Generative AI and New Business Models: How AI Can Help Predict Business Markets and Guide Investment Decisions

# Generative AI and New Business Models: How AI Can Help Predict Business Markets and Guide Investment Decisions

**Written By Seyed Amir Hossein Marashi Pour (Amir Marashi)**

MBA Candidate, University of Technology Sydney (UTS)

Founder – Elite Sports and Business Solutions (ESBS)

📅 Published: 28 October 2025

🌐 https://www.amirmarashi.com

Generative AI and New Business Models: How AI Can Help Predict Business Markets and Guide Investment Decisions

Date: 28/10/2025

Article

Abstract

Generative artificial intelligence (GenAI) is altering not only how firms create value but also how they allocate capital. Building on business‐model innovation and dynamic capabilities theory, this paper develops a managerial, non‑technical framework that explains how GenAI can (1) expand market foresight by extracting signals from unstructured data, (2) improve decision support through scenario generation and narrative synthesis, and (3) enable business‑model innovation in investment intelligence (e.g., “prediction‑as‑a‑service,” agentic advisory). We integrate fresh evidence from finance and macroeconomics—domain‑specific LLMs for markets, stock‑return predictability from LLM news interpretation, and macro nowcasting gains from LLM‑based sentiment—to derive testable propositions about where GenAI improves investment choices (e.g., digital assets, fast‑moving technology equities) and under what conditions (data availability, volatility, governance). The contribution is a conceptual, business‑major‑friendly roadmap for researchers and practitioners to evaluate GenAI’s role in predicting markets (e.g., Bitcoin, sector ETFs) and guiding capital allocation—while acknowledging ethical and governance risks. European Central Bank+3arXiv+3arXiv+3

Keywords: generative AI; business‑model innovation; investment decision‑making; forecasting; digital assets; dynamic capabilities

1. Introduction

Across leading business journals and conferences, editors are explicitly prioritizing research on AI for finance and business decisions, underscoring GenAI’s salience for investment strategy. INFORMS Pubs Online+2INFORMS Pubs Online+2 Recent market evidence suggests that domain‑specific language models (e.g., BloombergGPT trained on finance corpora) and general LLMs applied to financial text can add predictive content for asset returns and macro nowcasts. European Central Bank+3arXiv+3arXiv+3Research question: How can generative AI enable investors and businesses to identify the most promising market opportunities earlier, cheaper, and with less bias—without requiring coding?

2. Theoretical background

2.1 Business‑model innovation (BMI) with GenAI

A recent systematic review shows AI‑driven BMI research is expanding yet fragmented across units of analysis and mechanisms; GenAI appears to accelerate novel value propositions and revenue logics. (Jorzik et al., 2024). ScienceDirect We anchor the analysis in the Business Model Canvas (Osterwalder & Pigneur, 2010) to trace how GenAI affects value creation, delivery, and capture—especially the value proposition, channels, and revenue streams. Strategyzer

2.2 Dynamic capabilities and sensing for investment opportunities

Dynamic capabilities theory posits that superior sensing, seizing, and transforming under uncertainty drives superior performance. GenAI can enlarge the sensing aperture by converting unstructured text (news, filings, social media) into forward‑looking signals that inform seizing (capital allocation). (Teece, Pisano, & Shuen, 1997; Teece, 2020). josephmahoney.web.illinois.edu+1

3. A business‑centric framework (no coding)

Pillar	Managerial Objective	GenAI Role / Mechanism	Illustrative Examples and Evidence
1. Market Foresight	Detect investable trends early and anticipate sector shifts	• Convert unstructured text (news, PMI reports, earnings calls, policy statements) into structured foresight signals.• Generate scenario narratives (“optimistic / neutral / pessimistic”) for strategic planning.	• LLM-based interpretation of Purchasing Managers’ Index (PMI) narratives improves GDP nowcasts (European Central Bank, 2025).• Headline sentiment scores from LLMs predict cross-sectional stock returns (Lopez-Lira & Tang, 2024).
2. Decision Support	Compare investment alternatives, assess risk, and improve communication with stakeholders	• Use GenAI to synthesize due-diligence memos and management briefings.• Automate generation of multi-scenario reports with embedded probabilistic reasoning.• Support human-in-the-loop decision-making through narrative analysis dashboards.	• Big-4 consultancies (e.g., PwC 2024) embed ChatGPT Enterprise into client decision workflows for financial strategy.• Enterprises leverage GenAI for real-time ESG risk reporting and strategic forecasting.
3. Business-Model Innovation (BMI) in Investment Intelligence	Monetize predictive capabilities and re-design service models	• Enable new revenue logics such as Prediction-as-a-Service, on-chain analytics platforms, and agentic advisory systems.• Automate market insight delivery to investors and corporates.• Integrate data flywheels and user feedback loops into AI-driven financial products.	• On-chain analytics firms (e.g., Glassnode, CryptoQuant) commercialize narrative and sentiment data for digital-asset markets.• Fintechs and consulting alliances (PwC–OpenAI) demonstrate scalable deployment of AI-based market intelligence services.

4. Fresh empirical signals (what we know)

1. Finance‑domain LLMs. BloombergGPT, a 50B‑parameter model trained on a finance‑heavy corpus, demonstrates superior performance on finance tasks—evidence that specialized corpora matter for decision support. arXiv

2. Asset‑return prediction from news. LLM assessments of headlines significantly predict next‑day stock returns out‑of‑sample, outperforming traditional methods—suggesting incremental information extraction. (Lopez‑Lira & Tang, 2024). arXiv+1

3. Macro nowcasting with LLMs. ECB economists show that a ChatGPT‑derived activity score from PMI narratives improves euro‑area GDP nowcasts relative to strong benchmarks—using only a few pages of text. European Central Bank+1

4. Enterprise adoption. Large consultancies are institutionalizing GenAI for strategy/finance use cases (first reseller and largest enterprise deployment of ChatGPT Enterprise), signaling managerial relevance rather than only technical promise. PwC+1

5. Propositions for business research

• P1 (Foresight advantage). Firms that deploy GenAI to parse unstructured market text will identify investable themes (e.g., AI supply chain, green tech) earlier than comparable firms, yielding superior timing of capital allocations; the advantage is larger when traditional quantitative signals are weak. (Builds on ECB and LLM‑news evidence.) European Central Bank+2Reuters+2

• P2 (Digital‑asset salience). GenAI‑augmented sentiment and on‑chain narrative extraction improves allocation decisions more in high‑volatility, retail‑attention markets (e.g., Bitcoin/altcoins) than in low‑volatility bond markets, because information is text‑rich and rapidly evolving. (Motivated by on‑chain analytics business models.) Glassnode Insights+1

• P3 (Cost–speed trade‑off). Relative to traditional research, GenAI reduces the marginal cost and cycle time of generating investment theses (e.g., sector or token narratives), enabling more experiments per dollar; effectiveness depends on domain adaptation (finance‑specific corpora). (BloombergGPT evidence). arXiv

• P4 (Complementarity with capabilities). The performance impact of GenAI on investment outcomes is moderated by dynamic capabilities—notably data governance, human‑AI decision rules, and the ability to pivot business models. (Dynamic capabilities theory). josephmahoney.web.illinois.edu

• P5 (Governance constraint). Without robust AI governance (accuracy checks, audit trails), GenAI‑supported investment processes exhibit higher model risk and pro‑cyclical herding in sentiment‑driven markets; governance quality moderates realized alpha. (MISQ editorial concerns). AIS eLibrary+1

6. Research design guidance (business‑major friendly; no coding)

Designs.

• Event‑study with narrative treatment: Use GenAI to classify news (positive/negative/uncertain) around firm or token events; test abnormal returns relative to controls (cf. headline‑based LLM studies). arXiv

• Nowcasting/forecast comparison: Compare managerially relevant forecasts (e.g., sector sales growth, GDP‑sensitive ETFs) with/without GenAI‑based sentiment extracted from PMI narratives. European Central Bank

• Field evidence from enterprise rollouts: Study how GenAI deployment inside advisory firms (e.g., ChatGPT Enterprise at scale) changes client investment workflow cycle times and decision quality. PwC

Data sources (no programming required).

• Text: PMI reports, earnings call transcripts, central‑bank communications, major financial news.

• Market: Public equity/ETF prices; digital‑asset on‑chain analytics from institutional providers (illustrative: Glassnode; CryptoQuant). Glassnode Insights+1

Outcome metrics.

• Forecast accuracy (e.g., MAPE for nowcasts), timeliness/latency of insights, decision throughput (ideas vetted per month), and economic value (risk‑adjusted returns/net alpha for policy portfolios).

7. Illustrative cases (conceptual, not prescriptive)

• Digital coins (e.g., Bitcoin). A fund integrates GenAI‑summarized on‑chain narratives and macro text (PMI, policy speeches) to set entry/exit corridors; managers evaluate if the GenAI narrative signal adds value over a moving‑average baseline—consistent with the ECB and headline‑prediction evidence. European Central Bank+2Reuters+2

• Technology equities. An investor uses GenAI to synthesize semiconductor supply‑chain news and earnings commentary, producing scenario memos (optimistic/neutral/pessimistic) that feed investment committees. The enterprise pattern mirrors large‑scale deployments in consulting. PwC

(These are research/teaching illustrations, not investment advice.)

8. Managerial implications

1. Invest in capabilities, not just tools. Domain adaptation (finance corpora), human‑in‑the‑loop review, and governance determine ROI. arXiv

2. Use GenAI for narrative advantage. Markets with noisy, fast‑moving text (crypto, high‑growth tech) are fertile terrain for GenAI‑enabled sensing—provided oversight is rigorous. Glassnode Insights+1

3. Modernize the business model. Consider “prediction‑as‑a‑service” and agentic advisory models while tracking cost/benefit as enterprise deployments spread. Reuters

9. Ethics and governance

Leading IS/management outlets caution against over‑reliance, hallucination, and poor attribution in AI‑assisted work, reinforcing the need for auditability, data provenance, and model‑risk controls in investment settings. AIS eLibrary+1

10. Conclusion

GenAI can improve market foresight, decision support, and business‑model innovation for investment—particularly where signals reside in unstructured text and where volatility rewards early sensing. The agenda above converts recent empirical advances into a researchable, practice‑relevant program suitable for the highest‑tier business journals.

References

• AI & Finance evidenceBloomberg, S. et al. (2023/2024). BloombergGPT: A Large Language Model for Finance. arXiv. (Domain‑specific LLM for finance). arXivLopez‑Lira, A., & Tang, Y. (2024). Can ChatGPT Forecast Stock Price Movements? University of Florida working paper; arXiv/SSRN versions. (LLM scores predict returns from news). arXiv+1European Central Bank (Sun, 2025). Enhancing GDP nowcasts with ChatGPT: a novel application using PMI news releases. ECB Working Paper 3063; Reuters coverage. European Central Bank+1Nie, Y. (2024). A Survey of Large Language Models for Financial Applications. arXiv/RePEc. (Survey of LLMs in finance). IDEAS/RePEc

• Business‑model innovation & strategyJorzik, P., et al. (2024). “AI‑driven business‑model innovation: a systematic review.” Industrial Marketing Management. (State of the literature on AI‑BMI). ScienceDirectOsterwalder, A., & Pigneur, Y. (2010). Business Model Generation. Wiley. (Business Model Canvas). StrategyzerTeece, D. J., Pisano, G., & Shuen, A. (1997). “Dynamic capabilities and strategic management.” Strategic Management Journal, 18(7), 509–533; see also Teece (2020). josephmahoney.web.illinois.edu

• Editorials and calls (journal context)Ata, B., Cong, W., Giesecke, K., Sun, P., & Teo, C. P. (2024). “Call for Papers—Management Science Virtual Special Issue on AI for Finance and Business Decisions.” Management Science, 70(10). INFORMS Pubs Online+1MIS Quarterly Executive (2025). “Evolving thoughts on Generative AI, writing, research and MISQE.” AIS eLibrary

• Enterprise adoption and market infrastructurePwC (2024). “PwC becomes OpenAI’s first reseller and largest user of ChatGPT Enterprise.” Press release; corroborated by Reuters. PwC+1Glassnode (ongoing). On‑chain market intelligence platform. (Example of BMI in digital assets). Glassnode InsightsCryptoQuant (ongoing). On‑chain/market analytics platform for digital assets. (Example of BMI in digital assets). Cryptoquant

Appendix: Author guidance for an empirical paper

• Target outlets: Management Science (finance/OM/IS interfaces), MIS Quarterly, Strategic Management Journal (strategy + capabilities), Journal of Financial Economics (if heavy empirical), or Journal of Management Information Systems. INFORMS Pubs Online+1

• Ethics/IRB: If using GenAI to label texts, record prompts, guardrails, and human verification steps (per MISQ guidance). AIS eLibrary