MLS Future Hinges on Control of Housing Data by 2030

The Multiple Listing Service, built for a world where brokers were the primary gatekeepers of housing information, faces a fundamental transformation by 2030. AI and evolving data practices could redefine what it means to control property listings.

MLSs are shifting from systems that distribute listings among agents into infrastructure for governing and moving data among consumers, companies, and AI. A central, contentious question shaping this evolution is who should control how listings enter and move through that system.

Stephen Brobeck, a senior fellow at the Consumer Policy Center, argues for total listing transparency in the general consumer interest. "It is in the general consumer interest for there to be total listing transparency, for sellers to be able to market their listings broadly and for buyers to have access to up-to-date, important information about all listings," he told HousingWire. He cited a historical precedent in the U.S. Department of Justice's 2003 litigation, settled in 2008, which compelled the National Association of Realtors and MLSs to provide listing information to public portals, a move that reduced consumer reliance on agents alone.

Brokerage Strategies Diverge

Major brokerages hold different positions on seller control and market exposure. Compass advocates for greater seller flexibility, promoting a phased marketing strategy that begins with a private exclusive listing before broadening to a 'coming-soon' phase and eventually public websites. Keller Williams executive chairman Gary Keller has also supported a seller's right to choose private marketing but emphasizes the importance of broad exposure and full disclosure of the trade-offs involved. Both firms told HousingWire their views have not changed from previous public comments.

AI Reshapes the Interface

Victor Lund, co-founder of consulting firm WAV Group, argues the MLS does not need reinvention but requires an architecture ready for machines. "The MLS is already a live data repository with a front end for search and reports," said Lund. He predicts the data structure will become more AI-ready, a transition happening within the next couple of years, citing examples like NexusRE, FlexMCP, TrestleMCP, and Utah Real Estate.

He forecasts a conversational interface will emerge quickly. "Agents will not need to log in to find information or create work product." Lund believes tasks like setting up client searches or looking up listing history will be handled by an AI agent on behalf of the licensed Realtor, though the traditional MLS interface will remain.

Governance is the Hard Part

Lund describes an AI-ready MLS architecture using an egg analogy. "The data is the yoke in the middle. It is very easy, takes about 6 hours to move the MLS records onto an MCP server." He says connecting a database to an AI model is straightforward, but determining the rules for access is harder.

He also highlighted a key potential of the new infrastructure. "Where the infrastructure gets very interesting is when it starts to learn with every question prompt and every answer." However, Lund expects the listing-creation process to remain human-centric, with agents curating field options and verifying data, even as AI assists with image extraction and descriptions.

A Focus on Present Value

Richard Haggerty, CEO of OneKey MLS, offered a more pragmatic vision, arguing for building on current functions rather than technological conjecture. "I would build an MLS based upon fact, not conjecture," he said. He criticized the overemphasis on predicting where the technology 'puck' is going. For Haggerty, design must center on the people who rely on the system: the broker, then their agents, and finally the consumers they serve.

He sees creating a more efficient, less complicated ecosystem as a central opportunity, a goal that could align with AI transforming repetitive workflows. "I think if you’re creating flexibility and you’re simplify the process, you’re making access to the data more seamless, and you’re still ensuring the accuracy and completeness of the data," Haggerty stated, calling that the foundation of the MLS for 2026 or 2030.

Curbing 'Shadow AI'

A significant challenge for MLSs is managing unofficial AI use, or 'shadow AI,' where agents use tools without structured access. Lund views this as a consequence of MLSs failing to provide better alternatives. "Shadow AI is a workaround that does not work very well."

The solution, he suggests, is for MLSs to create authenticated pathways allowing licensed users to connect approved AI systems directly to MLS data, preserving controls and auditability. This would position the MLS as a trusted gateway rather than a locked vault. Lund concludes that MLSs do not need greater authority to fulfill this role, they need to become more effective facilitators of structured access.