AI readiness for property firms: the checklist before you spend anything

The short answer. A property firm is ready for AI when four dull things are true: someone owns it by name, your files and permissions are tidy enough that a tool pointed at them would not embarrass you, there is a one-page policy telling people what is allowed, and you have picked one repetitive job worth fixing. Nothing on that list is strategy, technology or budget. Firms that do those four first get value in weeks. Firms that start with a tool spend a year and a licence fee proving that nobody had the time.

Search "AI readiness" and you get assessment frameworks written for organisations with a chief data officer and a change management function. If your firm has forty people, one long-suffering office manager doing IT and a property management system bought in 2016, none of that is aimed at you. This is the version for the firm you have.

Readiness is housekeeping, not strategy

The blocker in a small property firm is almost never the technology and almost never the money. It is that nobody has an afternoon.

Enterprise readiness models score you on data maturity, governance, talent, culture and infrastructure. Those categories are not wrong, they are the wrong altitude. A twenty-person lettings business does not have a data strategy problem. It has a shared drive with nine folders called Admin, a system nobody has cleaned since a director left, and three people who would use AI properly if someone showed them how on their own work.

That is good news. Everything on the list below is achievable by one motivated person in a fortnight, and none of it requires you to buy anything.

The checklist, in order

Work top to bottom. The order matters more than the items.

1. Name an owner

One person, named, with time protected for it. Not a committee, not "IT", not the managing director in principle. Every stalled AI effort I have seen in a small firm traces back to this: everyone agreed it was a good idea and nobody owned it.

The right person is usually not the most technical. It is the person who knows where the work goes, is trusted by the team, and is organised enough to follow through. Give them a few hours a week and the authority to make decisions.

2. Find out where the hours go

Ask everyone for the three jobs that eat the most of their week. Do not lead them, do not mention AI, and do not make it a form nobody fills in. A short conversation with each person is better data than any survey.

Patterns appear fast. In a property business it is usually some version of the same list: chasing, re-keying between systems, answering the same questions, drafting the same documents, and pulling reports together by hand. That list is your roadmap, and you have it in a morning.

3. Tidy your files and permissions

This is the step that gets skipped and the one that hurts.

Any AI tool pointed at your own documents inherits your existing permissions. If everyone technically has access to the folder holding the salary review, the partnership accounts or the sensitive client matter, the tool will find it and surface it in an answer, and it will look like the tool leaked something when in fact it revealed something that was already true.

Before anything sees your files:

Work out which shared locations are open to everyone and whether they should be.
Find the folders containing personal data about staff, tenants or clients, and check who reaches them.
Deal with the leavers still holding access, which almost every firm has.
Decide where the documents you want AI to use are going to live, and get them there.

Nobody enjoys this. It is also the single highest-value fortnight in the whole exercise, and it is worth doing whether or not you ever adopt AI.

4. Write the one-page policy

One page. Longer will not be read.

Which tools are approved, and which are banned.
What is allowed in a prompt and what never is. Be concrete: client bank details, tenant financial circumstances, anything about an identifiable individual you would not email externally.
The rule that AI drafts and a person sends, for anything leaving the firm.
The rule that AI never states a fact. Every number, date, measurement and legal reference gets checked against a source document.
Who to ask when someone is unsure.

The purpose is not compliance theatre. It is to stop the well-meaning person pasting a client's file into a free consumer account because nobody had told them not to. That is the realistic failure, and it happens in week one without a policy.

5. Check your data protection position

Point your data protection responsibilities at this before you deploy, not after.

Get advice on whether an assessment is needed for what you are planning, particularly if AI will touch decisions about identifiable people, tenants, applicants and staff included. Update your privacy notices if what you do with personal information changes. If you use a third party, know where the data goes and what happens to it. If your firm holds a security certification or clients ask about your controls, work out what changes.

I am not going to summarise the current rules for you, because they move and because a summary in an article is not a defence. Take proper advice on your own circumstances. What I will say is that this step is much cheaper before rollout than after.

6. Pick one job

One. The temptation is a programme covering six departments. Resist it.

The right first job is repetitive, low-judgement, annoying, and owned by someone who wants it fixed. Arrears chasing, tenant and applicant queries, landlord reporting, lease abstraction and file notes are the usual candidates in property. Pick the one where you already know the hours, so the result is arguable with numbers rather than impressions.

7. Train the people, then build

In that order, and the order is the whole thing.

Teams trained first understand what the tool is doing and adopt what gets built afterwards. Teams handed a finished system they were not consulted on route around it politely for a quarter and then stop. Training also surfaces the good candidates for automation, because the people doing the work know which jobs repeat, and they will tell you once they see what is possible.

What readiness does not require

Three things you are allowed to skip.

A data warehouse. Your data does not need centralising before AI is useful. The first wins come from documents and inboxes you already have.

A new system. Replacing your property management system is a two-year project with a business case of its own. Do not attach it to this.

A strategy document. If the deliverable of your AI effort is a slide deck, something has gone wrong. The deliverable is a job that takes less time than it did.

How to tell you are ready

You are ready when you have answers to these, without hedging.

Who owns this, by name?
What are the three jobs eating the most time in this firm?
If I pointed a tool at your shared drive today, what would embarrass you?
Which tools are people allowed to use, and does the team know?
What is the one job we are fixing first, and how will we know it worked?

Five answers. If you have them, buy something and start. If you do not, that is your next fortnight, and it costs nothing.

Frequently asked questions

How long does getting ready take? For a firm of twenty to a hundred people, a fortnight of one person's attention covers the owner, the hours audit, the policy and the first job choice. The file and permissions tidy-up takes longer and runs alongside everything else.

Do we need to hire someone? No, and hiring an AI person before you know what you want is a reliable way to waste a salary. Give an existing person the remit and some outside help for the parts they have not done before.

What does it cost to get started? Less than firms expect. Most of the checklist above is time rather than money, and the first useful wins usually come from tools the firm already pays for. The spend decision is better made after the first job, when you know what you are buying.

We tried AI and it did not stick. What went wrong? Usually one of three things: nobody owned it, people were handed a tool with no training on their own work, or the first job chosen was too ambitious to show a result. All three are fixable, and the second is the most common by a distance.

Is our firm too small for this? Small firms tend to do better, because a decision reaches everyone in a week. The disadvantage is time, not size, which is why the checklist is built to be done by one person.

Where do we start if we only do one thing? Name the owner and give them the hours. Everything else follows from that, and nothing at all follows without it.

I run this as a session with property teams: the hours audit, the shortlist, the policy and the training, using your own work rather than a generic deck. If you want your team taken through it, book a session.