Real estate automation: which tasks to automate first and which ones not to.

Before choosing the technology, identify what work you want to reduce

When a real estate company begins exploring automation, it is easy to start with the tools: artificial intelligence, CRMs, WhatsApp automations, assistants, integrations, or new platforms. However, adding more technology does not necessarily make the company’s processes more efficient.

Technology adoption is growing rapidly across the real estate sector, but so are the challenges of turning that technology into tangible results.

In JLL’s Future of Work 2026 survey for Latin America, conducted with organizations in Argentina, Brazil, Colombia, and Mexico, AI-driven workplace transformation emerged as the top strategic priority for 47% of participants. At the same time, it was the area in which organizations reported the lowest level of preparedness.

The gap between adopting technology and actually generating value from it can also be seen globally.

JLL’s Global Real Estate Technology Survey 2025, which included more than 1,500 real estate decision-makers, found that nearly 90% of surveyed organizations were already experimenting with artificial intelligence, yet only 5% had achieved all the objectives they had set for those initiatives.

Evaluate how work is currently done

Before asking which tool to implement, it may be more useful to look at how the team currently works.

Is there information that someone copies every day from one tool to another? Are the same records created manually every time a new prospect comes in? Are there tasks that depend on someone remembering to complete them? Are there steps that follow essentially the same rules in every transaction?

These questions help shift the starting point from technology to the process itself. Automation can make sense when it removes repetitive work, reduces errors, or eliminates unnecessary manual steps. But that does not mean every activity within a real estate company should operate without human involvement.

Before automating, the first step is to identify which work is repetitive, what it costs to keep doing it manually, and where human judgment is still necessary.

The best candidates tend to share the same patterns

Not every time-consuming task is automatically a good candidate for automation. Some are repetitive and predictable; others may look similar from the outside but change constantly depending on the client, the property, or the circumstances of a transaction.

When analyzing the use of robotic process automation in Corporate Real Estate & Facilities Management, Deloitte identifies several characteristics that make a process especially suitable for automation:

• Highly transactional

• Time-consuming

• Scalable

• Based on low-complexity rules and sufficiently standardized

It also notes that this type of automation works best when there are few exceptions and the steps can follow clear rules without continuously depending on human judgment.

Applied to the day-to-day work of a real estate company, this makes it possible to look at a task from several angles before trying to automate it.

For example, it can be useful to ask:

• Does it happen frequently? Creating records, sending notifications, or moving certain data may happen several times throughout the day.

• Do the steps usually repeat in the same way? The more stable the process is, the easier it becomes to define what should happen after each event.

• Are there clear rules? If the task can be described using relatively precise conditions—for example, when A happens, do B—there is a stronger basis for automation.

• Is the necessary information already available? An automation needs data or events that allow the process to start and be completed correctly.

• Is there too much manual intervention for a predictable outcome? Copying information, creating a record, or sending an alert are examples of activities whose outcome can often be defined in advance.

• Are exceptions frequent? If almost every case requires interpreting a different situation, full automation becomes more difficult.

Let’s put it to the test

Consider two situations.

A prospect completes a form for a property. From there, the team creates the record, identifies where the prospect came from, assigns them to an agent, and sends a notification.

If that sequence repeats under the same conditions, there is a relatively stable process that can be evaluated for automation.

Now compare that with a conversation in which a property owner wants to change the asking price after receiving several offers. There may be data that helps inform the decision, but the circumstances, negotiation, and professional judgment carry much more weight.

The difference is not simply how much time each task takes. It is how much of the outcome can be defined in advance through sufficiently clear information and rules.

That is why, before automating an activity, it helps to describe three things precisely: what triggers it, what information it needs, and what should happen next. If those answers remain consistent each time the process repeats, it is probably a stronger candidate for automation.

When a person ends up acting as the connection between two systems

Some of the repetitive work inside a real estate company may not really belong to sales, property acquisition, or customer service. Sometimes it exists simply because information that already lives in one tool also needs to appear in another.

A prospect completes a form and someone later creates the contact in the CRM. A property visit is confirmed and someone manually adds the event to the calendar. A document arrives by email and someone has to download it, identify which property it belongs to, and save it where the rest of the team works.

In those cases, new information is not necessarily being created. Much of the time, existing information is simply being moved, recorded, or updated.

That is precisely one of the common uses of application integrations.

HubSpot applies this principle to information stored in a CRM. Its data sync functionality allows applications to be connected and synchronization to be configured in one or both directions so that, for example, contacts, companies, or deals created or modified in one system can also be updated in the other.

What does this mean for your real estate company?

For a real estate company, this makes it possible to look at some everyday processes differently.

For example:

Form → CRM
A prospect submits their information and, if the defined conditions are met, their record is created or updated.

CRM → notification
A prospect is assigned to an agent and the corresponding team receives an alert.

Confirmed property visit → calendar
An activity that already has a date and time can generate the corresponding calendar event without someone having to enter the same information again.

Email → property file
A received document can trigger a process to classify it or associate it with the corresponding record.

Record change → another system
A relevant update can be transferred to a connected tool without relying on someone remembering to copy it manually.

These are examples of possible processes. What should be automated and how it should be done depends on each organization’s tools, rules, and controls.

Connecting two applications also does not mean that absolutely everything should be synchronized. HubSpot, for example, allows teams to define the direction of the sync, determine which records participate, and configure how specific fields are mapped between applications.

A new prospect should not trigger a chain of manual tasks

When a new prospect comes in, receiving their information is often only the beginning. After that, the team may need to create or update the record, assign it to someone, generate a task, notify the team, and define what should happen next.

Many of those actions can already be configured through rules inside modern CRMs. Salesforce, for example, allows organizations to use lead assignment rules to automatically determine which user or queue should receive a prospect based on criteria defined by the organization.

HubSpot offers a similar principle through automation. A workflow can begin when a specific event occurs or when a record meets certain conditions and then execute different actions.

This makes it possible to think about the prospect journey as a sequence of events rather than a list of actions someone has to remember to complete manually.

For example:

A new prospect comes in
→ their record is created or updated.

They meet certain conditions
→ they are assigned to the appropriate agent.

The assignment is completed
→ the agent receives a notification.

The prospect requires follow-up
→ a task is created and associated with their record.

How does this apply to your real estate company?

In a real estate company, the criteria used for these automations will depend on how the team operates.

An agency might decide to assign prospects based on area, type of transaction, project, agent availability, or other properties recorded in its systems. What matters is that the rules reflect the way the organization actually works.

This also does not mean that every prospect should follow exactly the same journey.

That distinction matters because automating follow-up should not mean treating every prospect as if they were the same.

What can be automated is part of the operation around the prospect:

• making sure their record exists;

• making sure they have a defined owner;

• making sure the right person knows they arrived;

• creating tasks when certain actions occur;

• using relevant changes to trigger the next step defined by the process.

The conversation itself, the analysis of what the client needs, and the decision about how to move forward may still depend on the agent.

The goal is not to remove the professional from the process. It is to prevent predictable administrative activities from depending on someone remembering to complete them every time a prospect enters the pipeline or changes.

That is why a useful question is not simply “Can we automate follow-up?”, but something more specific:

What should always happen after a prospect enters, gets assigned, or changes?

The operation does not end when a prospect enters the CRM

Real estate automation is often associated with prospect acquisition and follow-up, but a significant part of the operational work continues around the property itself.

During property acquisition and marketing, new photos, owner documents, price changes, availability updates, new commercial conditions, or files that need to be associated with the corresponding property may appear.

Many of these actions can become events that trigger other processes.

Applied to a real estate operation, this could look like:

A document related to a property arrives
→ identify which record it belongs to
→ associate it with the corresponding file.

Important property information changes
→ update the corresponding record.

A property changes status
→ trigger the actions the organization has defined for that new status.

A new file is added
→ initiate a review or create a task if the process requires it.

This does not mean that a system can correctly determine on its own what to do with every document or change. For an automation to work predictably, it needs enough information to identify the correct record and rules that define what should happen next.

Automation and human involvement

A real estate company could automate the receipt, initial classification, or association of certain files while keeping human review in place whenever someone needs to confirm their content, interpret a particular situation, or make a decision.

The goal should not be for every change to automatically propagate everywhere.

First, the organization needs to determine which information is the authoritative source, which changes are relevant, and which systems need to know about them.

From there, the process can leave a clear point for human intervention, with cleaner processes, better-organized data, and a clearer view of the information needed to make decisions more quickly.

This is particularly important when working with documents and property information because an incorrect automation can also amplify an error.

That is why it can be more useful to begin by identifying simple and verifiable actions:

• associate a file with the corresponding record;

• create a task when a review is pending;

• generate a notification after a specific change;

• update a field when there is a clearly defined source and rule;

• detect when a document or file changes and trigger the corresponding process.

In this way, automation does not replace property management. It handles the administrative steps that can occur consistently around it, while decisions that require interpretation remain under the responsibility of a person.

When the process requires more than following a rule

So far, we have discussed repetitive tasks, records, notifications, documents, and information flows that can be executed under relatively clear conditions. But not every process inside a real estate company works that way.

Automating the execution of an action is one thing. Automating the decision that determines which action should be taken is another.

For example, a real estate company might automate the creation of a task after a specific event. That does not necessarily mean it should also automate a business decision whose outcome depends on incomplete information, a negotiation, or the particular circumstances of a transaction.

The same caution becomes especially relevant when automation incorporates artificial intelligence.

The AI Risk Management Framework from the National Institute of Standards and Technology (NIST) notes that AI systems can operate under very different configurations, ranging from fully autonomous decision-making to systems that simply provide additional information to a person. The framework recommends clearly defining which responsibilities belong to people and which belong to the system, as well as establishing human oversight according to the context and risks of the specific use case.

Applied to a real estate operation, this suggests paying particular attention to tasks such as:

• a negotiation with a property owner or prospect;

• a recommendation that depends on several specific circumstances;

• an exception that falls outside the usual process;

• a situation where the available information is ambiguous or contradictory;

• a decision whose consequences make a review necessary before it is executed.

This does not mean technology cannot participate.

A system can gather information, detect that a specific event has occurred, prepare data for someone, generate an alert, or even suggest an action.

What needs to be defined is how far the automation should go and at what point someone responsible for the process should step in.

What should you ask?

For each process, it is worth asking:

• Is the outcome sufficiently defined in advance?

• What happens when an exception occurs?

• What information might be missing?

• What would be the impact of executing the wrong action?

• At what point should a person review the situation or make the decision?

A real estate company can automate parts of a process without automating it from beginning to end. In many cases, the most reasonable approach may be exactly that: automate the predictable steps while keeping human involvement where context, exceptions, or professional judgment can materially change the decision.

Start with high-value, low-complexity processes

After identifying several tasks that could potentially be automated, another question comes up: where should you start?

Not every opportunity has the same impact or requires the same implementation effort. Trying to automate the most complex process in a real estate company first can consume significant resources before demonstrating whether automation is actually solving an important problem.

Deloitte proposes evaluating automation opportunities across two dimensions: the value automation can generate and the effort required to implement it. In its analysis of automation applied to real estate operations and facilities management, it recommends prioritizing processes that can deliver high value with relatively low automation effort.

This helps avoid an overly simple assumption such as:

This task takes a lot of time, so it should be automated first.

Time is an important variable, but it is also worth considering how frequently the task occurs, how many people are involved, how much manual work it creates, what happens when an error occurs, and how difficult it would be to define its rules correctly.

For example, a real estate company might identify situations such as:

• Creating a record when a new prospect comes in

• Notifying an agent when they receive an assignment

• Creating a task after a specific event

• Automatically associating a document when there is an unambiguous relationship

• Interpreting any document received and deciding what to do with it

• Automating a negotiation involving multiple variables

This list is not intended to establish a universal order of priority. A task that is simple for one agency may be complex for another depending on the tools it uses, the quality of its data, and how clearly its processes have been defined.

In fact, before automating anything, it may be necessary to understand more clearly how the work actually happens.

A company can begin by observing what the team does over several days and identifying activities that repeatedly appear:

• How often does this task occur?

• How much manual work does it require?

• What happens if someone forgets to do it?

• Are the steps usually the same?

• Are there many exceptions?

• Can the tools involved be connected?

• Does the outcome need to be reviewed by a person?

With that information, it becomes easier to distinguish between an automation that looks attractive because it is new and one that actually removes an everyday operational friction.

That is why the first process a real estate company automates does not need to be the most sophisticated one.

It may be something much simpler: making sure a prospect reaches the right place, creating a task when it is needed, associating a document with the correct property file, or ensuring that someone receives a notification without another person having to send it manually.

The ultimate goal is not to accumulate automations.

It is to make sure that tasks that can happen consistently no longer depend on someone having to repeat or remember them every time, while the team preserves its time and judgment for the parts of the work where they are genuinely needed.