In the rapidly evolving world of financial services, adopting advanced technologies such as generative AI is not just about keeping pace, it’s about staying ahead of the competition.
Recently, we worked with a leading financial services institution to explore how generative AI could reshape the investment experience, making it smarter, simpler, and more personalised. Our goal was clear: to turn complex data into actionable insights that empowers investors.
Here’s a behind-the-scenes look at our journey, the challenges we faced, and the innovations that followed.
The mission: elevating investing with AI
Our client, a major player in the personal investing space, wanted to differentiate themselves by offering advanced tools that empower investors. With competition intensifying, they needed a fresh approach to deliver value. Generative AI offered that opportunity—by transforming how investment information is consumed and used.
We focused on two primary use cases:
- Stock announcement summaries: Financial reports and stock announcements are packed with crucial information, but their length and complexity can overwhelm even the most seasoned investors. Our task was to harness AI to generate concise summaries that distil these reports into key points, making it easier for investors to access the information they need, quickly and confidently.
- Personalised portfolio insights: Beyond general market updates, we aimed to provide insights that were directly relevant to each investor’s portfolio. Imagine an AI tool that knows your specific holdings and serves up the most relevant news, trends, and opportunities. This approach was all about enhancing decision-making by delivering tailored, actionable insights.
The challenge: bridging technology and business needs
Implementing AI in a highly regulated financial services environment comes with its own set of challenges, including compliance, security, and integration with existing systems. It wasn’t just about developing a technically sound solution; it had to align with strict regulatory requirements and integrate seamlessly into the client’s ecosystem.
Iterative development and prompt engineering
Building effective AI summaries required a disciplined, iterative approach. We began with foundational prompts and refined them through multiple iterations based on continuous feedback. Early versions of the AI outputs were sometimes too broad or missed critical details, but by embracing a cycle of constant testing and improvement, we honed the AI’s performance.
One particularly impactful example was a financial report from a large tech company. Our initial attempts produced summaries that were too generic, but after several rounds of adjustments, the AI delivered focused, relevant insights that highlighted key financial metrics and strategic points. This iterative process demonstrated the power of refining AI models to achieve high-quality results.
Balancing technical feasibility and business viability
Throughout the project, we maintained a dual focus on technical feasibility and business viability. While AI offers vast capabilities, it was essential to ensure our solutions made practical business sense. For instance, personalised portfolio insights introduced challenges related to cost and performance when scaled across millions of users. By adjusting our approach to generate broader insights that could be reused, we struck a balance between delivering value and managing resource constraints.
The big showcase: demonstrating value
After weeks of development and testing, we showcased our AI solutions to a broad group of stakeholders within the organisation. The response was overwhelmingly positive. The AI’s capability to transform dense financial documents into actionable summaries resonated well, and there was clear enthusiasm for how these tools could be integrated into the client’s existing offerings.
Uncovering new use cases and future potential
User testing insights also sparked discussions about additional applications of the AI solution. The institution’s customer service team, which often handles inquiries from investors needing help to interpret stock announcements, saw potential in using AI-generated summaries as quick reference tools—demonstrating the value of undertaking user research to fully understand the desirability potential of the AI solution being built.
Key takeaways: strategic insights from the project
- Iterate and learn: The iterative development approach was critical to our success. By continually testing and refining the user experience, we were able to enhance the AI’s accuracy and relevance, aligning closely with both user needs and business objectives.
- Technical feasibility meets business strategy: Ensuring that our solutions were not just technically robust but also strategically aligned with business goals was key. Balancing cost, performance, and scalability allowed us to deliver a solution that was both innovative and practical for the client.
- Scalable innovation: Starting with targeted use cases and proving their value enabled us to build a strong foundation for future expansion. This approach not only built stakeholder confidence but also positioned the client to scale these innovations effectively across their broader operations.
What’s next?
This project demonstrated the significant potential of generative AI in enhancing the investment experience by simplifying complex information and providing targeted insights. The success of our AI solutions has also opened the door to further applications across different areas within the organisation.
Looking ahead, our focus will be on scaling these AI-driven capabilities, optimising performance, and ensuring alignment with regulatory standards and business goals. By continuing to refine and expand on these initial successes, we aim to further integrate AI into the core of the investment process, making it more efficient and impactful.