificial general intelligence or super AI, remain theoretical.

A mutual understanding of what AI is (and isn’t) is central to informed decision-making, building trust amongst end users, and supporting responsible use.

How artificial intelligence, machine learning, deep learning and generative AI are related.

Source: IBM

Using AI

From the participants, it was clear that AI is already embedded across financial services in a range of applications. Several types of narrow AI referenced included machine learning, natural language processing, computer vision and expert systems. Machine learning models are used to detect fraud and financial crime, assess credit risk and monitor transactions in real time. Natural language processing underpins chatbots and virtual assistants and computer vision supports document and identity verification.

These technologies are increasingly combined to create what is known as expert systems. A common example is KYC onboarding, where identity documents are verified, information is extracted and risk is assessed in a simple and quick workflow.

Key use cases

A wide variety of use cases were named, which are outlined below.

Fraud and financial crime: AI-driven fraud detection has significantly reduced false positives, allowing teams to focus on genuinely suspicious activity and improving overall outcomes.

Credit risk and lending: AI models are being used to move beyond traditional credit scoring by analysing a broader range of data, enabling faster, more accurate and more personalised lending decisions.

Customer service: Chatbots and voice assistants can provide instant and around-the-clock support, improving response times to queries and reducing pressure on call centres while allowing staff to focus on complex or sensitive cases.

Internal productivity: Tools such as Copilot are being used to streamline routine tasks, analyse data and support document creation, saving employees several hours each week.

Digital identity verification: AI is also being deployed to counter advanced threats such as deepfakes and voice cloning, supporting authentication processes and reducing fraud risk.

Use Cases of AI in Financial Services

Source: Bank of England, 2024

Barriers to adoption

Despite growing adoption, several challenges still need to be overcome to realise the full benefits of AI.

Data quality: AI systems are only as effective as the data they are trained on. Fragmented systems, inconsistent formats and incomplete datasets limit the performance of AI tools.

Change and culture: AI adoption can change how work is done and how decisions are made, requiring employees to adapt to new processes and trust AI-supported insights. Without clear communication, training and reassurance, resistance (often fuelled by fears of job displacement) can slow adoption and limit the potential benefits.

Build, buy or partner decisions: Organisations face trade-offs between developing AI in-house, buying off-the-shelf solutions or partnering with specialist providers. Each approach has implications for cost, speed, control and risk.

Overreliance on AI: There is a risk that people place too much trust in AI outputs without applying critical thinking. Even with AI support, human judgment and accountability are essential, especially in regulated environments where decisions have profound consequences.

Cyber security risks: AI can increase the attack surface through system complexity, third-party integrations and more sophisticated cyber attacks, including model manipulation, prompt injection and AI-enhanced phishing.

Responsible AI

Participants noted that responsible deployment is critical to effective and resilient AI adoption. Many organisations are developing internal AI policies, strengthening governance, and investing in training and testing.

A notable recent development is the FCA’s AI Live Testing programme, which aims to define what safe and responsible AI looks like in practice through real-world testing. The programme reflects the UK’s ongoing interest in guiding responsible AI deployment, with regulators exploring principles such as transparency, fairness, accountability, and contestability.

Source: Responsible AI in financial services, EY

Conclusion

The forum highlighted that AI is delivering value across a wide variety of areas in financial services, including fraud prevention, lending, customer service and internal operations. Adoption is driven by clear goals: improving outcomes, reducing costs and managing risk.

The key to progress now is responsible implementation. High-quality data, strong governance and meaningful human oversight will be essential to realising AI’s benefits while keeping trust, fairness and accountability across the sector.