KINGSTON, Jamaica: Something important happened in September 2026, and it was bigger than the launch of another chatbot. Meta launched Muse, an artificial intelligence agent designed not merely to answer questions, but to actually do things. That distinction matters. Muse can operate through a dedicated virtual computer, use a browser, work across connected applications, send emails, make bookings and pursue longer tasks on behalf of its user. Meta says people remain in control of which services it can access. Within less than two weeks, Muse reportedly reached 2.5 million downloads and overtook ChatGPT in US and Canadian app download rankings. Meta’s share price jumped more than 3 per cent around the excitement, adding an estimated US$56 billion to its market value.
Strip away the Silicon Valley fireworks and the message is simple. AI is leaving the chat box. And real estate should pay attention.
From answering to acting
For the past few years, artificial intelligence has mostly behaved like a very clever person sitting behind a desk. Ask a question and get an answer. Write a listing description. Summarise a valuation. Explain a mortgage. Draft an email. Useful, certainly. But the next generation of AI is moving from information to action.
Muse can already access applications covering email, calendars, payments, shopping and other services. Meta has subsequently announced connections involving Shopify, Stripe, QuickBooks and Canva for businesses. It is also extending Muse into its AI glasses and has unveiled the pocket sized Muse Charm. OpenAI is pursuing the same broader direction. ChatGPT Work can research across websites, work with connected applications and files, and continue multi stage projects until it produces finished documents, spreadsheets, presentations and other outputs.
So this is not really Muse versus ChatGPT. It is agents versus the old internet. And property search is almost perfectly suited to disruption.
Goodbye, 27 open tabs
Buying a home remains one of the strangest consumer experiences around. Someone looking for a property may visit several websites, scroll through social media, message three agents, calculate mortgage repayments, search Google Maps, investigate schools, compare neighbourhoods and then try to remember which house had the leaking ceiling and which one had the impossible driveway.
AI agents could compress much of that into a conversation. “Find me a three bedroom home within 30 minutes of Kingston for less than J$40 million.” “Remove anything without parking.” “Compare the remaining five.” “Estimate monthly mortgage payments.” “Check the journey to work.” “Tell me which listings appear overpriced.” “Contact the agents and arrange viewings.”
The buyer has suddenly moved from searching databases to delegating an objective. That is a fundamental change. The traditional property portal assumes humans want thousands of listings. The AI agent assumes humans want the right answer. Those are not the same business.
Real estate’s next gatekeeper
For decades, estate agents worried about property websites replacing them. Then websites worried about Google. Now Google, property portals and agents may all have to worry about the software sitting between the buyer and the internet.
If an AI agent searches 40,000 properties and recommends six, the six matter enormously. So does the question of why the other 39,994 disappeared. Was the price wrong? Was the information incomplete? Was the listing duplicated? Did the AI distrust the source? Was the location poorly described? Or did another platform simply structure its information better?
The future battle in real estate may therefore be less about who has the prettiest website and more about whose data machines can understand. That means accurate prices, currencies, coordinates, property types, availability, floor areas, ownership information where appropriate, agent credentials, photographs, modification dates and clear listing status.
Bad data will not suddenly become good because AI reads it faster. Garbage in. Luxury apartment out.
Jamaica is more ready than it looks
It would be a mistake to dismiss this as something for London, New York or Silicon Valley. Jamaica is already a highly connected society.
DataReportal estimated 2.54 million internet users in Jamaica at the end of 2025, representing 89.5 per cent of the population. There were approximately 3.18 million mobile connections, equivalent to 112 per cent of the population, reflecting people with multiple connections. Jamaica also had about 1.81 million social media user identities.
Instagram’s potential Jamaican advertising audience alone was around 1.10 million, while Facebook’s stood at approximately 928,000. That matters because Meta owns Instagram, Facebook, Messenger and WhatsApp.
Picture somebody watching an Instagram video of a house in St Ann. Instead of clicking a link, opening a browser and completing a form, the future interaction could become: “What house is that?” “Can I afford it?” “Find something similar J$8 million cheaper.” “Show me the neighbourhood.” “Message the agent.”
That is not science fiction anymore. The individual pieces already exist.
Jamaica also exposes AI’s weakness
The Jamaican property market, however, has characteristics that should make anybody cautious about handing everything to a machine. Properties can be described differently across several listings. Addresses may be vague. Road conditions cannot always be understood from a map. Titles, access arrangements, boundary issues and infrastructure can require human investigation.
And sometimes the most important information about a property has never appeared online. A local professional may know that a road regularly floods, that access becomes difficult after heavy rain, that infrastructure remains unfinished, or simply that an asking price bears little resemblance to transactions nearby.
An algorithm cannot retrieve information nobody has recorded. That is where humans remain important.
Agents are not dead. Lazy agency might be.
AI will probably automate a substantial part of what estate agents currently regard as work. Lead qualification. Routine enquiries. Appointment coordination. Property comparisons. Follow up emails. Mortgage calculations. Listing descriptions. Market research. Basic neighbourhood summaries.
That does not necessarily eliminate the agent. It eliminates the value of being the person whose principal contribution is forwarding information from one person to another.
Human value moves upwards. Negotiation. Trust. Inspection. Local knowledge. Strategy. Due diligence. Reading people. Recognising when something simply does not feel right.
The best agent may eventually have an AI doing the work of several assistants while the human concentrates on decisions requiring experience. That is a plus. The minus is obvious. Fewer administrative tasks may mean fewer administrative jobs.
Then there is the uncomfortable part
Muse is impressive, but agents introduce risks that ordinary chatbots never carried. A chatbot can give you the wrong answer. An agent can potentially do the wrong thing.
Reuters reported that Muse’s development included reliability and security concerns, with Meta delaying its original planned launch while additional safeguards were developed. That distinction becomes serious when property, money and contracts are involved.
Would you allow an AI to negotiate an offer? Transfer a deposit? Choose a mortgage? Reject properties on your behalf? Decide which neighbourhood you should live in?
At some point convenience collides with accountability.
Is this becoming godlike?
No. Muse is not God. It is software. It requires chips, electricity, servers, networks, training data and human created infrastructure. Switch those off and the miracle gets very quiet.
But there is a more interesting question. What happens when something remembers what you want, knows your finances, sees what you see through smart glasses, reads your calendar, searches the world’s information and begins taking actions before you even ask?
Meta openly uses the language of “personal superintelligence.” The danger is not that AI becomes divine. It is that humans begin treating it as infallible.
That matters enormously in real estate because a home is not another Amazon purchase. It is shelter, debt, land, family, location and often the largest financial commitment a person will ever make.
AI can narrow the search. AI can crunch the numbers. AI can even arrange the viewing. But somewhere between “find me a house” and “this is where I will build my life,” human judgment still deserves a seat at the table.
Muse’s launch is therefore not the death of real estate. It is a warning that the interface to real estate is changing. The next property revolution may not happen inside the house.
It may happen before the buyer ever reaches the front door.
Visit our YouTube Community ↗