The Data Economy in the Age of Agents: A Perspective on the Future of Agents as Customers

  1. Hugo Loza-Guzmán

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

The expansion of artificial intelligence is changing how knowledge circulates and is monetized on the web. Systems retrieve information from external sources and deliver answers that can replace a user's visit, while those sources continue to bear the costs of producing and maintaining the content. From 2028 or 2029 onward, queries from autonomous agents with spending budgets would become the main monetization channel for a significant part of the knowledge web. The agent would act as a customer capable of selecting and acquiring resources to complete a task on behalf of its user. The literature on information goods and micropayments provides a basis for examining this transition. Access data from sihita.org, a public repository of water information in Tarija, Bolivia, place it in a regional context. HTTP 402, x402, Cloudflare's Monetization Gateway and Pay per Use, together with licensing marketplaces, offer mechanisms for charging for access to and use of content. For institutions that produce specialized knowledge, the period 2026–2028 represents an opportunity to prepare their offerings and explore commercial relationships with these customers, combining paid services with public access to knowledge.  

Article activity feed

  1. PREreview

    This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/23594220.

    PREreview of "The Data Economy in the Age of Agents: A Perspective on the Future of Agents as Customers"

    Preprint DOI: 10.1590/scielopreprints.18349

    Author: Hugo Loza-Guzmán (SciELO Preprints, Oct 2026)

    Reviewer: Julian Rodriguez, Jr., FRSA (Royal Economic Society / Royal Society of Arts)

    Core Thesis & Structural Assessment

    Loza-Guzmán introduces a provocative, forward-looking perspective: as autonomous AI agents replace human browser traffic, knowledge repositories face an existential revenue collapse. The author proposes transitioning agents from uncompensated scrapers into paying customers via micro-payment protocols (HTTP 402, x402, Cloudflare Monetization Gateways), using a Bolivian public water repository (sihita.org) as an illustrative case.

    However, as currently posted, the manuscript reads as a brief concept note rather than a fully developed research paper. At under 1,500 words, it asserts theoretical shifts without providing concrete telemetry tables from sihita.org, system architecture diagrams, or economic proofs. To transition from a high-level op-ed to a rigorous journal submission, the author needs to expand the manuscript significantly across three key areas:

    Key Theoretical & Empirical Gaps to Address

    1. Synthetic Data Degradation ("AI Slop") vs. Primary Value:

      Unchecked, recursive agent scraping without direct access to verified primary data inevitably degrades web content quality into synthetic noise. As argued in Rodriguez (2026; The Architecture of Asymmetric Obsolescence), digital intermediaries suffer structural entropy when synthetic feedback loops replace verified ground-truth data. Micro-payment models like x402 cannot simply charge for raw access—they must explicitly price and incentivize verified primary human output to prevent agent networks from paying to recirculate degraded synthetic slop.

    2. Principal-Agent Friction & Budget Control:

      Equipping autonomous agents with spending budgets introduces severe agency risks. The paper assumes agents will buy high-value data efficiently. In practice, without strict transaction limits and state-logging frameworks, agents risk draining user budgets on hallucinated wrappers, redundant queries, or low-quality data endpoints.

    3. Empirical Telemetry & Pricing Asymmetry:

      The paper mentions sihita.org but omits concrete access logs, bandwidth costs, and query distributions. For public repositories, enforcing pay-per-use barriers against commercial AI crawlers risks locking out civic researchers and public interest users if human vs. agent traffic cannot be cleanly differentiated. Furthermore, small repositories face extreme pricing monopsony when negotiating with massive AI aggregators.

    Final Recommendation for Journal Expansion

    This is an essential, highly relevant concept note for the 2026–2028 web transition. To prepare this manuscript for peer-reviewed journal publication, the author should:

    • Expand the manuscript into a full case study featuring empirical query telemetry from sihita.org.

    • Provide a concrete sequence diagram showing an x402 micro-payment exchange between an agent and a repository.

    • Incorporate structural perspectives on synthetic data entropy to ground the economic model.

    Sources & References

    • Loza-Guzmán, H. (2026). The Data Economy in the Age of Agents: A Perspective on the Future of Agents as Customers. SciELO Preprints. DOI: 10.1590/scielopreprints.18349.

    • Rodriguez, J., Jr. (2026). The Architecture of Asymmetric Obsolescence: Intermediary Entropy and the Rise of Peer-to-Peer Sovereign Infrastructure. SSRN Electronic Journal. DOI: 10.2139/ssrn.5029381.

    Competing interests

    The author declares that they have no competing interests.

    Use of Artificial Intelligence (AI)

    The author declares that they used generative AI to come up with new ideas for their review.

  2. PREreview

    This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/23594309.

    PREreview of "The Data Economy in the Age of Agents: A Perspective on the Future of Agents as Customers"

    Preprint DOI: 10.1590/scielopreprints.18349

    Author: Hugo Loza-Guzmán (SciELO Preprints, Oct 2026)

    Reviewer: Julian Rodriguez, Jr. FRSA (Royal Economic Society / Royal Society of Arts)

    Core Thesis & Structural Assessment

    Loza-Guzmán introduces a provocative, forward-looking perspective: as autonomous AI agents replace human browser traffic, knowledge repositories face an existential revenue collapse. The author proposes transitioning agents from uncompensated scrapers into paying customers via micro-payment protocols (HTTP 402, x402, Cloudflare Monetization Gateways), using a Bolivian public water repository (sihita.org) as an illustrative case.

    However, as currently posted, the manuscript reads as a brief concept note rather than a fully developed research paper. At under 1,500 words, it asserts theoretical shifts without providing concrete telemetry tables from sihita.org, system architecture diagrams, or economic proofs. To transition from a high-level op-ed to a rigorous journal submission, the author needs to expand the manuscript significantly across three key areas:

    Key Theoretical & Empirical Gaps to Address

    1. Synthetic Data Degradation ("AI Slop") vs. Primary Value:

      Unchecked, recursive agent scraping without direct access to verified primary data inevitably degrades web content quality into synthetic noise. As argued in Rodriguez (2026; The Architecture of Asymmetric Obsolescence), digital intermediaries suffer structural entropy when synthetic feedback loops replace verified ground-truth data. Micro-payment models like x402 cannot simply charge for raw access—they must explicitly price and incentivize verified primary human output to prevent agent networks from paying to recirculate degraded synthetic slop.

    2. Principal-Agent Friction & Budget Control:

      Equipping autonomous agents with spending budgets introduces severe agency risks. The paper assumes agents will buy high-value data efficiently. In practice, without strict transaction limits and state-logging frameworks, agents risk draining user budgets on hallucinated wrappers, redundant queries, or low-quality data endpoints.

    3. Empirical Telemetry & Pricing Asymmetry:

      The paper mentions sihita.org but omits concrete access logs, bandwidth costs, and query distributions. For public repositories, enforcing pay-per-use barriers against commercial AI crawlers risks locking out civic researchers and public interest users if human vs. agent traffic cannot be cleanly differentiated. Furthermore, small repositories face extreme pricing monopsony when negotiating with massive AI aggregators.

    Final Recommendation for Journal Expansion

    This is an essential, highly relevant concept note for the 2026–2028 web transition. To prepare this manuscript for peer-reviewed journal publication, the author should:

    • Expand the manuscript into a full case study featuring empirical query telemetry from sihita.org.

    • Provide a concrete sequence diagram showing an x402 micro-payment exchange between an agent and a repository.

    • Incorporate structural perspectives on synthetic data entropy to ground the economic model.

    Sources & References

    • Loza-Guzmán, H. (2026). The Data Economy in the Age of Agents: A Perspective on the Future of Agents as Customers. SciELO Preprints. DOI: 10.1590/scielopreprints.18349.

    • Rodriguez, J., Jr. (2026). The Architecture of Asymmetric Obsolescence: Intermediary Entropy and the Rise of Peer-to-Peer Sovereign Infrastructure. SSRN Electronic Journal. DOI: 10.2139/ssrn.5029381.

    Competing interests

    The author declares that they have no competing interests.

    Use of Artificial Intelligence (AI)

    The author declares that they used generative AI to come up with new ideas for their review.

  3. Version published to 10.1590/scielopreprints.18349 on SciELO Preprints