Parag Agrawal’s AI startup aims to compensate publishers when AI agents utilize their content

(SeaPRwire) –   Parag Agrawal’s latest venture aims to address one of AI’s most contentious issues: fairly paying content creators as AI agents, not people, increasingly browse the internet.

His firm, Parallel Web Systems, is introducing Index. This platform offers publishers, data providers, and independent creators insight into how AI agents utilize their material and a novel method to get paid for it. Initial partners encompass publishers and distributors like The Atlantic and PR Newswire; business and data intelligence firms such as PitchBook, Enigma, RocketReach, and ZoomInfo; and independent creators like Alex Heath’s Sources, Packy McCormick’s Not Boring, and Mario Gabriele’s The Generalist.

“The company’s fundamental belief was that agents will utilize the web far more than humans, and consequently, everything about the web will transform—both its technology and its business models,” Agrawal stated to .

Unlike human readers who click through pages, AI agents can simultaneously draw from dozens or hundreds of sources to fulfill user requests. Parallel already provides web access infrastructure to AI firms and developers, including Harvey, Notion, and Opendoor.

Index seeks to assist content owners in comprehending how AI agents employ their work. The platform centers on a concept known as Shapley value, a game theory principle for assessing each participant’s contribution to a collective result. Parallel’s approach, instead of paying merely for access or citations, attempts to gauge how much a specific source contributed to an AI agent’s finished task and the value of that agent’s overall output. In theory, a more unique source or one used in higher-value work should get greater payment.

Parallel plans to compensate participating publishers using this value-estimation framework. Initially, the system will cover AI agents using Parallel’s own tools, but the company aims for Index to eventually function with externally built agents as well. Agrawal notes AI companies are incentivized to join because their agents require high-quality content, particularly as more publishers erect barriers against AI crawlers and agents.

“AI agents are emerging as the next primary interface for information access, but the web’s economics haven’t adapted to this shift,” said Nicholas Thompson, CEO of The Atlantic, in a statement. “Parallel addresses this by developing a dynamic, scalable model for acknowledging and paying publishers.” ( is part of Parallel’s Index platform. Under the program, the startup distributes revenue to publications when AI agents use their content.)

This method contrasts with the fixed-fee licensing agreements that have so far characterized AI company-publisher relations. For instance, OpenAI has inked deals with publishers like the Associated Press, Axel Springer, and News Corp.

Agrawal argues such fixed-price arrangements are unsuitable for the AI agent era and could marginalize smaller publishers and AI startups.

“If only a handful of large corporations can access premium content, how can anyone else compete?” he asked.

The agent economy

The launch occurs against a backdrop of multi-year friction between AI companies and content owners.

The New York Times sued OpenAI and Microsoft in 2023, claiming millions of its articles trained AI systems without consent. Dow Jones and the New York Post have also filed suit against AI startup Perplexity, accusing it of unlawfully using copyrighted material. Authors have similarly contested AI companies’ use of books in training large models. Last year, Anthropic consented to a significant settlement in a copyright lawsuit filed by authors.

Parallel isn’t the sole company attempting to forge a new economic framework for AI web access. Last year, Cloudflare unveiled Pay Per Crawl in private beta, a product enabling publishers to charge AI crawlers a standard fee per request across their domain.

Both Index and Cloudflare’s solutions aim to offer an alternative to simply blocking AI crawlers or granting free access. The key distinction is that Cloudflare’s model is founded on crawl access, whereas Parallel’s Index strives to link payment to the value of the agent’s work and the source’s role in it.

Index’s framework relies on the challenging endeavor of determining which sources were most influential for an AI’s output. Publishers, already cautious about AI’s impact on conventional revenue like web traffic, might also worry about over-reliance on another intermediary.

Agrawal states that this uncertainty partly motivates beginning with transparency. Index will allow site owners to input their domain and observe how Parallel’s agents are using their content. The company anticipates more partners will join over time and offer feedback on the model.

Parallel, established roughly two and a half years ago, secured a $100 million Series B round at a $2 billion valuation in April, following a $100 million Series A at a $740 million valuation five months earlier.