I recently reviewed and compared (with the help of multiple LLMs) hundreds of video interview transcripts featuring short-term rental executives, property managers, technology founders, consultants, and specialist suppliers.

Many of you will know these podcast channels and recordings, and the challenge of listening to them all while filtering out the commercial focus, which is sometimes inevitably coloured by the speaker’s interests. Pricing companies see pricing problems. Direct-booking specialists see OTA dependence. Operations platforms see inefficient processes. AI businesses see an agentic future, etc. Taken together, however, the bias begins to average out, and a much clearer picture emerges.

The blunt conclusion is this:

There is remarkably little that is genuinely new in short-term rentals. Is there anything genuinely new in short-term or vacation rentals?

The industry is still discussing dynamic pricing, direct bookings, better guest communication, professional operations, owner retention, property quality, automation, distribution, payments, insurance, screening and regulatory conformity, amongst others. These are important topics, but most were important 10 or even 20 years ago.

What has changed is not so much what a professional manager must do. It is how those activities can now be connected, interpreted and executed. That distinction may define the industry’s next era.

An industry repeating itself. Much of today’s conversation sounds new because familiar activities have acquired new terminology.

• Automated messages have become “AI guest engagement.”
• Revenue management has become “predictive intelligence.”
• A unified inbox has become an “AI communications hub.”
• Standard integrations have become an “open ecosystem.”
• Data dashboards have become “business intelligence.”
• Process automation has become “agentic operations.”

Sometimes these descriptions are justified. Often, they are a polished vocabulary wrapped around an established function, and the interviews contained endless variations of the same advice:

1. Build a brand.
2. Diversify distribution.
3. Increase direct bookings.
4. Improve conversion.
5. Automate repetitive work.
6. Standardise operations.
7. Use dynamic pricing.
8. Understand the numbers.
9. Protect owner relationships.
10. Deliver better hospitality.

All correct but none revolutionary. The approximate assessment of the material was that 55–60% of these transcripts concerned established management principles, another 25–30% covered existing technology improved by AI or automation, and perhaps 10–15% pointed towards genuinely different capabilities. Less than 5% demonstrated a completely new and proven source of bookings or operational transformation at meaningful scale. The STR industry appears not to lack software, but connected understanding.

The genuine shift is towards an AI-mediated business.

The major development isn’t another application doing another task. It’s the emergence of an intelligence layer that can see across the business architecture, understand what is happening, and coordinate the response. Instead of a manager opening multiple dashboards, the future interface looks to be a conversation.

Ask questions such as: “Which properties are falling behind the market, why is this happening, what will it cost us, which owners need to be contacted, and what changes should we make?”

A capable system would not simply produce a persuasive answer. It would draw upon reservations, prices, availability, guest conversations, reviews, maintenance, cleaning, payments, marketing and owner history. It might then prepare rate changes, amend restrictions, generate owner communications, reschedule operations and present the proposed actions for approval. That is materially different from using an LLM to write a guest message. This progression is simple:

Answer → recommend → execute → orchestrate.

Most products remain between answering and recommending. The meaningful transition will occur when systems can safely execute and coordinate actions across multiple functions. But this raises a much bigger question:

Who will control this intelligence layer? The PMS has an extraordinary opportunity and the next STR gatekeeper may already be inside your business.

The property management system began as the operational record for properties, availability and reservations. The larger platforms are now expanding far beyond that role. Modern multi-category PMSs increasingly encompass or influence:

• Channel distribution
• Websites and booking engines
• Payments
• Revenue management
• Guest communications
• Cleaning and maintenance
• Reviews and reputation
• CRM and marketing
• Owner reporting
• Guest loyalty
• Screening and identity
• Insurance and damage protection
• Financial analysis

This creates an opportunity of enormous strategic significance. The PMS already sits between the manager, the property, the guest and the booking channel. It controls the core records, user permissions, integrations and many of the actions that can be performed. If it can add sufficiently advanced AI, it can become more than the system of record.

It can become the system of decision and control, not just a data recorder! The likely evolution is: Reservation system → operating platform → commercial ecosystem → intelligence layer.

Once that happens, the PMS no longer merely supports the property manager’s business. It mediates much of it.

Connected suppliers may be training their future competitors.

For years, PMS companies used integrations to fill gaps in their products. Specialist suppliers brought better pricing, communications, operations, CRM, screening, insurance and marketing capabilities. The PMS benefited because customers could assemble the technology stack they wanted.

But integrations also show the PMS exactly which products customers adopt, which functions they use and where additional revenue exists. They could be considered a reverse Trojan Horse that pays to enter the city! The pattern is increasingly easy to recognise:

1. A specialist connects to the PMS.
2. Customers demonstrate demand for the product.
3. The PMS observes how the category works.
4. It builds, buys or bundles a similar capability.
5. The external connection becomes more expensive or restricted.
6. The native alternative becomes the default.

The specialist may believe it has gained distribution. The PMS may see temporary outsourced product development. A connection does not have to be formally terminated. It can be rendered commercially unattractive through marketplace commission, API charges, limited write access, payment restrictions or preferential treatment for native modules.

Perhaps “Good enough and included” can be more powerful than “superior but separate and extra fees”

This means many standalone STR products have weaker moats than their founders imagine. If their main advantage is a feature, interface or workflow, a well-funded PMS can probably cross the bridge.

The harder categories to absorb are those protected by something beyond software: regulation, underwriting capacity, proprietary data, risk assessment, a substantial network, independent demand, etc. Even then, the PMS may own the customer interface while the specialist becomes an invisible infrastructure provider.

The next industry gatekeeper may not be an OTA.

Property managers rightly worry about dependence on large reservation platforms. Yet a vertically integrated PMS could become equally fundamental to their business.

Major PMSs are increasingly participating economically in:

• The software subscriptions
• Payment processing
• Booking-engine transactions
• Insurance or damage protection
• Guest verification
• Revenue optimisation
• Marketing activity
• Marketplace commissions
• Financing and other services

They also receive a flow of data, expose endpoints to others who can perform actions, and must pay for the privilege. The PMS could therefore become the operating equivalent of an OTA: difficult to leave, commercially embedded and positioned between the manager and essential parts of the transaction.

The lock-in may be deeper because changing a booking channel affects distribution. Changing an embedded PMS can affect almost every part of the business.

Owning the data is not enough.

None of this guarantees that PMS companies will succeed. Possessing data and customer distribution does not automatically produce intelligence. A genuine intelligence layer requires:

• Clean and consistently structured data
• Broad read and write permissions
• A sophisticated AI architecture
• Reliable identity and property mapping
• Real-time operational context
• Explainable recommendations
• Defined approval thresholds
• Complete audit trails
• Reversible actions
• Strong security and privacy controls
• Measurable business outcomes

Legacy systems

Many PMS platforms still contain legacy technology, fragmented acquisitions and data structures built for administration rather than intelligence. Adding a chatbot to an old database does not create an AI operating system.

Nor does acquiring multiple products guarantee integration. A suite can be a collection of modules sharing a logo while the underlying data stays disconnected. The winning PMS will need more than a wide product catalogue. It must make its categories work as one coherent system. That is a substantially harder challenge, and unavoidable conflicts remain.

An intelligence layer should recommend the best action for the manager. A vertically integrated PMS may profit from particular actions. Would it impartially recommend:

• A competing payment provider?
• An external booking engine?
• A different pricing platform?
• Lower use of a product from which it earns transaction income?
• Moving data to another system?
• Replacing the PMS itself?

This creates a distinction between intelligence and commercially directed intelligence. The more revenue streams a PMS controls, the more valuable its ecosystem becomes, but the harder it becomes to claim neutrality. Managers will eventually need to ask not only, “Is this recommendation accurate?” but also, “Who benefits if I follow it?”

Can an independent AI layer survive?

A generic third-party AI product faces a difficult position. To outperform the PMS, it may need to connect to the PMS, pricing, communications, operations, payments, accounting, reviews, CRM, smart devices, insurance and screening. Each integration includes permissions, commercial terms, data mapping and continuing maintenance. The independent layer may also depend on the PMS for access to the most important transactional information.

If its proposition is “one AI interface across your technology stack,” the PMS can potentially reproduce it with better access, native permissions and an existing customer relationship. Independent intelligence providers will need something the PMS cannot easily absorb or restrict:

• Cross-PMS portfolio intelligence
• Proprietary external data
• Independent demand generation
• A regulated or underwritten capability
• A defensible supplier or guest network
• Neutral benchmarking
• Independent compliance and assurance
• AI governance and decision auditing
• A direct relationship valuable enough to withstand platform pressure

Being cleverer will not be sufficient. The company must own an asset, relationship or authority that cannot be recreated by adding another module.

AI discovery may create another gatekeeper, not remove one!

The same caution applies to the guest side of the market. Conversational search could meaningfully improve how guests discover accommodation. Guests may describe what they need rather than navigate pages of filters.

But AI discovery does not automatically weaken existing platforms. AI agents require live availability, prices, property information, reviews, payments, cancellations, identity and support. The largest reservation platforms already possess these assets at scale. AI may become a discovery interface while the same companies continue supplying inventory and processing transactions.

Alternatively, it may place a new gatekeeper above them. Managers should therefore prepare their inventory for machine-led discovery without assuming AI will deliver inexpensive direct bookings. Structured data, verified amenities, clear terms and machine-readable availability will become essential to participate. GEO is important preparation. It is not yet, however, a proven fourth booking channel.

What should managers do now?

Managers should resist distraction by the latest AI-labelled feature and focus on the structure of their future dependency. The important question is no longer, “Does the PMS have an API?”

An API allows systems to exchange defined information and commands. But an AI-mediated business requires something broader: a controlled way for intelligent agents to discover what information and actions are available, understand their context and use them safely. This is why MCP servers and similar agent-access frameworks could become so important.

An MCP server does not replace an API or magically solve poor data. It acts more like a governed doorway through which an AI system can discover and use approved data, tools and actions. Through that doorway, an authorised AI agent might be permitted to:

• Examine reservations, rates and availability.
• Analyse guest conversations and reviews.
• Check cleaning and maintenance activity.
• Compare property and owner performance.
• Prepare pricing or minimum-stay changes.
• Create operational tasks.
• Draft owner communications.
• Update selected information across connected systems.
• Execute approved actions and record what happened.

That could remove much of the technical work involved in connecting an AI independently to every application. Instead of building numerous bespoke integrations, the intelligence layer could access the tools and information exposed through one governed interface. But this convenience creates a new point of control. If the PMS owns the MCP server, it may decide:

• Which data an external AI can see.
• Whether access is read-only or includes actions.
• Which functions are reserved for its own AI.
• Which connected suppliers can expose their information.
• What external developers must pay.
• Whether access can be withdrawn.
• How much operational history can be exported.
• Whether competing intelligence platforms can operate effectively.

The MCP server could therefore become much more than a technical connector. It may become the gatekeeper to the company’s operational intelligence. A genuinely open MCP environment could give managers greater choice. They may retain the PMS as the transactional core while selecting their preferred AI system to analyse information and coordinate activity across pricing, operations, finance, communications and marketing.

A closed environment could produce the opposite result. The manager may technically own the data but remain unable to use it freely outside the PMS provider’s approved systems. This means managers should examine not only whether AI and MCP access are offered, but the commercial and operational rules surrounding them. The important questions are becoming:

• Who owns the underlying data?
• Can the manager export it in a complete and usable form?
• Does the PMS supply a comprehensive API?
• Is there an official MCP or equivalent agent-access layer?
• Can independent AI systems connect, or only the PMS provider’s own assistant?
• What data and tools does it expose?
• Is access read-only, or can authorised systems write and execute actions?
• Can permissions be limited by user, property, function and financial value?
• Are sensitive owner, guest and payment records separately protected?
• Is every automated instruction and resulting action logged?
• Can an incorrect action be reversed?
• Who accepts responsibility when an automated decision causes loss?
• Can the PMS restrict, price or terminate access later?
• Can connected suppliers expose their full intelligence, or only the information the PMS selects?
• Can the business change PMS without losing its business history, AI context and accumulated learning?
• Are recommendations commercially neutral?
• Can the AI recommend competing payment, insurance, pricing or distribution products?
• What happens when a connected supplier becomes a competitor?
• Does the platform genuinely integrate its modules or merely bundle them?
• Is the system improving profitability, quality and productivity—or simply adding functionality?

Managers should also recognise that accumulated AI context may become almost as important the same as the original reservation data. Over time, an intelligence system may learn:

• How managers price different property types.
• Which owners accept investment recommendations.
• Which guests or bookings require additional attention.
• How service problems are resolved.
• Which contractors perform reliably.
• Where operational exceptions occur.
• Which properties consume disproportionate resources.
• Which decisions improve contribution margin.

If this knowledge exists only inside a proprietary PMS intelligence layer, leaving the platform may mean losing far more than booking records. The business could lose years of accumulated operating context.

Data portability must therefore expand into intelligence portability. Managers should ask whether the rules, history, classifications, prompts, decisions and measured outcomes created by AI can be retained and transferred, not simply whether reservations can be downloaded as a CSV file.

The cheapest PMS may become very expensive if it captures payments, distribution, insurance, marketing, agent access and accumulated business intelligence. Equally, the most comprehensive PMS may be extraordinarily valuable if it creates genuine operational intelligence, reduces complexity and allows managers to run better businesses with fewer administrative resources.

The issue is not whether consolidation is inherently good or bad. It is whether managers understand the bargain they are making, and whether access to their own business intelligence is still a right or becomes a service controlled by their technology provider.

The real transformation

After reviewing hundreds of video interviews, I do not believe we are witnessing the reinvention of short-term rentals. The fundamentals remain familiar: A desirable property. The right guest. Appropriate pricing. Reliable operations. Owner confidence. Strong distribution. Genuine hospitality. Sustainable profit. AI does not change those fundamentals. (until robots appear)

What it changes is who can see across them, who can coordinate them and who controls the permissions required to act upon them.

• APIs connected the systems.
• MCP and similar frameworks might connect the intelligence.

The next strategic question is who controls that connection. That is why the major contest in STR may not be between another generation of specialist applications. The industry’s next gatekeeper may not arrive through the front door with a new booking channel. It may already be installed at the centre of the business, with the keys to every other door.

The major challenge in the STR industry has always been that it is easy to enter, hard to scale, in both technology and management: extensive competition, hundreds of countries, fragmented approaches and orchestration. Having it all under one roof makes it easier to win in this market; it provides commercial viability, pairs well with automation, and lets companies improve operations and identify opportunities and leakages.

A Word of Warning

Security must also move to the centre of software selection. Recent system breaches show that every new integration, API, MCP server, and AI agent can expand the attack surface, particularly in platforms that hold guest identities, owner bank details, payment information, access codes, and permission to perform operational actions.

Managers should examine encryption, multifactor authentication, role-based and least-privilege access, supplier security testing, audit records, incident response, data separation, backup and recovery, not simply features and price. As systems become increasingly connected and autonomous, a compromised account may do more than expose data: it could alter rates, redirect payments, unlock properties or communicate fraudulently with guests and owners. Connectivity creates intelligence and efficiency, but without disciplined security, it also concentrates risk.

Glaring Omissions

But there is one element of note before we leave this article-

One striking omission from much of the industry conversation is company culture. Considerable attention is given to other business categories, but far less to the people expected to use these systems responsibly and deliver the hospitality behind them.

Technology adoption must be accompanied by proper training, employee engagement, fair pay and rewards, psychological and employment security, clear accountability and confidence that automation is intended to strengthen the business rather than remove jobs. Staff must understand not only how to operate new systems, but when to question their recommendations, recognise security risks and intervene when automation produces the wrong result.

An unmotivated or insecure workforce will not create better hospitality simply because it has better software; sustainable performance comes from combining intelligent technology with skilled, trusted, and fairly rewarded people.

---