When Every Tenant Has a Compliance Expert in Their Pocket

There has always been an uncomfortable truth within the private rented sector. Agents and landlords sometimes get away with compliance failures simply because the tenant doesn't know that something has gone wrong. That doesn't automatically mean somebody has deliberately ignored the law or set out to take advantage of a resident. Residential lettings is already complicated, regulation continues to grow and genuine mistakes happen. There is, however, no escaping the fact that the balance of knowledge has historically sat heavily in favour of the professional.

For a tenant to challenge something, they first need to know that a particular legal requirement exists. They then need to understand whether it applies to their tenancy, recognise that it may have been breached, establish what their rights are and decide what they can do about it. Even after all of that, they need the time, confidence and inclination to actually pursue the matter.

AI has the potential to remove a significant amount of that friction, and I think the consequences for our industry could be much bigger than many agents currently appreciate.

This came up during my recent Viking Chats conversation with Sandra, whose career has involved taking huge amounts of property data and turning them into information that people can actually understand and use. We covered rental evidence, market statistics, rent increases, regulation and some of the unintended consequences that can occur when government tries to influence the behaviour of a housing market. The part I kept returning to afterwards, though, was what happens when technology doesn't simply give consumers easier access to information, but begins interpreting that information and acting upon it for them.

We've spent an enormous amount of time discussing what AI might mean for agents. I think we need to spend considerably more time thinking about what it will mean for tenants.

The knowledge gap is going to shrink very quickly

At the moment, if a tenant suspects something isn't right, there is still a fair amount of work involved in establishing what has happened. They might search Google, visit a government website, read something from Shelter or Citizens Advice, or disappear down a rabbit hole of online forums. They then have to work out whether what they've found is current, whether it applies in the part of the UK where they live and whether their individual circumstances are covered by it.

Now imagine the same resident using an AI assistant that already has access to their tenancy documents and correspondence. Rather than requiring the tenant to understand the legislation, the technology can potentially interpret it for them. Instead of relying on the tenant to remember that something should have happened by a particular date, the system can identify the deadline itself and notice that the relevant evidence appears to be missing.

The resulting interaction could be remarkably simple: "Your landlord or agent was required to provide this by yesterday and I can't find any evidence that they have. Would you like me to contact them?"

That changes things.

The tenant hasn't suddenly developed years of property-management or legal knowledge. They simply no longer need that knowledge in order to identify a potential issue. As agentic AI develops, they may increasingly not need to put much effort into pursuing it either. Technology can potentially identify the problem, explain the available options, prepare the correspondence and, with permission, take the first steps on the consumer's behalf.

Historically, consumer rights in many industries have been softened by friction. People often don't pursue relatively small issues because the time and hassle involved aren't worth it. You can be entirely justified in complaining about something and still decide that spending your evening researching legislation and writing emails isn't how you want to use your time.

If that effort is reduced to reviewing an AI-generated recommendation and pressing a button, behaviour is likely to change.

For letting agents, this is arriving at a particularly interesting moment because the regulatory environment isn't becoming simpler. Quite the opposite. Agents and landlords are facing greater regulation, more obligations and an increasing need to demonstrate that processes have been followed correctly. In that environment, relying on a resident being unaware of their rights isn't simply poor practice. It is becoming an increasingly dangerous operational weakness.

Doing the right thing and proving it are not the same thing

One of the biggest implications, in my view, is that record-keeping is going to become even more important.

A good agency may be doing almost everything correctly and still leave itself unnecessarily exposed if it cannot demonstrate what happened. It is no longer enough to be confident that somebody probably sent the document, made the call or completed the check. If an issue is challenged, the business needs to be able to establish what happened, when it happened and ideally produce the evidence without somebody spending half a day reconstructing events.

That sounds obvious, yet many agencies still operate with important information scattered across multiple systems, individual inboxes, WhatsApp conversations, locally stored spreadsheets and people's memories. Experienced staff often compensate for that fragmentation because they know the portfolio so well. Ask them what happened with a particular tenancy and they can probably tell you.

The difficulty comes when that person isn't available, leaves the business or simply can't remember the details of something that happened nine months ago.

AI is likely to make that weakness much more visible. If a tenant's technology can identify a potential failure in seconds, it is difficult to justify an agency taking three days to establish whether the failure happened because somebody needs to search four different systems and an archived inbox.

This is why I don't really see the coming challenge as an AI problem. It is fundamentally an operational one.

The agencies likely to struggle most may not be those employing bad people or deliberately ignoring regulation. They may simply be businesses that have accumulated years of manual processes, disconnected technology and informal workarounds. Their teams have kept everything functioning through experience and effort, but the underlying operation has never been designed for a world in which every action can be questioned quickly and evidence may be expected almost immediately.

The least technologically prepared agencies could therefore face the greatest pressure. They may receive more enquiries and challenges because AI makes them easier for consumers to generate, while simultaneously lacking the systems and records needed to deal with them efficiently.

Better-informed tenants should be good for good agents

It would be very easy to frame all of this as the next battle between landlords, agents and tenants. I don't think that's particularly helpful, and it certainly isn't where I'd like to see the industry go.

A better-informed tenant should not frighten a good agent. If legislation requires us to do something, our objective should be to do it whether the resident understands the legislation or not. If the business model depends on customers failing to notice when we get something wrong, the arrival of better-informed customers isn't the fundamental problem.

There is also a much more positive side to better information. Used properly, evidence can make difficult conversations more human rather than less.

Rent reviews are a good example, and they formed an interesting part of my conversation with Sandra. Rent is a highly emotive subject because it concerns somebody's home and usually represents one of their largest monthly outgoings. Simply telling a resident that their rent is going up because "that's where the market is" isn't a particularly constructive way to begin the conversation.

I've always preferred to start somewhere else. Ask how they are, whether they're happy in the property, whether they're planning to stay and whether there is anything that needs dealing with. Once you've established that, you can have the conversation about rent.

The evidence then matters.

Sandra's work around rental evidence is useful because it allows the discussion to move beyond a handful of asking prices on a property portal. If comparable homes have actually achieved a particular level of rent, that provides a much stronger basis for explaining why an increase is being proposed.

The resident may still disagree. They certainly don't have to be delighted about paying more. But a reasonable proposal supported by credible evidence is fundamentally different from an arbitrary number accompanied by a demand.

As regulation grows, I think substantiated and reasonable proposals will become increasingly important. They give both sides something objective to discuss and start what might otherwise be a confrontational interaction from a position of mutual respect.

There is a nice contradiction here too. We often worry that more technology and data will make the industry less human. In situations like this, good evidence can actually help preserve the human element because less time is spent arguing about the basic facts.

AI will also create plenty of rubbish

Of course, giving everybody access to an AI compliance expert doesn't mean that expert will always be right.

AI can misunderstand context, work from incomplete information or apply the wrong interpretation. A resident may inadvertently provide inaccurate information and receive an answer based on it. As these tools become capable of initiating correspondence rather than simply answering questions, agents are likely to receive a mixture of legitimate complaints, misunderstandings and potentially quite a lot of automated noise.

We need to prepare for that too.

If AI contributes to a significant increase in incoming queries, I don't think the sensible response is simply to throw more people at the problem. There is a perfectly reasonable case for agents using their own technology to perform an initial triage. An AI layer could identify the nature of an enquiry, locate the relevant records and deal with straightforward requests or prepare an initial response for review.

If a resident's AI asks whether a document was served, there is no particular value in an experienced property manager manually searching through a tenancy file if the agency's systems can locate the document, confirm when it was sent and provide the evidence almost instantly.

I would, however, be cautious about allowing that process to become completely autonomous. The last thing anybody needs is a tenant's AI and an agent's AI arguing endlessly with one another while a relatively minor issue becomes increasingly confrontational.

Any automated response system needs careful oversight. The information has to be accurate, escalation needs to be sensible and there should be clear points at which a person steps in because the situation requires judgement, empathy or a proper conversation.

The objective isn't to automate conflict. It's to deal with the routine efficiently enough that our people have time for the situations where being a person actually helps.

What the rental market tells us about evidence

Our conversation also moved into the wider rental market and the consequences of trying to regulate market behaviour. This may sound like a separate issue, but I think it demonstrates exactly why good evidence matters.

Take the decision to prevent prospective tenants bidding above the advertised rent. The intention is understandable. In a market where demand significantly exceeds supply, nobody particularly likes seeing people forced into increasingly frantic competition for a home.

The difficulty is that we operate fundamentally within an open-market economy. Whether somebody is buying oranges, alcohol, a holiday or a house, supply and demand influence value, and consumers ultimately decide whether the price being asked is one they're prepared to pay.

Housing doesn't stop responding to those forces because legislation changes the way a transaction can take place.

Sandra shared some particularly interesting historical data during our conversation. Across the wider dataset, around one in ten lettings had agreed above asking rent, while at the height of competition in London that figure had risen to around four in ten.

Those higher offers weren't created by the mechanism of bidding itself. They were a symptom of the demand that existed for those homes.

If an agent can no longer accept an offer above the advertised rent, the obvious risk is that pricing behaviour changes at the beginning of the process. Agents and landlords may become more cautious about marketing a property too low because they know there is no opportunity for competing demand to establish a higher figure. The property may therefore be advertised at a more ambitious price and reduced if consumers don't respond.

People still ultimately decide what they're prepared to pay. We have simply changed the direction from which price discovery occurs.

Couple that with restrictions around rent in advance at the start of a tenancy and there is arguably even greater focus on the headline rent itself. Whether these changes ultimately produce higher asking rents remains to be seen, but they demonstrate why accurate market evidence is going to matter so much. Asking prices alone may become an increasingly imperfect measure of what is actually happening.

Good policymaking needs to understand achieved outcomes, not just advertised intentions. So does good agency.

Compliance needs to be part of the process, not an extra job

There is a wider operational issue that runs through all of this.

In too many businesses, compliance still sits alongside the process rather than being embedded within it. Somebody completes the operational task, then records that it happened, ticks a box, saves a document somewhere and hopes the evidence can be found if it is ever needed.

That creates unnecessary work and unnecessary risk.

As the regulatory environment becomes more demanding, the better approach is to design compliance into the workflow itself. Completing the correct action should create the record automatically wherever possible. The operational event and the audit trail shouldn't be two separate jobs requiring somebody to remember both.

This is where technology can genuinely improve the industry. Not because another dashboard full of green ticks is particularly exciting, but because well-designed systems reduce the opportunity for important steps to be missed and create reliable records as part of doing the work.

It is also closely aligned with how we think about tenancy conclusions at The Depositary.

The end of a tenancy can become contentious because the parties often start from very different perspectives. A landlord may believe a deduction is entirely justified, while the resident may feel equally strongly that it isn't. The answer isn't to make information difficult to access or to rely on one party understanding the process less well than the other. A better process brings the relevant information together, creates transparency around what is being proposed and why, and gives everybody a clearer route towards an outcome.

Evidence doesn't guarantee agreement, but it generally improves the quality of the discussion. If the matter does escalate, having a clear record of what happened is considerably better than trying to reconstruct events from memory and an assortment of email chains.

The best preparation for AI is probably quite boring

Whenever a major technological shift arrives, there is an understandable temptation to focus on the exciting bits. We debate which jobs AI might replace, what new tools agencies should buy and whether property management will eventually become autonomous.

Those are interesting conversations, but for many letting agents I suspect the most valuable preparation for the AI age is far less glamorous.

Get your processes in order. Know where your information is held and make sure important actions are recorded consistently. Reduce your dependence on individual inboxes, memory and locally stored spreadsheets. Integrate systems where it makes sense, automate repetitive tasks where the technology can reliably do them, and make sure the business can produce evidence without launching an internal investigation every time somebody asks a question.

Most importantly, don't wait for a complaint to discover whether you can prove you're compliant.

Businesses that get those fundamentals right will be much better prepared for an environment in which consumers have increasingly sophisticated technology working on their behalf. They will have structured records, clearer processes and the ability to respond quickly when something is challenged.

Businesses that don't may discover a rather different version of the AI revolution. AI might not replace their employees at all. Instead, it could create so much additional administrative work that already stretched teams find themselves struggling to keep up.

That isn't the futuristic AI story that gets people onto conference stages, but I suspect it's a much more immediate risk for a lot of agencies.

More technology should allow us to be more human

There is one final point that I think is important because none of this needs to result in a colder, more adversarial rental sector.

Quite the opposite.

If technology can establish facts quickly, people can spend more time dealing with the circumstances around them. If our records can clearly demonstrate what happened, less time needs to be wasted arguing about the past. If good data can show that a proposal is reasonable, the conversation can start from evidence rather than suspicion.

That's where I'd like to see the industry go.

AI will increasingly give tenants access to knowledge that once sat mainly with professionals. Before long, those same tools may be capable of acting on that knowledge with very little effort from the resident. In an increasingly regulated sector, that is a significant shift and one agents need to take seriously.

I don't think the answer is to be frightened of better-informed tenants. Nor should we look for ways to make it more difficult for them to challenge us. We should make sure our own businesses can withstand the scrutiny.

That means better processes, stronger record-keeping, clearer evidence and systems capable of showing not only that we're doing the right things, but that we can prove we did them.

The irony is that getting those foundations right should also give our people more time to do the things technology isn't particularly good at. They can listen, explain, exercise judgement, resolve complicated situations and have proper conversations with landlords and residents when something isn't straightforward.

For all the talk about AI making property less human, perhaps the opportunity is to use it to remove the administrative noise that so often gets in the way of being human in the first place.

If every tenant is going to have a compliance expert in their pocket, good agents shouldn't be trying to hide from it. They should be building businesses that are ready for it.