Asset Management
Mar 31, 2025
Using Data Reasoning For Portfolio Research & Construction
Creating an investment portfolio is an arduous process often filled with manual, repetitive tasks. Fortunately, modern portfolio managers (PMs) can leverage AI-powered data reasoning to expedite the security selection process, produce insights to generate investment performance, and much more. This article will explore why PMs stand to benefit immensely from AI-powered data reasoning.
What is Portfolio Research and Construction?
Portfolio research and construction is the process of analyzing and building a collection of investments to achieve specific financial goals while balancing risk and return. The exact portfolio construction process varies depending on the PM’s goals. But, it usually follows these basic steps:
Define Investment Objectives & Constraints: Establish the portfolio’s investment goals, risk tolerance, time horizon, and potential compliance or regulatory constraints.
Conduct Market & Macroeconomic Research: Factor in broader market sentiment, industry trends, and economic indicators like interest rates, inflation, or GDP growth.
Screen & Select Assets: Research assets based on quantitative or fundamental criteria. This usually involves filtering assets based on financial ratios or valuation metrics to identify securities that could be a good fit for the portfolio. This process could also include technical analysis to maximize asset entry points.
Construct the Portfolio & Allocate Assets: Purchase the assets in the proper quantities, implementing weighting strategies based on risk-return objectives.
Manage and Optimize the Portfolio: Continue to perform ongoing portfolio management activities like backtesting, scenario analysis, rebalancing, tax-loss harvesting, compliance, reporting, and much more.
PMs must complete the above tasks in a highly efficient manner, as the industry has both a high bar for success and a large opportunity for error. Even a small mistake or delay could be the difference between success or failure.
For example, PMs must constantly read and react to new information to make quick investment decisions and optimize their return. However, market-moving data is often sourced from a wide variety of data sets, including news media (Bloomberg, Seeking Alpha, etc.), regulatory reports (SEC filings or governmental updates), or company communications (earnings reports, press releases).
Even just aggregating this data is a highly time-consuming and error-prone process, let alone analyzing and responding to it. These are all areas where data reasoning can provide an immediate benefit.
What is Data Reasoning?
Data reasoning is the process of using artificial intelligence to analyze, interpret, and apply logic to data. Solutions like Gradient’s Finance Reasoning platform go one step further than most traditional artificial intelligence (AI) or business intelligence (BI) tools by understanding the why behind the what.
AI-powered data reasoning is emerging as a game-changing solution for the finance sector, capable of automating high-level processes and providing value to many financial verticals, from asset management to banking.
Continue reading to learn more about how data reasoning works.
Accelerating Portfolio Research & Construction: Expediting the Security Selection Process
The security selection process typically involves conducting extensive research on all companies that could be included in the portfolio. This involves tasks like:
Analyzing Fundamental Data: PMs often need to review balance sheets, income statements, and cash flow statements from 10-K and 10-Q reports to identify companies that should be included in the portfolio.
Searching For Qualitative Insights: Since numbers only share part of the story, PMs also need to manually parse for qualitative insights through additional documentation like earnings calls, investor presentations, and conference call transcripts.
Considering Broader Macroeconomic Conditions: Portfolio companies do not exist in a vacuum, which means that PMs must remain up-to-date on economic reports to consider how general market conditions or trends may impact their selected companies.
Gradient helps expedite these tasks by quickly analyzing data sources to provide PMs with the necessary takeaways – saving hours of manual effort spent searching for that information. This is possible thanks to Gradient’s ability to analyze unstructured data sets.
Most artificial or business intelligence tools require perfectly organized, structured data to deliver trusted insights. However, real-world data is almost never perfectly organized. It’s usually disorganized and comes in many different formats, which presents a number of challenges. Fortunately for asset managers, Gradient can easily digest raw, unstructured data sets while still generating reliable outputs.
Gradient’s Finance Reasoning Platform can quickly and reliably analyze data sources like:
Financial statements (balance sheets, cash flow, etc.)
Earnings call transcripts
Market commentary from sources like Bloomberg, Reuters, or Seeking Alpha
Regulatory filings and legal documents
Corporate communications or press releases
Social media posts
Instead of poring through thousands of pages of earnings transcripts each quarter, many portfolio managers leverage data reasoning to analyze these transcripts and report the key takeaways.
This automation allows portfolio managers to expedite the security selection process while ensuring that only the best companies are included.
Produce Insights to Generate Investment Performance
Data reasoning also plays a pivotal role in helping portfolio managers track market-moving updates to generate investment performance.
Data reasoning excels at analyzing vast amounts of data, uncovering trends, and delivering actionable insights to drive performance. This makes it an ideal solution for PMs who spend countless hours tuning into the news, inflation reports, Fed meeting minutes, company conference calls, or geopolitical updates in search of market-moving news.
Instead of trying to keep up with the nonstop barrage of information, many portfolio managers use data reasoning to monitor relevant information sources and automatically extract the updates with the biggest market-moving potential.
Gradient has already helped one US-based investment manager with $40 billion in AUM improve investment performance by using AI for market analysis.
Broadening the Universe of Securities PMs Can Choose From
Data reasoning can also help portfolio managers (PM) broaden the universe of names they plan on researching.
Under normal conditions, it would take the PM a substantial amount of time to analyze additional securities they may want to add to the portfolio. Each new stock requires hours of research to ensure that it meets the goals and criteria outlined by the PM, creating an immense bottleneck.
This manual process not only limits the number of companies that can be analyzed but also increases the potential for oversight and error, impacting the overall investment strategy's effectiveness and performance.
However, data reasoning can drastically reduce the time required to analyze securities by automatically filtering large databases for potential portfolio companies and surfacing securities that meet certain criteria. This allows the PM to place higher scrutiny on the securities that get added to the portfolio, which could potentially help them achieve better investment outcomes.
Accelerate Portfolio Construction With Gradient
Gradient’s Finance Reasoning Platform is a transformative solution for portfolio managers, capable of providing immense value at all phases of the research and construction process. A few – but not all – of data reasoning’s capabilities include:
Expediting the security selection process
Produce insights to generate investment performance
Broadening the universe of securities that PMs can choose from
FAQ: Data Reasoning For Portfolio Construction
How does data reasoning differ from traditional data analysis in portfolio management?
Traditional data analysis relies on structured datasets, while data reasoning can interpret both structured and unstructured data to extract key insights. It applies logic to messy, real-world data, helping portfolio managers make more informed investment decisions by automating tedious research tasks.
What types of unstructured data can AI-powered data reasoning analyze for investment insights?
AI-powered data reasoning can process unstructured sources like earnings call transcripts, regulatory filings, market commentary, press releases, social media, stock screeners, and more. Unlike traditional systems, it can extract key takeaways from messy, unformatted data.
How can portfolio managers use data reasoning to enhance security selection?
Portfolio managers can automate financial analysis, scan earnings reports for qualitative insights, and track macroeconomic trends. This reduces manual research time and helps them react to market-moving data faster.
Share
Tags
