Cint and Potloc deploy MCP to automate research workflows via AI agents
Cint, the programmatic research and measurement platform, is collaborating with survey technology firm Potloc to develop and test a Model Context Protocol (MCP) server that connects AI agents directly to Cint's research infrastructure. The move is an early-stage but concrete signal of how the market research industry is beginning to restructure itself around agentic AI, where software agents execute complex, multi-step workflows in response to natural-language instructions rather than manual configuration.
The Model Context Protocol is an open standard, originally developed by Anthropic, that allows large language models (LLMs) to query and interact with external tools and data sources in a standardised way. By building on this standard, Cint is positioning its platform to be callable from within any LLM-based environment a client already uses, whether that is a proprietary AI assistant, an orchestration layer, or an emerging class of agentic research tools.
From console clicks to conversational commands
The practical gains are already visible in Potloc's production environment. In a single automated session, the firm's fieldwork managers configured a 20-market consumer banking study on the Cint Exchange, setting up 20 separate target groups with locale specifications, feasibility checks and pricing bids in one pass. Previously, each market required manual console configuration. "Our fieldwork managers draft, price and validate target groups conversationally, in a fraction of the time," said Mathieu Dussart, Director of Sampling Strategy at Potloc.
The Cint MCP is designed to cover the full research lifecycle: defining target audiences through plain-language prompts, checking feasibility across geographies, generating pricing guidance, launching studies, and monitoring fieldwork in real time. On the media measurement side, Cint says the MCP will allow clients to activate measurement studies and generate tailored performance insights at scale, though those capabilities remain in active development.
Convergence read-across: the agentic layer reaches market intelligence
The significance of this collaboration extends well beyond the survey industry. For cross-sector strategists, the story is about where the agentic AI layer is landing next. Financial services firms and management consultancies, both named as core Potloc clients, have already absorbed agentic workflows into legal document review, financial modelling, and supply-chain monitoring. Market intelligence and primary research, which have historically required skilled human coordination across vendors, geographies, and data formats, are now being pulled into the same automation paradigm.
This matters for capital allocation. The enterprise software firms and research technology vendors that fail to expose their platforms via open protocols such as MCP risk being disintermediated by AI orchestration layers that route around them to sources that are machine-readable by default. Conversely, platforms that become natively accessible to AI agents gain distribution through every client's AI stack rather than just through their own interface. Cint, which operates one of the world's largest programmatic survey panels spanning 130-plus countries and hundreds of millions of respondents, is making a deliberate bet that becoming an AI-native data source is a stronger competitive moat than any proprietary interface.
The broader investment landscape reflects this logic. Enterprise AI infrastructure spending, covering the orchestration, data-connectivity, and agentic tooling layers, is attracting capital from both specialist venture investors and the corporate balance sheets of large software incumbents. The MCP standard itself is gaining rapid adoption as a lingua franca for agent-to-tool communication, with Microsoft, Salesforce and a growing roster of enterprise vendors publishing their own MCP servers. Cint's move positions it within that ecosystem rather than outside it.
Cint is now inviting additional AI-forward companies to join as beta partners, using their feedback to shape the MCP's continued development. The company describes the MCP as complementing, rather than replacing, its existing API and platform interfaces, which reduces switching risk for clients already integrated via conventional routes. Whether the capability matures into a durable competitive differentiator or becomes table-stakes infrastructure, as APIs themselves eventually did, will depend on how quickly rivals in the research technology market follow suit.