UX Research · 1:1 Interviews · Thematic Analysis · 2024

The Impact of AI on Foreign Exchange Traders at Work

Exploring how AI can enhance the workflow and decision-making of foreign exchange traders by combining financial expertise with user-centered design insights.

AI and Foreign Exchange Traders research

Type

Sponsored Project with Bentley Hughey Center for Financial Services

Role

UX Researcher

Team

1 Project Manager, 2 UX Researchers

Date

09/13/2024 – 12/12/2024

Description

I served as a UX Researcher, responsible for deciding on the research approach and methods to first understand the life of a foreign exchange trader. I was also in charge of recruiting foreign exchange traders, conducting usability testing, and creating participant interview scripts for my team to follow.

Through this collaboration between Bentley's Human Factors in Information Design Program and the Trading Room, our team aimed to understand the FX trader's user experience, focusing on their workflow and exploring how AI can improve their routines.

Skills

Moderating & Note-takingUsability TestingPersonaJourney MapThematic AnalysisAffinity Mapping

TL;DR 🧾

Problem Context

Foreign exchange traders work in high-pressure environments where they rely on a combination of technical analysis, intuition, and real-time data to make decisions. However, the overwhelming amount of data and the challenge of navigating multiple platforms often hinder decision-making efficiency.

Through the collaboration between Bentley's Human Factors in Information Design Program and the Trading Room, our team aimed to understand the foreign exchange trader's user experience, focusing on their workflow and exploring how AI can improve their routines.

Traders Interviewed

Both institutional and individual FX traders across varied markets and experience levels

AI Design Recommendations

Surfaced to better align the Bentley Trading Room with how real FX traders work

Context

Expected Business Impact

This project begins the HFID program's collaboration with the Bentley Trading Room. Our goal was to align the trading room's user experience with real-world foreign exchange practices.

By understanding how FX traders actually work — their tools, decision-making processes, and pain points — we could recommend meaningful improvements and strengthen the HFID program's partnership with the Bentley Trading Room.

📈 Align with real-world FX practices

Align the trading room's user experience with real-world foreign exchange practices to make it more authentic.

🎯 Recommend UX improvements

Recommend improvements to the Bentley Trading Room's user experience based on foreign exchange traders' work routines.

🤝 Strengthen the HFID partnership

Create a realistic, immersive trading environment for students by grounding the curriculum in real trader insights.

Research Questions

What we set out to understand

To better understand key pain points and opportunities for improvement in foreign exchange trader workflow and decision-making.

Our research questions aim to uncover the key activities, challenges, and emotional experiences traders encounter in their work and how they utilize technology and data to support their decision-making.

Workflow & Activities

What does a typical trading day look like — tools, routines, and workflows?

Pain Points

What are the biggest challenges and friction points traders face in real-time decisions?

Technology Use

How do traders currently use technology and data to support their decisions?

AI Attitudes

What is their perception of AI tools, and how open are they to AI-assisted trading?

Process

How we got there

Approach

Our Approach

Qualitative Research Method: 1:1 Interview

I chose this method because it allows our group to gather in-depth insights by asking questions to explore traders' decision-making processes. A one-on-one interview also allows us to ask follow-up questions to deepen our understanding of the varied experiences of traders and clarify their responses.

To understand how we could improve foreign exchange traders' workflow and what technology could help with their decision-making process, we found experienced foreign exchange traders and gathered insights through interviews.

Foreign exchange traders' experiences, workflows, and tools in their day-to-day activities

Understanding the daily activities of FX traders helps us better understand the difficulties they encounter, the resources they use, and how they interact with different data sources.

Technology, data usage, and views on AI tools in decision-making

By researching the technology and data traders use, we can identify the most trusted tools, the types of data used in decision-making, and whether the technology supports their needs.

Methodology

Usability Testing — Participant Screening

Both types of foreign exchange traders are important to explore — individual traders use their own capital while institutional traders rely on firm capital. Their day-to-day work varies due to differences in work environments, resources, and access to data and technology. This approach helped us understand how AI technology could better integrate into their decision-making processes.

Recruitment Process

UserTesting

We chose this platform because we can easily screen for a specific number of foreign exchange traders with 3+ years of experience, ensuring they have varied expertise across different markets and are open to sharing their responsibilities.

Alumni Network

Jay, founding director of the Bentley Hughey Center for Financial Services/Trading Room, introduced us to Lucy Slowe, a Bentley alumna currently working as an institutional FX trader at State Street Bank, bringing invaluable real-world experience.

Methods

Research Methods Question

Coming from a design background, I don't have a deep understanding of what a foreign exchange trader specifically does. I did some of my own research through watching videos to understand FX trading before writing the interview script.

Goal: To prepare for user interviews, I created an interview script for my group with questions to understand foreign exchange traders — how the two types differ in workflow, resources, background, decision-making, and their use of AI.

Why do I need to ask these types of questions?

After observing these experiences and learning more about foreign exchange traders, it was time to get to work on consolidating our data and providing helpful recommendations for improvement to the forex traders.

Analysis

Analysis and Report Toolkit

While UserTesting automatically recorded the meeting, I also took notes on a Google Doc to capture emotions, reactions, and key phrases that stood out as I wrote them down. I was multitasking during the interview instead of just reading from the script.

Data Analysis Process

After each of us finished our user interviews, we met via Zoom to share insights gathered from both institutional and individual Forex Traders to a FigJam table, which included details like company, specialization, work experience, tools, decision-making, challenges, and thoughts on AI.

Table: Individual Foreign Exchange Trader Interview Insights

Individual FX trader interview insights table

Table: Institutional Foreign Exchange Trader Interview Insights

Insights

From Table to Key Insights

After hearing each insight gathered from each participant, we identified common themes and extracted insights for both institutional and individual traders.

Institutional FX Traders

1.Heavy Reliance on Real-Time Data & News for Trading Decisions

Institutional insight 1 - real-time data reliance

2.Risk Management and Trading Strategies Are Central to Daily Operations

Institutional insight 2 - risk management

3.Emotion and Psychological Pressure in Trading

Institutional insight 3 - emotion and pressure

4.Integration of AI and Automation in Trading Operations

Institutional insight 4 - AI integration

Individual FX Traders

1.Balancing Trading with Personal Life or Full-Time Jobs is a Key Challenge

Individual insight 1 - work-life balance

2.Challenges with Information Latency and Market Timing

Individual insight 2 - information latency

3.Emotional Control and Discipline are Key Challenges

Individual insight 3 - emotional control

4.Integration of AI and Automation in Trading Operations

Individual insight 4 - AI and automation

Deliverables

Key Insights Into Persona

Based on our participant interviews and the common themes we identified during our discussions, two main personas emerged. Each persona includes a brief bio detailing their work experience and areas of specialization, along with a summary of their daily workflow. I incorporated personality traits to provide a deeper understanding of their character and how their work shapes their personality.

Persona: Institutional Trader

Persona: Individual Trader

Journey Mapping

Personas to User Journey Map

With a better understanding of the two personas and insights from participant interviews, we proceeded to create a journey map showing the various stages of both individual and institutional foreign exchange traders' daily workflows.

User Journey Map: Institutional Trader

In general, they went through 4 main stages: gathering information, using that information to build strategy with their team, executing that strategy using specialized trading software, and monitoring the trade and adjusting the strategy as needed.

User Journey Map: Individual Trader

In general, they went through four main stages, different from institutional traders. They started by checking market movements, then analyzed past trends and developed strategies. In the execution phase, traders place and adjust trades using platforms like MetaTrader. In the monitoring phase, traders evaluate performance, refine strategies, and manage stress.

Recommendations

AI Recommendations 💡

Based on interview insights, I surfaced six AI design recommendations to better align tools with how FX traders actually work and make decisions.

Confidence Score Display

Show a real-time confidence level for each AI-generated suggestion to help traders assess reliability at a glance.

Verified Sources

Add an expandable section that explains why the AI is suggesting a specific action ("based on historical volatility patterns / this source").

Adaptive Alerts

Introduce smart alerts that learn a trader's preferences and trading style, reducing unnecessary noise and surfacing high-priority signals.

Simulation Mode

Let users test AI-driven recommendations in a risk-free "what-if" environment, enabling learning and exploration before committing to real trades.

Human Override Logging

Allow users to override AI suggestions and annotate why, then use this data to improve AI feedback loops and create a record for performance review.

Decision Timeline Visualization

Provide a timeline view that overlays AI suggestions, user actions, and market data, helping traders reflect on outcomes and learn from past decisions.

Future Directions

Next Steps 🔭

Balancing Efficiency and Skepticism in AI

AI and automation are being used to improve trading efficiency, with most effort put into improving data analysis and speeding up trade execution.
For now, institutional FX traders view AI with a degree of skepticism about reliability, preferring "human in the loop" due to the sensitivity of trading decisions.

Interview Trading Room Managers and Students

Interview students to understand how the current curriculum aligns with industry practice.
Explore opportunities to better prepare students for real-world challenges by aligning the curriculum with the evolving needs of the trading industry, especially in terms of AI integration.

Explore Other Finance Positions

Conduct additional research on various finance positions to uncover further opportunities for aligning education with industry expectations and addressing the broader impact of AI and automation across different sectors.

Takeaways

Reflection 🪞

What I learned

Throughout this project, I dedicated significant effort and gained valuable insights, notably by taking a role as a UX Researcher. I moderated 3 user testing sessions and created the user interview script.

Combining observational research and interviews

Combining interviews with observations helped us better understand how traders make quick decisions, especially when it's hard to explain things verbally.

Screener specificity matters

I found in our first round of recruitment that it was much easier to recruit individual traders than institutional traders. As a result, I advised the team to run a second round of research with more specific screener questions to better balance the participant pool.

Domain learning is part of the research process

Coming from a design background without deep finance knowledge pushed me to do background research before writing the interview guide. This made my questions more credible and helped participants engage more openly in the sessions.