How AI Is Changing Life for Tenants

Short answer: The biggest change AI brings to a tenant's life is balance of information. Until now, one side of every rent negotiation held all the facts and the other side held none. AI can now tell a tenant whether the asking rent sits above the market rate, whether a rent-increase demand is justified, and whether a new flat is genuinely better than their current one. Autolinium's AI team analysed real data from nearly 5,000 flats in Chattogram to show what the picture actually looks like on the ground.

Where the data in this article comes from

Most writing about rent is guesswork. This article is different.

Autolinium built this analysis on data collected by Jilani Home, an established rental service in Chattogram — more than 4,900 property entries recorded between 2020 and 2026. Each entry carries bedroom count, floor, whether the gas is piped or cylinder, water source, rent, and address.

The most valuable part is that the data has continuity over time. When the same flat became vacant again years later, it was listed again. That makes it possible, for the first time, to see how much rent in Chattogram actually rises.

This kind of time-series rental record is rare in Bangladesh. Most platforms hold only today's listings, with no trace of what last year looked like.

Does rent in Chattogram really rise every year?

This is the most widely held belief. The data says it is wrong.

We identified 158 flats that appeared on the rental list twice, at least six months apart. The results:

In other words, more than half of these flats saw no rent increase at all.

And when rent does rise, by how much? Usually 2,000 taka. 15,000 to 17,000. 10,000 to 12,000. Landlords do not think in percentages; they think in round numbers.

The wider market tells the same story. Median asking rent from early 2025 to late 2026:

• 1 bedroom: 6,000 → 6,000 (unchanged)
• 2 bedroom: 11,000 → 12,000
• 3 bedroom: 22,000 → 22,000

Barely any movement in eighteen months. Adjusted for inflation, real rents have slightly fallen.

For a tenant, this single fact is enough. When a landlord says "rents are going up everywhere," that statement is not supported by the data. Knowing it can save 20,000 to 25,000 taka a year.

How much does rent vary within the same city?

Almost a twofold gap inside one city. The pattern repeats for 2-bedroom flats — 14,500 in GEC, 9,000 in Muradpur.

This is where one of AI's simplest but most powerful uses sits. If someone asks 26,000 taka for a 3-bed in Chandgaon, a tenant can now see that this is 37% above the median. Until now there was no way to make that comparison at all, other than asking a neighbour what they pay.

What exactly does AI make easier for tenants?

Checking whether the rent is fair

Enter the area, bedroom count and rent, and the system shows where that figure sits against the market. The negotiating table now has numbers on it instead of guesses.

Working out the real monthly cost

This is where the biggest budgeting mistake happens in Bangladesh. The data shows entries like "14,000 — including gas, water, electricity and maintenance" alongside "8,000 — bills separate." The second looks cheaper. In practice it often is not.

AI reads this messy free text and works out what the month will actually cost — rent, service charge, the extra burden of cylinder gas, all of it.

Making sense of unstructured addresses

In Chattogram an address is not a format, it is local memory — "the second lane off Mayor Goli," "Kalu Sawdagar's house," "the last flat at the end of the alley." Autolinium's analysis found that only 40% of addresses could be mapped to any recognised area name.

This is where AI's language ability earns its place: reading a loose description, identifying the area, and matching it to the market rate.

Asking in Bangla, answering in Bangla

"Two bedrooms under twelve thousand in Muradpur, must have a lift, piped gas" — typed or spoken, the system understands it. No form-filling. Mixed Bangla and English numerals are not a problem.

Saving time

The data shows many flats listed more than once in the same month — meaning nobody actually knows which ones are still available. Status tracking and instant replies are two ordinary features, and they save tenants more time than anything clever.

A new way to compare flats: your current home as the benchmark

Ordinary platforms ask: "Which is the best flat?"

But that is not what a tenant wants to know. What they want to know is: "Is this better than where I already live?"

Autolinium is building on exactly that idea. The structure is simple:

First, the tenant describes their current home — rent, floor, piped gas or cylinder, which way the balcony faces, whether the ground floor is residential or shops, how dense the surroundings are. By voice, by typing, or through a form.

Then they enter the flats they have viewed — two of them, or five.

The system scores each one out of 100 and compares it against the current home, showing which is an upgrade, which is roughly equivalent, and which is actually a step down.

The calculation runs in three stages

Stage one — bringing every feature onto the same scale. Piped gas or cylinder, lift or no lift, balcony direction — these are different kinds of information. Before they can be combined, they have to be measured the same way.

The rent score comes from the most important place of all — comparison against the actual median rent for that area. 22,000 in Chandgaon and 22,000 in Khulshi are not the same thing. Without knowing the local market rate this stage is impossible, and that is precisely where years of accumulated data earn their value.

Stage two — weighting by the tenant's priorities. Whether cost matters most, or facilities, or comfort, is decided by the user. A lift means nothing to a bachelor; to a family with elderly parents it matters more than anything else.

Stage three — subtracting from the current home. The difference is the real answer. +13 means there is good reason to move; -3 means the hassle is not worth it.

The most useful number: the price of an upgrade

This is the part almost everyone leaves out, and it influences the decision more than anything else.

Suppose your current home scores 72. You have seen two flats:

Flat A scores higher. But each point of improvement costs three times as much there. In practice Flat B is the better decision — 72,000 taka less over a year.

That single line does what pages of comparison cannot.

The system does not decide for anyone — it makes the basis of the decision visible.

How much AI is actually involved in this?

One misconception is worth clearing up. Many assume a system like this means handing everything over to AI. It does not, and doing so would make the system unreliable.

The principle Autolinium builds on:

The maths is done by formula. There is no language model anywhere in the three stages above — it is arithmetic. The same inputs always produce exactly the same result. Score a flat with a language model and you would get 78 one time and 81 the next. The moment a user notices that, trust in the whole system is gone.

AI is used in exactly three places: extracting structured information from loose human language or voice, identifying the area from an address, and explaining the result in plain Bangla.

Drawing this line matters. AI's job is to understand and to explain; the arithmetic behind the decision stays in verifiable formulas. That distinction is what separates a demo from a product people can rely on.

What Autolinium does in real estate

Autolinium is an AI-first software company — AI is not bolted on afterwards, it sits at the centre of the system from the start. We have delivered more than 55 projects across over 13 countries, and real estate and property is one of our primary sectors.

What our work in this sector looks like:

Property and rental ERP

Listings, tenants, contracts, collections and maintenance in one system. Our own property management platform, EstateX, is being built for exactly this sector.

Turning data into decisions

The analysis you have just read is a sample of the work itself — converting messy form data into usable market intelligence.

AI voice and messaging agents

Round-the-clock answers, visit bookings, follow-ups. In real estate, most prospects are lost to slow replies.

Custom tools

Rent checkers, flat comparators, cost calculators — solutions to specific problems that off-the-shelf software does not address.

Our position is simple: make use of the data an organisation has already accumulated over the years. Not collecting new data — extracting value from what already exists.

What AI cannot do

Honesty matters here. Some things exist in no dataset:

• What kind of person the landlord is
• Whether bachelor tenants will be accepted
• What the neighbours are like
• Whether the advance deposit will actually be returned

A large share of rental decisions in Chattogram rests on exactly these unwritten factors. AI cannot reach them, and should not pretend otherwise.

So the honest position is this: AI does not decide for the tenant. It only makes sure the decision is not being made in the dark.

Frequently asked questions

Can AI tell me whether my rent is too high?

Yes, provided there is enough real data for that area. Without at least a dozen comparable samples matched on area and bedroom count, a reliable comparison is not possible — and in that case saying nothing is the correct answer.

How does AI help when a landlord wants to raise the rent?

By showing the actual market trend. The Chattogram data shows that 52% of re-listed flats saw no rent increase, and when rent does rise it is typically by 2,000 taka.

Is this kind of system expensive to build?

The core calculations are arithmetic, so no heavy AI model is required. Costs are far lower than most people assume. They only rise when voice input or image analysis is added.

Can Dhaka data be used for Chattogram?

No. Chattogram's rental market has its own character — working families around the port and EPZ, expatriate households in Khulshi and Nasirabad, bachelor messes around the university. Without local data the numbers will be wrong.

Could something like this work for my organisation?

If you hold a few years of transaction or customer records — even in a spreadsheet or a Google Form — then yes. This is exactly the work Autolinium does.

Closing thought

AI is not inventing anything new in real estate. What it is doing is replacing guesswork with evidence.

And the biggest beneficiary of that shift is the tenant, because the information gap has always sat on their side.

There is one condition, though. AI is only as good as its data. We were able to show this picture of Chattogram because a local company patiently collected information for six years. The foundation is not the technology — it is that patience.

Your organisation's data is probably sitting somewhere much like it: in a spreadsheet, in a form response sheet. Turning it into a decision-making tool is what we do.

Autolinium is an AI-first software and automation company — enterprise software and ERP, AI automation, AI voice and messaging agents, and web and mobile apps. More than 55 projects delivered across over 13 countries. Real estate is one of our primary sectors.

To find out how your organisation's data can be put to work, get in touch — autolinium.com