AI in Finance: Transforming an Industry Through Automation, Trust, and New Skills 

Artificial intelligence is rapidly reshaping industries worldwide, but few sectors stand to gain—or lose—as much as finance. Highly regulated, data‑intensive, and dependent on accuracy, the financial sector finds itself at a pivotal moment: AI promises substantial productivity gains, yet demands strong governance, trust, and new capabilities for its workforce. 

A recent discussion with Associate Professor Ke-Wei Huang, Executive Director at the Asian Institute of Digital Finance offers a rare, in‑depth look at how AI is unfolding across financial services, where the technology is headed, and what financial professionals must do to thrive alongside it. This article is generated from the points brought up during the discussion. 

Catch the conversation in the latest episode of The AIMX Podcast. 

AI Adoption in Finance: Hype vs. Reality 

AI adoption forces us to confront a recurring tension: AI is both over‑hyped and genuinely transformative. 

On the one hand, media narratives portray AI as nearly superhuman—capable of replacing entire departments or running entire companies as autonomous agents. On the other hand, financial institutions experimenting with AI quickly learn that: 

• AI makes errors, 
• It cannot reason without historical data, 
• It cannot handle ambiguity as humans do, 
• And it requires significant oversight. 

Most large institutions have invested heavily in AI projects to understand where the technology brings true ROI. Many early assumptions—that AI would be flawless or self‑driving—have given way to practical experimentation. Organizations are now evaluating which business functions benefit most from AI and which require human‑centric reasoning or judgment. 

Small and medium-sized firms, meanwhile, often lack the resources for full-scale AI projects. Yet employees have organically embraced tools like ChatGPT or Gemini to increase personal productivity. 

Across the board, the belief is that AI will eventually become as ubiquitous as the internet—something every business uses, whether or not it makes headlines. 

Where AI Is Making the Biggest Immediate Impact 

One area seeing rapid transformation is financial reporting and auditing. 

AI now excels at tasks that involve: 

• Reading large volumes of documents 
• Extracting information 
• Summarising content 
• Drafting reports 

These capabilities map closely to repetitive, document-heavy processes that have traditionally consumed vast amounts of human time. 

Other areas also show strong AI potential, including: 

• Credit risk assessment 
• Fraud detection 
• Compliance checks 
• Simple financial analysis 

Yet AI struggles where reasoning, judgment, or sparse historical data are required. 

Pushing the Frontier: Multimodal AI in Earnings Calls 

A fascinating use case for advanced innovation is developing multimodal AI systems to analyze earnings calls by companies. 

Earnings calls include: 

• Numerical financial data 
• Written transcripts 
• Audio recordings revealing tone, emotion, hesitation, and sentiment 

AI can now merge these inputs to detect patterns and signals beyond human capacity. For example: 

• Tracking whether forward‑looking statements across quarters ever came true 
• Identifying mismatches between positive verbal content and negative emotional tone 
• Detecting subtle cues in speech that humans may miss 

Such systems could help investors, analysts, and compliance teams uncover hidden risks or inconsistencies from company executives’ statements and responses during their quarterly earnings calls. 

The Trust Problem: Making AI Transparent and Auditable 

A major barrier in finance is trust. AI models are “black boxes,” and this is incompatible with regulatory expectations for transparency, traceability, and auditability. 

Researchers are tackling this challenge on several fronts: 

1. Simplifyingblack-boxmodels 

Developing methods that approximate complex models using transparent mathematical expressions—sacrificing some accuracy for interpretability. 

2. Hybrid models

Starting with traditional financial econometrics and layering AI on top, preserving transparency where it matters most. 

3. Guardrails and safety layers

Implementing filters that: 

• Block harmful or irrelevant user inputs 
• Validate outputs for correctness 
• Flag anomalies 

Even with these advances, AI is far from perfect. Accuracy varies dramatically depending on the task, and human oversight remains essential. 

As noted, “zero‑error tolerance” is the norm in finance—making human‑in‑the‑loop design mandatory for years to come. 

What Will Happen to Jobs? 

AI is automating many entry‑level white-collar tasks in finance, especially those involving repetitive paperwork. This mirrors what is happening in programming: AI can write simple code extremely well, reducing the need for junior developers. 

However, several categories of work remain safe: 

• Relationship management 
• Sales roles requiring human trust 
• Team leadership 
• Complex decision-making 
• Roles involving nuanced judgment 

For mid-level and senior workers, AI will boost productivity rather than replace jobs. 

What skills will finance professionals need? 

Not coding ability—but: 

• Understanding AI’s capabilities and limitations 
• Knowing how to incorporate AI into workflows 
• Being able to supervise AI tools effectively 
• Skills in change management and digital strategy 

In short: working with AI, not competing against it. 

A Glimpse Into the Future: Software Without Buttons 

Looking five years ahead, one bold prediction emerged: Software interfaces as we know them may disappear. 

Instead of navigating complex menus, users will simply tell systems what they want—similar to interacting with ChatGPT or voice assistants today. Every enterprise tool may become prompt‑driven, with an AI agent executing tasks behind the scenes. 

It is a vision where: 

• Complex systems become conversational 
• Learning software becomes unnecessary 
• AI serves as the universal interface 

Parts of this future are already emerging in enterprise AI platforms and automation tools. 

Conclusion 

AI is reshaping the finance sector gradually but profoundly. Rather than replacing professionals wholesale, it is automating repetitive tasks, demanding new governance frameworks, driving new skillsets, and accelerating digital transformation. 

The journey will take time—five to ten years of sustained research, implementation, and cultural change—but the trajectory is clear: AI will become as fundamental to finance as the internet itself.