How to "Win" Earnings Season with Generative AI
- BCMstrategy, Inc.
- Jun 5
- 6 min read
Earnings season arrives in July. Are you ready?
Equity analysts will face unprecedented pressures to provide solid guidance regarding corporate earnings prospects to drive a variety of portfolio investment strategies. Second quarter corporate books will close and corporate managers will brief analysts amid unprecedented policy volatility.
Key drivers of uncertainty regarding corporate outlooks at midyear 2025 will include:
Lack of clarity regarding US tariff policy
Significant shifts in bilateral and regional trade policies in Asia, Europe, Latin America, and Africa
Bruising U.S. fiscal policy battles that include rescinding a range of green subsidies
Difficult U.S. fiscal policy battles and brinksmanship regarding the debt ceiling
Fiscal pressures in Europe as the continent shifts to a war economy footing while continuing to execute a fast-track pivot away from fossil fuels
Ongoing if not intensifying tensions regarding access to the critical minerals that drive modern appliances as well as climate-friendly components for cars and energy generation.
Every equity analyst must now become a global macro expert. Global macro and fixed income analysts will be paying more attention than usual to corporate earnings calls in order to identify potential early warning indicators of growth trajectories for the second half of 2025 and beyond.
Dismissing public policy risks as random is both irresponsible and incorrect.
Those that view the current policy environment as delivering random, exogenous shocks from only one source are short-changing themselves and their customers because this mindset ignores factoring into the analysis the full range of reaction functions...many of which fail to generate headlines.
The volume, velocity, intensity, and volatility of the policy cycle will peak during summer 2025 just as earnings season hits high gear.
This post provides our key recommendations for optimal deployments that maximize analyst capacities to assess corporate prospects in light of rapidly changing public policy realities. Check out the video summary as well:
How To "Win" Earnings Season with Generative AI
Deploy Generative AI
If you have not yet done so already, start using an internal or third-party Generative AI solution to parse through the high volume of text-based content that drives capital market analysis.
These advanced language processes can sift through millions of words within earnings call transcripts, regulatory filings, market data, news stories, and other input sources to spot important earnings signals in a timely manner, eliminating many of the delays and inconsistencies that characterize manual processes.
Knowledge is power, particularly pricing power in capital markets. Faster analysis translates into real value and informational advantages for traders.
Deploying generative AI to stay on top of the earnings calls, news cycle, and other document releases is a necessary but not sufficient condition for successful earnings season analysis.
Unfortunately, generative AI solutions have not been trained to answer the key question that every equity analyst must answer during summer 2025: is management accurately reading the policy trend?
Never has it been more important to have access to solid policy intelligence. Spotting a technical opportunity can deliver the difference between survival and oblivion for many companies during 2025-26 in particular. Being able to compare in an instant whether an earnings call shows management is ready for the next 12-36 months will be crucial in finding companies that are making good choices amid a challenging environment.
Some concrete examples drawn from PolicyScope data during 2Q2025 help illustrate the point:
Does the company know that the President of the European Commission and the Secretary-General of the World Trade Organization publicly agreed with President Trump's criticism of the global trading system even when they disagreed with using unilateral reciprocal tariffs to address the issues? Knowing these facts can expand a company's strategic options and suggest a longer time horizon for trade policy negotiations beyond summer 2025.
Does the company know that a large number of tariffs imposed by the United States during 2Q2025 relied on traditional trade policy foundations rather than economic emergency powers? Understanding which tariffs are at risk for invalidation and which ones are likely to withstand legal challenge is crucial to formulating a resilient business strategy.
Does the company know that the UK, India, Japan, Australia, and the EU have accelerated bilateral free trade negotiations with selected partners and that the new Canadian government is prioritizing the creation of an internal free trade market across its own provinces?
If the only frame of reference for strategic analysis consists of news headlines that react to White House public comments, the analytical foundation is too narrow to provide potential investors with solid data-driven information to support investment decisions. Specific use cases include:
PolicyScope structured text training data provides the perfect add-on for analysts seeking to provide the best analysis possible during summer 2025 and beyond. GenerativeAI trained on PolicyScope data extends the depth and scope of earnings transcript analysis, identifies technical policy shifts important to corporate earnings, and can flag alignment/divergence with technical trading tools like MACD analysis.
Extend Traditional Equity and Market Analysis with Granular Public Policy Text Data
Traditional equity analysis performs technical analysis on a company’s financial statements and fundamental analysis regarding the company’s competitive position in its sector. Tax and public policy issues can impact these analytical functions, but usually at the margin.
Analysts evaluating the market risks associated with holding a security often incorporate a formal factor to account for country risk. However, the country risk premium is associated only with the country of incorporation. Risk premia are set in relation to the spread between the sovereign bond of the issuer’s home country compared with the benchmark 10-year U.S. Treasury bond which serves as a proxy for the risk-free rate.
Traditional market analysis is thus inadequate to price for potential instability, volatility, and risks associated with U.S. corporates much less other equities because traditional market analysis does not incorporate factors for specific policy issues that impact individual corporates. Nor does it provide the foundation to identify the three distinct reaction functions outside the public policy process: media cycle reaction functions, market reaction functions.
PolicyScope quantitative data was specifically designed to support granular market risk analysis by enabling markets to set prices for discrete policy risks. PolicyScope data is also structured to support dynamic reaction function analysis by a wide range of actors both inside and outside policy formation processes.
Training generative AI solutions on PolicyScope data thus enables investment analysts to access precision-crafted generative AI solutions on a par with the precision provided from training runs on prior earnings transcripts and corporate filings.
The PolicyScope dataset in machine-readable .JSON format extends back to 2006 and refreshes twice a day on a going-forward basis.
Every day, twice a day, our award-winning, patented process scours the globe for policy shifts important to capital markets. We measure, tag, and store the text. The inputs are Golden Source. The quantitative outputs and other metadata tags are also Golden Source because they are normatively neutral. In addition, the tagging structure creates a robust expert-driven ontology even as the quantitative values provide an additional measure of grounding for Knowledge Graphs that support Retrieval Augmented Generation processes.
In other words: PolicyScope training data for generative AI solutions delivers the premier, definitive training data cover the policy issues that drive price action today regarding monetary policy, trade policy, digital currency/stablecoins policy, energy policy and climate-related policy.
It is the ideal training data for investment analysts seeking to optimize their existing generative AI solutions to address public policy risks.
Deploy PolicyScope Text Data in a Knowledge Graph
Knowledge Graphs provide grounding and context for Retrieval Augmented Generation Processes. By tokenizing the relationships among words through correlation coefficients, knowledge graphs decrease the incidence of hallucination and increase the accuracy of generative AI outputs.
PolicyScope data arrives to customers immediately ready for deployment into knowledge graphs. Our award-winning, patented process delivers more than a proprietary ontology based on decades of leadership service within public policy on the global stage. The quantification layer embedded within the .JSON objects provides immediate context to language models, facilitating their capacity to calculate correlation coefficients across words.
The quantification layer also increases the operational efficiency of deployment into generative AI processes and the subsequent training runs because your agentic AI does not have to guess and learn which concepts are related to each other and your human team does not have to allocate hours to data preparation.
It is no secret that generative AI training runs are far from cheap. They cost real capital and real time to implement. The cost savings associated with deploying PolicyScope data are thus real
Get started today in order to be ready for the marathon that awaits in July when earnings season begins!
