Organizations need access to the right inputs at the right time and processes capable of carrying that information through the TP lifecycle. Progress comes when TP design, operation, documentation and defense draw from the same evidentiary foundation.
The difference today is the ability to access and connect underlying facts at scale. For years, teams built positions from a limited set of interviews, transactions and manually selected comparables because that was all they could realistically gather and reconcile. Because relevant evidence resides across ERP systems, contracts, emails and operational records, technology now makes it easier to trace comparable data back to its source.
Rather than assembling support after the fact, organizations can monitor positions as business conditions evolve, identify events that affect outcomes and capture the reasoning behind decisions when they occur.
“Design informs operation, operation feeds documentation, and documentation — ideally grounded in one consistent data lake — strengthens defense,” says Volker Trautmann, EY Global Transfer Pricing Documentation CoE Co-leader.
When teams redesign their processes to capture facts as decisions occur, they can reduce the reconciliation and reconstruction that consume so much year-end effort. By the time a tax authority requests information, much of the supporting narrative already exists. In that sense, the organizations making the most progress no longer treat audit readiness as a separate exercise but build it into the TP lifecycle itself.
AI and standard data earn their place
Tax leaders see the potential of AI clearly. As noted previously, half of respondents expect advancements in AI to have a significant impact on TP over the next three years, ranking it ahead of every regulatory and commercial pressure they face. Yet the same survey shows how early the shift still is. Only 17% deploy agents or machine learning to any meaningful degree, and among those deploying AI, 11% use it extensively in audit preparation.
That distance between expectation and practice is a challenge that many companies face today. AI delivers on its promise only when the inputs and processes beneath it are ready.
“Using agentic workflows with AI-ready data is transformative,” says Rebecca Coke, EY Global Transfer Pricing Product Portfolio Leader. “Using AI with unstructured data of unknown provenance and context often leads to high token costs and very wrong answers.”
On the inputs, AI pulls a fact from one system, a decision from another and a comparable from a third, and holds them together or flags when data is not connected. Both inputs and process depend entirely on data quality and accessibility — which is precisely the main obstacle cited by 38% of survey respondents, making it their biggest AI readiness risk.
This is why the work starts with reframing and reevaluating the inputs and data, not the technology. A company that learns to mine data across the enterprise builds a reliable base of connected evidence and creates one source of truth, where each trusted input can compound and make the next connection possible. That is the first sign of an environment where the enterprise’s TP evidence works as a whole rather than in pieces and where trusted sources are built.