For over a decade, property management platforms have been judged by their openness. We at Thesis Driven have written extensively about real estate software's tug of war around open systems. But in those conversations, "openness" usually meant one thing: how many third-party point solutions a platform integrated with, and how good its API was.
Artificial intelligence is now scrambling that definition. As more operators prefer to run their work directly inside foundation models, openness is taking on a new dimension: how fluidly data and action move in and out of tools like Claude and Copilot.
AppFolio offers the sharpest illustration of the shift. It has long been an outlier in the traditional “openness” conversation, betting on its own native tooling rather than a sprawl of third-party integrations. Earlier this summer the company shipped what it bills as the real estate industry's first bidirectional MCP connector to Claude, a move that reframes the openness question and could hand smaller and middle-market operators an edge over larger rivals in the race to adopt AI.
The openness question
For much of the past decade, most property management software platforms treated third-party point solutions as both a revenue line and a feature-coverage strategy. This required balancing competing, and in some cases outright contradictory, priorities: satisfying client desires to integrate specific point solutions, protecting their moat as the definitive source of truth for tenant ledgers and payments, and generating revenue by selling API access to third-party software companies.
Yardi, for instance, integrates broadly but charges software providers upwards of $25,000 for the privilege, and endpoints are limited for certain key functionality such as rent payments. RealPage has a standardized integration marketplace, RPX, but many Thesis Driven subscribers have found the API lacking. Either way, the openness scoreboard was roughly: how many tools, how good the API, and how much it cost to connect one system to another.
AppFolio always played that game differently. Its Stack Marketplace holds fewer than 70 integrations, and the company is deliberate about what gets in. It is not a blanket exclusion: the editorial logic is about protecting the platform's foundational control points, not locking out an entire category.
Where RealPage and Yardi will happily carry a tool that overlaps their own products, AppFolio treats the agentic workflows for its core property-management domains as something to own rather than rent. It is an editorial decision as much as a technical one, and AppFolio has been more opinionated than its peers about what belongs inside the platform and what stays outside it.
But in a post-AI, agentic world, what does openness look like? And is “which platform is more open” even the right question to ask?
“Openness used to mean access to things. Now we think it means something different, which is really about agency,” says AppFolio CTO Matt Baird. “The most open platforms are not the ones with the most connectors anymore. They are the ones that give customers the most freedom to shape how this new intelligence layer operates within the context of their business.”
In practice, agency means giving clients the option of doing the actual work inside a foundation model like Claude rather than in AppFolio’s own UI. And doing actual work means that Claude, whether human-powered or operating via an AI agent, isn’t just fetching data through a read-only connector, but writing things back to the source of truth as well.
AppFolio’s new connector puts Claude in conversation with AppFolio's own agent, Realm-X. Claude states what the operator wants; Realm-X works out what it needs, asks its own questions, and carries out the task inside AppFolio, under AppFolio's rules. The distinction is the one between handing a contractor the keys to a building along with a copy of every document (data exposure), and calling the property manager, describing the outcome, and trusting them to carry it out correctly within the rules (agent-to-agent).
As AppFolio sees it, this allows operators the best of both worlds: the intelligence of a best-in-class generalist model combined with the sector-specific understanding and guardrails of the property management platform. While a generalist model can read a database and reason an answer, harnesses add value by actually understanding the core business.
“Realm-X understands things that we cannot export,” Baird says. “The real alpha is having it be a conversation between two peers: a high-performing generalist in Claude and a very high-performing specialist in Realm-X.”
This arrangement also limits some of the risks of using foundation models out of the box to perform critical tasks. When calling back to AppFolio, every action the connector takes inherits the same permissions and compliance checks that would apply if a person did it by hand inside AppFolio, so an agent can never reach beyond what the underlying employee is allowed to do. “The guardrails, the permissions, everything the customer has set up automatically cascades,” says Kyle Triplett, Chief Product Officer at AppFolio. Triplett joined AppFolio in 2019 when the company acquired his AI technology platform, Dynasty. “That takes the burden of building outside security and business logic off their plate.”
AppFolio Chief Product Office Kyle Triplett (left) and CTO Matt Baird
To date, most customers are still using the connector for relatively straightforward tasks like reporting and analysis. “Custom reporting, even custom dashboarding, is one of our top use cases today,” says Triplett, describing operators who pull data out of AppFolio and do the last mile of work — an owner report or an investor deck — inside Claude.
More agentic use cases are beginning to take shape. Managing resident churn, for instance, is an example of AI using that business context to make better decisions.
“If you ask AI today who is likely to churn, it looks at when leases are up, because that is the logical decision point,” Baird says. “But the real decision was made seven months earlier, because of noise complaints and maintenance requests that never got fixed on time. That is a broader, more holistic view of the data and the relationships within it.”
Arming the rebels
Strategy aside, adoption is clearly happening: AI actions on AppFolio’s platform are up 7x year over year, a measure of work executed on the operator's behalf rather than typical chat interface querying.
But this technological shift is far more interesting in the context of AppFolio’s position in the market. Much of the company’s traditional customer base sits in the smaller and middle-market tier, multifamily operators who do not employ a CIO to hand-pick point solutions and cannot bankroll the multimillion-dollar systems integrators that large institutions lean on to stitch a stack together.
The timing is fortunate, because the ground is shifting beneath those middle-market operators. Artificial intelligence is conferring significant advantage onto the largest managers, which are able to build the data moats and engineering capabilities that a 10,000-unit fee manager could never assemble on its own. The growing role of centralized back-office teams compounds the problem for small fee managers; a centralized leasing or customer support team is far more efficient at 50,000 units than 5,000.
The pressure is already visible in the wave of rollups and private equity capital entering the property management sector. As centralization and AI make the benefits of scale greater, the logic goes, the sustainability of smaller operators is up for debate.
AppFolio is betting that its technology can help smaller operators keep pace. By drawing on shared training data and capability, a middle-market multifamily manager can effectively rent a data-and-automation moat it could never build alone, which turns the consolidated platform into less of a threat to the small operator than a lifeline against the giants' in-house advantage.
“Part of AppFolio's charter, given the wide customer base we serve, is to give smaller companies access to the same tools that larger companies thrive on,” says Triplett. “Out of the box, you have the opportunity to run a best-in-class business.”
Toward agentic management
Last year, we ran a thought experiment: what would it take to build a multifamily property management firm with zero human employees? While we wouldn’t recommend zeroing out the payroll line quite yet, bidirectional connectors to foundation models push AI in property management past task execution and into the kind of reasoning and decision-making work that would be required to at least move humans a tad further out of the loop.
“I would be disappointed if in 12 months we did not have workflows and reasoning models operating on the operator's goals, not just on how they have codified their business,” says Baird. “If you are trying to increase occupancy, the harness and the foundation model can say, I can change your workflow this way, or make this decision differently, as I guide the hands of the people in your business who are using me as a copilot.”
“A CEO of a real estate operating firm comes in on Monday and gets not a report full of analytics, but genuinely great ideas for small tune-ups to how he runs his business,” adds Triplett. “Even something as small as, you should send that text three days after instead of five. He looks at it and says, I will make these changes, but not those. The software helps him get a little better every week, because it is looking at his last week, last month, year over year, and merging it all together.”
The same trajectory reshapes the interface itself. As more of the work moves into natural language and delegated agents, the traditional web portal, dense with its toggles and fields, begins to lose its claim on the operator's attention.
“More and more customers will become UI-agnostic,” says Triplett. “So much of their work is done in a terminal-style or natural-language experience that what made UI a fun challenge in the past begins to change. Some number of customers will move into the bucket of UI being a second thought. It is not the core thing they are navigating between.”
In its place, the interface becomes something summoned for the moment rather than a permanent fixture.
“There will be a lot more ephemeral UIs created to solve a specific problem in the moment,” says Baird. “They exist for a short time and then go away. You say, I want to do this, and suddenly a UI appears that lets you do that and just that, tied into all the guardrails.”
As decision-making grows smarter and more automated, the economics of multifamily management will shift with it. The operators rolling up the market may come to compete less on their access to capital than on their ability to point automation and intelligence at an ever-larger book of units. After all, platforms like AppFolio don’t solely scope their ambitions to the middle market. “It supercharges great operators too,” says Triplett. “Both things can be true. There is an opportunity for companies ahead of the curve to massively level up.”
But if high-quality reasoning with industry context becomes a commodity, it remains to be seen how managers will continue to differentiate. “If expertise is essentially free, what is the new constraint on growth for these smaller organizations?” Baird asks. “That is a serious question, and I do not know the answer yet.”
Openness isn't dead, but what it means to be open will look very different in 12 months than it does today.
Brad is the founder and editor of Thesis Driven (www.thesisdriven.com), a publication covering innovation in the built world. Over the past decade he has co-founded General Assembly and Common which generated a combined $500M+ in enterprise value.
