Developing Climate-Friendly Investment Strategies: A Use Case how Asset Managers approach it with AI tools

 Morningstar is providing you with their document, "The Asset Manager’s Guide to Climate Reporting," which aims to guide asset managers through the evolving landscape of climate-related reporting. It highlights the increasing pressure from regulations, investor preferences, and the accelerating climate crisis, emphasising the need for asset managers to transparently disclose their climate considerations and align their portfolios with net-zero goals.

As a prompt engineer of this fund company, how can you support asset managers in addressing these questions by leveraging AI models?

Here are three concrete examples of how I would approach this task:

Integrating climate risk assessment and reporting into investment processes and decision-making frameworks:

What to do next: Develop an AI-powered tool that can automatically extract and analyze climate-related data from publicly available sources, such as company reports, sustainability disclosures, and regulatory filings. This tool should be able to identify and categorize relevant information according to the TCFD recommendations and other applicable regulations.

Data needed: Official company reports, sustainability disclosures, regulatory filings, and other publicly accessible data sources.

Prompt engineering technique: Use a combination of keyword-based prompts and semantic similarity techniques to guide the AI model in identifying and extracting relevant climate-related information from the data sources. For example:

Prompt: "Extract information related to the company's governance of climate-related risks and opportunities from the following text: [input text]. Categorize the information according to the TCFD recommendations."

The AI model would then identify and categorize the relevant information based on predefined keywords and semantic similarity to the TCFD categories.

Expected outcome: An automated tool that streamlines the process of integrating climate risk assessment and reporting into investment processes, saving time and resources for asset managers while ensuring compliance with regulatory requirements.

Identifying reliable and comprehensive data sources and tools for measuring and reporting on climate-related metrics:

What to do next: Create a knowledge base of reliable and comprehensive data sources and tools for measuring and reporting on carbon footprint, transition risks, physical risks, and climate-related opportunities. Use AI models to continuously update and expand this knowledge base based on new information and user feedback.

Data needed: Publicly available information on climate-related data sources and tools, including their methodologies, coverage, and user reviews.

Prompt engineering technique: Use a question-answering prompt to guide the AI model in providing relevant information from the knowledge base. For example:

Prompt: "What are the most reliable and comprehensive data sources for measuring the carbon footprint of a portfolio? Provide a brief description of each source and its key features."

The AI model would then search the knowledge base and provide a structured response based on the available information.

Expected outcome: A continuously updated knowledge base that helps asset managers identify the most suitable data sources and tools for their specific needs, ensuring accurate and consistent climate-related reporting.

Leveraging climate-related data and insights to develop innovative investment products and strategies:

What to do next: Develop an AI-powered recommendation engine that suggests potential climate-friendly investment opportunities and strategies based on a combination of climate-related data, market trends, and client preferences.

Data needed: Climate-related data (e.g., carbon footprint, transition risks, physical risks, and opportunities), market data, and anonymized client preference data (if available).

Prompt engineering technique: Use a combination of data-driven prompts and conditional language to guide the AI model in generating personalized investment recommendations. For example:

Prompt: "Based on the following climate-related data and market trends: [input data], recommend three climate-friendly investment opportunities for a client with a moderate risk tolerance and a preference for companies with strong ESG performance."

The AI model would then analyze the input data and generate personalized recommendations based on predefined criteria and conditions.

Expected outcome: An AI-powered recommendation engine that helps asset managers identify and capitalize on climate-friendly investment opportunities, tailored to the specific needs and preferences of their clients.

By implementing these AI-powered solutions, a prompt engineer can support asset managers in effectively addressing the key questions related to climate risk assessment, reporting, and investment strategy development, ultimately enabling them to make more informed decisions and meet the growing demand for climate-friendly investments.