Running a campaign on TikTok used to mean a media buyer inside a dashboard, manually setting up creatives, adjusting bids and shifting budgets. TikTok wants artificial intelligence agents to own that work.
At TikTok World, it’s sixth annual global ad product summit held Tuesday (May 12), the platform launched its TikTok Ads Model Context Protocol (MCP). The system connects AI agents directly to TikTok’s ad platform, giving them the ability to plan, launch and optimize campaigns without human intervention. Jose Villalobos, global head of product marketing for platform and core ads, Digiday reported, said the goal is to combine automation, control and AI to drive better performance for advertising partners.
What MCP Does to an Ad Platform
MCP is an open standard introduced by Anthropic in late 2024. PYMNTS reported that MCP lets AI models connect to live business tools and systems, moving them from passive assistants to active agents that can retrieve data, update records, and execute actions within approved boundaries. Before MCP, each connection between an AI system and a business application had to be custom built. The protocol standardizes that integration layer.
Applied to advertising, MCP turns the connection between an AI agent and an ad platform into a structured, programmable interface. Campaign setup, bid adjustments, budget reallocation, and targeting changes all become operations an agent can execute directly rather than tasks a human has to perform through a dashboard.
TikTok isn’t alone. Google has released an open-source Google Ads MCP server. Meta launched one that lets advertisers manage ad accounts through Claude and ChatGPT without touching its ads manager. Amazon has done the same. Digiday reported that the direction is consistent across every major platform: build the infrastructure to let AI agents, not human media buyers, handle the operational work of running campaigns.
Control the Connection, Control the Data
Adweek reported that the TikTok announcement also included support for third-party API agents, letting advertisers build custom infrastructure for their own workflows on top of TikTok’s ad systems. That interoperability extends the platform beyond its own AI tools into what Adweek described as AI-native ad infrastructure.
Advertisement: Scroll to Continue
Adtech expert Shirley Marschall told Digiday the data angle is what’s actually driving the MCP race among major platforms. A platform that routes agent traffic through a third-party MCP loses visibility into how agents are querying it, what they’re comparing it against, and what’s driving conversion. In an agentic advertising environment, that signal is the most valuable data a platform holds. Large platforms are moving to control their own MCP connections before that data flows elsewhere.
The split, Marschall said, runs along scale: Large platforms race to set their own MCP standards; everyone else races not to depend on them.
The Ad Stack Underneath
TikTok paired the MCP launch with a broader slate of ad product updates at TikTok World. TopReach combines TopView and TopFeed placements into a single buy. TopReach Sequencing lets advertisers hold the first two high-visibility placements to run a continuous narrative. Search Hubs places branded pages at the top of TikTok search results. Creator AI Search, inside TikTok One, interprets campaign briefs and surfaces relevant creators automatically.
The company also integrated Dreamina Seedance 2.0, ByteDance’s video generation model, into its Symphony AI suite, and added profit optimization features to TikTok Shop’s GMV Max.
TikTok World came four months after the platform’s U.S. future was resolved through a U.S.-China deal that ended months of legal uncertainty. Three new executives led the stage alongside Villalobos, including Isobel Sita Lumsden, who took over as global head of business marketing last month.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
