Removing inefficiency that comes from humans having to understand where every piece of information lives
Thought Leadership (Pictured) by Fredérick Wakim, Founder of ImmoAdmin
“For most of the software era, making a property management platform better meant adding something to it. Another dashboard. Another report. Another filter. Another workflow. Another integration. The logic made sense. Property management is complicated, so the software became complicated too. But I think we are approaching a point where that relationship starts to reverse.
The next generation of property software may not win because it gives users more screens. It may win because users need to visit fewer of them. That is a quite different idea from simply adding an AI chatbot to an existing SaaS product.
We built software around navigation
Think about how a property manager completes a relatively ordinary task today. A tenant has not paid. The manager opens the rent module, finds the tenant, checks the balance, looks at previous payments, possibly opens the communication history, prepares a reminder, sends it, then records or follows up on what happened.
None of those steps is especially difficult. The inefficiency comes from the human having to understand where every piece of information lives and how every part of the software is supposed to be operated.
For years, that has simply been accepted as the interface between people and business software. You learn the system. AI potentially flips that relationship. The system starts learning how the work is done.
A chatbot beside the software is not the same thing
There is an important distinction here. putting a conversational box inside a property management product does not automatically make the product intelligent.
If the AI can explain how to create a lease but the user still has to leave the conversation, find the leasing module and create it manually, extraordinarily little has changed. The much more interesting model appears when the AI is connected to the underlying application itself.
A manager should be able to ask which tenants are behind on rent, understand the answer, then instruct the same system to prepare the appropriate follow-up. It should already understand which building, lease, tenant and transaction are being discussed.
That requires far more than a language model. It requires structured data, permissions, business rules, workflows, and a system capable of actually carrying out the resulting action.
This is the distinction we have been working through while building ImmoAdmin.
The platform itself handles operational areas such as leases, rent, accounting, maintenance, documents, and communications. Its AI assistant sits inside that environment rather than beside it.
It can work with the context already present in the system and help users move through real operational tasks, while sensitive actions remain subject to user confirmation. That last part matters more than it might seem.
The future is not AI does everything
There is a temptation in technology to make autonomy the headline. I think that is the wrong target for property management. Property operations involve money, contracts, personal information, and communications that can have legal consequences.
The objective should not be removing the human from every decision. It should be removing unnecessary human navigation and repetitive execution. There is a major difference between an AI preparing an action and an AI being allowed to execute every action without oversight.
Good operational AI needs boundaries. The system needs to know who the user is, what that person is allowed to access, which property they are working on and when confirmation is required.
Without that layer, an AI agent with more capabilities can simply become a more capable way of making mistakes.
Fragmented software becomes an AI problem
There is another consequence that I think PropTech companies should pay attention to. Fragmentation was already inconvenient when humans were operating software manually. It becomes a structural limitation when AI is expected to operate across it.
Imagine the rent ledger is in one product, maintenance is in another, documents are sitting in cloud storage, tenant conversations are in email, and accounting information lives somewhere else.
A human property manager can mentally connect those systems because they understand the business. AI cannot reliably act across that environment unless the context is made available to it.
This makes the architecture underneath the AI increasingly important. You have already argued in Proptech-X that AI is becoming an operating layer and that traditional SaaS could face pressure as intelligence becomes more important than the interface itself.
I agree with that direction. Where I think the next battle gets interesting is this:
Who owns enough of the operational context for that intelligence to be useful?
The value may shift away from having the nicest individual module and toward having the system that understands how the modules relate to one another.
Property management makes this particularly obvious
Property management is a useful test case because almost nothing happens in isolation.
A lease affects rent. Rent affects accounting. A tenant affects communications. A building affects maintenance. A document may affect a legal or financial workflow months after it was originally uploaded.
In Quebec, where we started ImmoAdmin, the operational layer also has to understand local processes such as official Tribunal administratif du logement leases, renewals, notices, and RL-31 tax documents. That complexity is exactly why conversational AI by itself is not enough. The AI needs the system underneath it.
Dashboards will not disappear tomorrow
I do not believe dashboards suddenly vanish. There are many things that are still better visually: financial statements, portfolio comparisons, large tables, document review, and anything where a user wants to scan a lot of information at once.
But the dashboard may stop being the starting point for every action. Instead of thinking, ‘Where do I go in the software to do this? the user increasingly starts with. ‘This is what I need done.’ That sounds like a minor change. It is not.
It changes how products are designed, how features are discovered and potentially how SaaS companies compete. If the intelligence layer becomes the primary interface, users may care less about remembering where a feature sits in a menu and more about whether the underlying system has enough context and permission to complete the task correctly.
This is already moving beyond theory
We are still early in this transition, and I would be suspicious of anyone claiming that autonomous property management has already been solved. But there is enough happening in production to see the direction.
ImmoAdmin is live today. Its founding program represents more than 8,000 units, while Elevate Real Estate operates more than 1,500 units in Quebec as an enterprise client using the platform.
Our Elevate Real Estate case study is one example of what it looks like when a real property management team uses a unified operating platform at portfolio scale. What interests me most is not whether people will use AI in property management. That question is becoming less interesting by the month.
The better question is what happens to property software once the AI no longer sits beside the workflow but becomes the way users enter it. For twenty years we trained people to operate software. We may now be entering the period where the software learns to operate around people.”
Author bio
Fredérick Wakim is the 19-year-old founder of ImmoAdmin, a Quebec-built property management software platform that connects AI directly to operational property management workflows.
