💰 quantitative finance

Organizational Technology Ladders: Remote Work and Generative AI Adoption

This study demonstrates that firms' prior adoption of remote work technology during the pandemic accelerated their subsequent uptake of generative AI by fostering organizational capital and shifting hiring toward technical and managerial capabilities that lower the barriers to adopting new technologies.

Original authors: Gregor Schubert

Published 2026-08-13

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Gregor Schubert

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the world of business as a giant, bustling construction site where companies are constantly trying to build better, faster, and smarter ways to do their work. For a long time, economists and business experts have known that you can't just throw a new tool at a problem and expect magic to happen. If you want to build a skyscraper, you first need a solid foundation, the right cranes, and a crew that knows how to use them. This idea is called "complementarity": new technology works best when it's paired with the right people and the right organization. Think of it like trying to play a video game on a brand-new console; if you don't have the controller or the power cord, the fancy machine is just a expensive paperweight. The big question researchers have been asking is: does fixing one part of the "machine" (like letting employees work from home) actually make it easier to add the next cool upgrade (like artificial intelligence)?

This paper, written by Gregor Schubert from UCLA, dives into exactly that question. It looks at two massive waves of change that hit companies recently: the sudden shift to remote work during the pandemic, and the explosion of Generative AI (the kind of smart computer programs that can write code, draft emails, and create art). The author proposes a fun idea called the "Organizational Technology Ladder." Imagine a ladder where each rung is a new technology. The theory is that climbing the first rung (remote work) doesn't just get you higher; it actually changes the shape of the ladder itself, making the next rung (Generative AI) easier to grab. The study uses a massive amount of data from job postings to see if companies that got good at remote work were also the ones that got good at using AI later. The findings suggest a clear path: yes, the order matters. Companies that successfully adapted to remote work built the necessary skills and digital tools that made adopting AI much cheaper and faster for them. It's not just that smart companies do everything; it's that doing one thing well actually helps you do the next thing better.

The Ladder Effect: How Remote Work Paved the Way for AI

The core discovery of this paper is that companies don't adopt technologies in isolation; they climb a ladder. When firms were forced to figure out how to let employees work from home during the pandemic, they didn't just buy a few laptops. They had to rewire their entire organization. They had to hire people who knew how to manage digital workflows, invest in better cloud systems, and learn how to make decisions without everyone sitting in the same room. The author argues that these specific investments became the "rung" that allowed them to climb to the next level: Generative AI.

To prove this wasn't just a coincidence (like two smart companies happening to pick both technologies), the author used a clever statistical trick. They looked at how much pressure different companies felt to offer remote work based on their location and the local job market, rather than just their own choices. This helped isolate the effect of remote work itself. The results were clear: for every 10 percentage point increase in a company's remote hiring during 2021–2022, that company increased its hiring for Generative AI skills by about 0.4 percentage points by 2023–2024. When looking at specific jobs within the same company, the effect was even stronger: a 10 percentage point difference in remote work across different roles led to a 0.7 percentage point increase in AI-related hiring.

Why Does This Happen? The Two-Step Mechanism

The paper suggests two main reasons why climbing the "remote work rung" helps you reach the "AI rung."

First, it's about building the right muscles. To make remote work function, companies had to hire technical experts and managers who could handle complex digital coordination. They invested in data infrastructure and security. The paper finds that these exact skills and tools are the same ones needed to run Generative AI. It's like if you had to learn how to drive a manual transmission car to get to work; once you mastered that, learning to drive a high-tech electric car was much easier because you already understood the basics of gears and steering. The data shows that companies that hired more for technical roles (like data science and machine learning) and managerial roles during the remote shift were the ones who moved fastest to adopt AI later.

Second, it's about fixing the friction. Some companies struggled with remote work; they felt like their teams were disconnected, decisions were slow, and creativity was suffering. The paper found that these specific companies—those that eventually issued "return-to-office" mandates because they felt remote work wasn't working—were actually the ones that most aggressively adopted Generative AI. It seems they turned to AI as a way to fix the problems remote work created. If a team was struggling to coordinate over Zoom, they used AI to automate the tasks that were causing the bottlenecks. In this sense, the struggle with remote work acted as a push, forcing them to find a technological solution.

Who Benefits the Most?

Not everyone climbs the ladder at the same speed. The study found that the "technology ladder" effect was strongest in industries where work is heavy on information and decision-making, specifically in Finance, Business Services, and the Tech sector. In these fields, the quality of a decision matters a lot. If remote work makes decision-making harder, the value of using AI to help process information and automate routine tasks becomes huge. The paper suggests that in these sectors, the combination of remote work experience and AI adoption creates a "compound advantage," where the two technologies boost each other's effectiveness.

What the Paper Rules Out

It is important to note what this study says is not the main driver. The author argues against the idea that these companies were just "tech-savvy" from the start and happened to pick both technologies by chance. The statistical methods used to control for a company's pre-pandemic tech skills show that the remote work experience itself caused the shift toward AI. Furthermore, the study suggests that simply having a job that could be done remotely doesn't automatically mean it will be done with AI; the company has to actually go through the process of adapting to remote work first. The paper also notes that while job postings show a clear trend of hiring for AI skills, this measures the company's official intent and formal hiring, not necessarily every single employee casually using AI tools on their own.

The Bottom Line

The story here is one of preparation. The pandemic forced companies to climb the first rung of a ladder by mastering remote work. In doing so, they accidentally (or perhaps intentionally) built the digital infrastructure and hired the talent needed to grab the next rung: Generative AI. Companies that skipped the first rung or struggled to adapt to it found the second rung much harder to reach. As new technologies continue to emerge, this research suggests that a company's ability to innovate isn't just about the new tool itself, but about the foundation they built with the tools they used yesterday. For business leaders, the takeaway is that investing in digital adaptation today might be the best way to ensure they are ready for the next big wave of technology tomorrow.

Technical Summary: Organizational Technology Ladders: Remote Work and Generative AI Adoption

1. Problem and Motivation

This study investigates the causal dynamic between the adoption of remote work technologies during the COVID-19 pandemic and the subsequent uptake of generative AI (GenAI) by firms. While GenAI adoption has been rapid, it remains strikingly uneven across organizations. The paper posits that this heterogeneity is not solely a function of task exposure to automation but is driven by an "organizational technology ladder."

The core hypothesis is that adopting one technology (remote work) transforms a firm's hiring patterns, organizational capital, and technical infrastructure. These transformations alter the cost-benefit calculus for adopting subsequent technologies (GenAI). Specifically, the paper argues that firms that successfully adapted to remote work built the necessary digital and managerial capabilities to lower the implementation costs of GenAI, thereby climbing the ladder to the next rung of technological adoption.

2. Methodology

Data Sources

The analysis utilizes a comprehensive dataset of U.S. job postings from Lightcast (January 2010 – September 2024), filtered for full-time, non-internship roles.

Key Variables:
- Remote Work Share: The proportion of job postings indicating remote or hybrid work arrangements (2021–2022, excluding Q4 2022 to avoid ChatGPT contamination).
- GenAI Adoption: Measured by the share of job postings mentioning GenAI skills or tools (2023–2024).
- Exposure: Occupation-level GenAI exposure scores (Eisfeldt et al., 2025) aggregated to the firm level.
- Controls: Pre-pandemic (2019) hiring characteristics, industry fixed effects, and labor market variables.
Supplementary Data: Earnings call transcripts (Capital IQ/WRDS, 2019–2025) analyzed via a multi-stage LLM pipeline to validate technology investment narratives; Flex Index data for Return-to-Office (RTO) mandates; and O*Net/BLS data for occupation characteristics.

Identification Strategy

To establish causality and address endogeneity (e.g., firms that are "tech-savvy" might adopt both remote work and GenAI simultaneously), the author employs an Instrumental Variables (IV) strategy.

Instruments: The study exploits exogenous variation in labor market pressure to offer remote work.
1. Competition Instrument: Interacts a firm's pre-pandemic (2019) teleworkability with the average teleworkability of other firms in the labor markets (MSAs) where the firm hires. The logic is that firms in markets with high remote-work competition are pressured to offer remote options to retain talent.
2. Commuting Instrument: Interacts firm teleworkability with pre-pandemic average commuting times in hiring MSAs. Longer commutes increase the perceived value of remote work.
Variation: The analysis utilizes both cross-firm variation and within-firm variation across occupations (firm-by-occupation level), controlling for firm and occupation fixed effects to isolate the effect of remote work adoption on GenAI hiring within the same organization.

Empirical Model

The baseline specification estimates:
$\text{GenAIJobShare}_{i} = \alpha + \beta \text{RemoteWorkShare}_{i} + \text{Controls}_{i} + \epsilon_{i}$
Where $i$ represents a firm or a firm-occupation cell. The coefficient $\beta$ captures the causal effect of remote work share on the subsequent share of GenAI-related job postings.

3. Key Results

Causal Effects

Cross-Firm: A 10 percentage point (pp) increase in remote hiring share during 2021–2022 causes a 0.4 pp increase in the share of job postings mentioning GenAI skills by 2023–2024.
Within-Firm: A 10 pp difference in remote work share across occupations within the same firm is associated with a 0.7 pp increase in GenAI mentions for those specific occupations.
Extensive Margin: A 10 pp increase in remote work raises the likelihood of a firm mentioning GenAI at all by approximately 0.9 pp.

Heterogeneity and Mechanisms

The study identifies specific channels through which the "technology ladder" operates:

Capability Accumulation: Remote work adoption induces firms to hire more for technical skills (data science, machine learning, deep learning) and managerial/coordination roles. These capabilities are shown to predict a higher conversion of GenAI exposure into actual adoption.
Sector Specificity: The ladder effect is most pronounced in Technology and Financial Activities & Business Services sectors. These are information-intensive industries where decision quality is critical, suggesting that the organizational adaptations required for remote work (digital workflows, data infrastructure) are highly complementary to GenAI automation.
Return-to-Office (RTO) Mandates: Firms that issued RTO mandates (interpreted as revealing low perceived remote productivity or high coordination frictions) exhibit a substantially larger response to remote work adoption in terms of GenAI hiring. This suggests these firms use GenAI as a substitute to mitigate the productivity losses associated with distributed teams.
Investor Narratives: Analysis of earnings calls confirms that firms discussing remote work investments frequently mention data infrastructure, cloud access, and security. These same capabilities are explicitly cited as "enablers" for GenAI adoption, providing narrative evidence of the reusability of investments.

Theoretical Framework

The paper formalizes the mechanism using a task-based model of sequential technology investment. The model demonstrates that if remote work adoption leads to investments in reusable digital infrastructure and skills ( $S_R$ ), the marginal cost of implementing GenAI ( $S_G$ ) decreases. The "ladder" exists if the cost savings from reusability and the productivity gains from automation outweigh the potential decision-quality penalties of remote work.

4. Contributions and Significance

Theoretical Contribution

The paper introduces the concept of an "organizational technology ladder," arguing that technology adoption is path-dependent. It extends the literature on organizational capital and technology (e.g., Garicano and Rossi-Hansberg, 2006; Bresnahan et al., 2002) by empirically linking the adoption of one wave of technology (remote work) to the readiness and likelihood of adopting the next (GenAI). It highlights that organizational adaptation—specifically in skills and infrastructure—is a critical determinant of technological success.

Empirical Contribution

Causal Link: It provides the first causal evidence linking remote work adoption to GenAI uptake, distinguishing this from simple correlation driven by "tech-savviness."
Mechanism Evidence: It moves beyond correlation to identify specific mechanisms: the accumulation of technical/managerial skills and the role of coordination frictions (RTO mandates) in driving automation.
Data Innovation: The use of job posting data as a proxy for formalized organizational intent, combined with LLM-classified earnings call transcripts, offers a novel way to measure organizational technology adoption and investment narratives.

Policy and Managerial Implications

The findings suggest that "AI readiness" is not merely a function of task exposure but is deeply rooted in prior organizational investments. Firms that treated remote work as a capability-building exercise (investing in infrastructure and talent) are better positioned to leverage GenAI. Conversely, path dependency may exacerbate productivity gaps, as firms that successfully navigated the first technological shock (remote work) gain compound advantages, potentially locking in persistent differences between organizational leaders and laggards.

Limitations

The author notes that GenAI adoption is proxied by job postings, which captures formal hiring intent but may miss informal or decentralized usage by incumbent workers. Additionally, while the study identifies organizational capability shifts, it does not directly measure post-adoption productivity or decision quality outcomes.