The Trade Desk confirmed it is testing AI-driven campaign creation built on Anthropic's Claude, according to a headline published by trade outlet ADWEEK.

The Trade Desk operates one of the largest independent demand-side platforms (DSPs) in programmatic advertising, buying digital inventory across display, video, connected TV (CTV), mobile, audio, and digital out-of-home on behalf of agencies and advertisers. Claude is the family of large language models developed by Anthropic, an AI vendor that has positioned itself around enterprise deployments.

What does the headline actually say?

Only this: The Trade Desk is testing AI campaign creation with Claude. No additional detail on the test's scope, the model variant in use, participating agencies, pilot length, pricing structure, or commercial launch date appeared in the available source material.

Operators reading past the headline should treat the confirmation as a signal of strategic direction, not as a procurement-ready product announcement.

What would AI campaign creation actually do?

In a DSP context, "campaign creation" typically covers the workflow a media planner executes today by hand:

• Reading a client brief
• Selecting inventory channels and supply sources
• Setting audience targeting parameters
• Allocating budgets across line items
• Drafting initial bid strategies

A language model that ingests the brief and outputs a structured plan would compress that workflow. Whether Claude performs any of those steps inside the Trade Desk test, or sits behind the interface handling only one function, is not addressed by the source.

Where does this fit the AI-in-ad-tech arc?

The Trade Desk is not the first major ad-tech platform to publicly discuss large language models inside the buying stack. Industry coverage through 2024 and 2025 has tracked similar experiments from the supply side, from rival DSPs, and from agency-holding-company in-house tools.

The Trade Desk's specific bet on Claude signals a model-vendor choice. Anthropic has marketed Claude around longer context windows, code execution, and tool use, features that align with structured campaign outputs. Whether the test will extend to other Anthropic models or remain on a single Claude variant is unstated.

What details would an operator need?

Procurement-relevant information that typically follows this kind of announcement:

• Pilot participants, meaning which agencies or advertisers
• Duration and the success metrics in play
• Pricing pass-through to clients
• Integration with existing Trade Desk platform features
• Brand-safety, audit, and latency guardrails
• Commercial launch date

None of those points appear in the source headline. Operators planning 2025 budgets should wait for the next The Trade Desk earnings disclosure, the vendor's newsroom, or an industry-event announcement before evaluating this test as a buyable product.

Bottom line

The Trade Desk has confirmed an AI campaign creation test on Claude. The confirmation is thin, the commercial product is unannounced, and the timeline is open. For agency operators, the signal worth tracking is which DSP closes the gap between LLM-assisted planning and a procurement-ready product first.