Multinationals sharing proprietary AI agents among their subsidiaries risk the possibility of more disputes with tax authorities as questions arise about how to price that sharing. <\/p>\n
The fact that any person with access to the technology can make or add to an agent raises issues about where the value of the agent should be assigned, especially because there is no law or international standard governing the transfer pricing \u2014 valuation of affiliate transactions \u2014 of AI. <\/p>\n
Companies may find themselves in a situation where multiple countries claim they\u2019re owed more tax because they think the value of an AI agent was created in their jurisdiction. <\/p>\n
An AI agent is software coded to perform autonomous tasks such as collecting and analyzing data or researching and writing memos. <\/p>\n
A fundamental issue is whether sharing the artificial intelligence agent is sale of a service, or sale or licensing of intellectual property, which is usually of higher value. The determination isn\u2019t always easy, but is critical under transfer pricing rules, which govern intragroup transactions and have enormous implications for profit allocation and tax bills.<\/p>\n
For example, Grant Thornton partner Samit Shah said, a sales team dispersed through 30 countries contributes to development of an agent that finds new leads to sell products.<\/p>\n
\u201cYou\u2019ve got a collective team that has built an agent that\u2019s used around the world. Is this a shared service allocation? Is this a cost share?\u201d he said, referring to ways companies allocate costs under transfer pricing rules.<\/p>\n
\u201cThat\u2019s one of my major questions,\u201d he said. \u201cIs this truly IP development or is this a service? Those are two different things.\u201d<\/p>\n
And even if that distinction is made, allocating the value, profit, and tax due to the correct jurisdiction is difficult. <\/p>\n
Companies must be able to figure out how much profit each entity deserves, Shah said. If governments are left to their own devices, \u201cthey\u2019re going to claim the biggest piece\u201d of profit they can. <\/p>\n
Back to Basics<\/p>\n
In the absence of guidance, practitioners said they\u2019re going back to transfer pricing basics: looking at companies\u2019 AI investments, the functions the technology performs, and how it benefits the business.<\/p>\n
Some of the investments companies are making in AI are routine and \u201cenhance capabilities that the company may already have,\u201d said Brian Burt, a leader in PwC\u2019s transfer pricing practice. He cited an example of an agent that\u2019s used to spot new consumer trends.<\/p>\n
In that case, a services categorization is warranted, he said.<\/p>\n
But other investments in AI may provide a \u201cnew channel to business\u201d or \u201cnew channel to market,\u201d Burt said, potentially warranting designation as an IP transaction.<\/p>\n
The difference matters. Services \u2014 IT assistance, software use, and legal advice exchanged for a fee \u2014 are generally valued lower than IP, where a royalty payment or licensing fee is due to the affiliate designated as the IP\u2019s owner. <\/p>\n
Firas Zebian, a leader at Deloitte\u2019s US transfer pricing and value chain alignment group, said a subsidiary\u2019s ability to change the model impacts how transactions are characterized. <\/p>\n
Like Burt, Zebian said many of the AI transactions he sees at the moment are SaaS-like \u2014 providing software as a service. The parent company shares a \u201clocked\u201d AI model with a subsidiary in another jurisdiction, but the subsidiary can\u2019t retrain the agent. <\/p>\n
But if affiliates share an agent that\u2019s customizable, where coding can be changed, this activity \u201ccan be close to a license\u201d for IP, he said. <\/p>\n
Business Impact<\/p>\n
Burt said it\u2019s important for companies to look at the impact AI investment has on the business to accurately characterize the transaction. <\/p>\n
\u201cIs it increased revenue? Is it decreased cost? Is it just back-office efficiency? Is it speed to market?\u201d <\/p>\n
Answers will help a company figure out whether an agent is a separate asset, he said. <\/p>\n
Anthony Pastore, a partner at Mayer Brown, pointed to 2025 US Treasury Department final regulations that, in the absence of AI-specific guidance, could help companies characterize the agent for pricing purposes. <\/p>\n