Our Technology Leadership in Investment Decision-Making

Imagine a future where investment decisions are tailored to fit investor data and beliefs with analytic rigor of financial quant but explained in prose. In this world, you are not looking at tabular data or long reports where investment thesis assumptions are buried. In the Growthsphere system of investment decision-making, tabular and textual data is…

Written by

Dr. Christopher Thomas

Published on


Imagine a future where investment decisions are tailored to fit investor data and beliefs with analytic rigor of financial quant but explained in prose. In this world, you are not looking at tabular data or long reports where investment thesis assumptions are buried. In the Growthsphere system of investment decision-making, tabular and textual data is mapped to knowledge and converted to artifacts like memos and reports using your insights and assumptions.

At the core of our vision is the commitment to tailor the investment decision-making process to the unique needs and idiosyncrasies of each client. This involves enabling clients to build an assumption library—a repository of their analyses, reasoning, recommendations, and decisions. Our AI leverages these assumptions to derive bespoke analyses, recommendations, and decisions, ensuring that every piece of advice is closely aligned with the client’s beliefs and expectations.

Moreover, we’re dedicated to reducing decision-making biases and errors. Our AI does this by critically evaluating the assumptions and underlying data, highlighting potential issues, and suggesting alternatives. This collaborative approach between AI and human analysts aims to enhance the latter’s capabilities, transforming them into super-analysts who can make informed decisions with a holistic view of relevant data, enriched by their own creativity and experience.

Technology Driving Innovation

The technological backbone of our initiative is underpinned by the cutting-edge capabilities of large language models (LLMs) like Llama2 and Mistral, which have democratized access to sophisticated AI tools. Our tech stack is robust, incorporating both commercial and custom-tuned open-source models, enriched with assumption libraries and grounded in comprehensive ontologies and knowledge graphs.

Advanced Model Training

We are advancing the training of LLMs to achieve a profound understanding not only of the finance domain at large but also of the unique characteristics and needs of each individual client. This involves not just programming our models with general financial knowledge but also customizing them to integrate the specific beliefs, strategies, and preferences of each investor. By doing so, we create a personalized AI advisor capable of providing insights that resonate deeply with the client’s own perspective and objectives.

Integration of LLMs with Ontologies and Knowledge Graphs

The synergy between LLMs and our ontologies allows us to offer solutions that reason with greater depth and fewer errors compared to those powered by LLMs alone. These ontologies and knowledge graphs encode essential domain knowledge and interconnections within financial concepts, enabling our AI to navigate complex reasoning paths that would otherwise be out of reach for standard AI applications. This integration enhances the reliability and accuracy of the AI’s analyses and recommendations, fostering greater confidence among users.

Human-First AI Implementation

Implementing a human-first approach while heavily utilizing AI is a significant challenge. Our goal is to make the AI’s operations not only effective but also transparent and understandable to human analysts. The AI is designed to adopt the worldview and reasoning methodologies of its human users, making its processes and conclusions easy to follow, explain, and verify. This transparency is crucial for analysts who rely on AI to make informed decisions, ensuring they can trust and effectively interpret the AI’s output. Our approach emphasizes the augmentation of human decision-making, enhancing rather than replacing the intuitive and creative capacities of human analysts.

Through these innovations, we are setting a new standard for how technology can enhance financial analysis, combining the precision and analytical power of AI with the nuanced understanding and strategic insight of human expertise.

Challenges and Solutions

Integrating LLMs into financial analysis presents unique challenges, especially when dealing with complex structured data, customer-specific data, and the need for personalized recommendations. 

Our method involves a dual approach: leveraging AI’s broad analytical capabilities while honing in on the specifics that define the financial domain and the individual investor.  

By developing domain-specific ontologies and ensuring personalized, client-specific model training, we provide AI-driven solutions that are not only financially sound but also deeply relevant to each client’s financial context, contributing to a future where financial decisions are increasingly strategic, informed, and personalized.

Data Isolation and Historical Analysis

Incorporating LLMs into financial analysis brings its own set of challenges, notably the need for significant computational resources and a variety of data. The imperative to maintain each institutional investor’s data as a separate entity, to safeguard confidentiality, results in smaller, distinct datasets for fine-tuning each model. This data isolation is a critical step in ensuring personalized and secure financial analysis.

Moreover, a key obstacle is the ability to generate actionable insights from the history of investment outcomes. Our AI is designed to identify trends and patterns in past investments, leveraging these insights to enhance future strategy development. This capability enables us to not only learn from previous successes and failures but also to apply these lessons in crafting more informed and strategic investment decisions moving forward.

Conclusion: A Future Powered by AI and Human Synergy

By combining the holistic data analysis capabilities of AI with the creativity, experience, and intuition of human analysts, we’re not just providing investment advice; we’re redefining what it means to be a financial analyst in the modern world. Our technology, vision, and approach aim to create a future where investment strategies are smarter, more personalized, and aligned with sustainable, long-term goals—benefitting everyone from individual investors to the broader global economy.

Thank You and Looking Ahead

As we officially launch our company and step into the light after a year of hard work behind the scenes, we’re more committed than ever to our vision to reshape the landscape of investment decision-making, making it more intelligent, sustainable, and inclusive. Our journey has just begun, and we’re excited about the possibilities that lie ahead.