Blog · September 17, 2026 · 4 min read

What AI actually changed about finding tenants, and what it did not

By the Atlas team. Research and practical guides for commercial landlords and leasing brokers, from the people building tenant-search software.

Every leasing tool acquired an AI label in the last two years, which makes the category almost impossible to evaluate. Some of it removed real hours from real work. A good deal of it renamed a search box.

The distinction matters because the money is not trivial and the claims are loud, so it is worth being precise about which part of finding a tenant a machine is actually good at.

What it genuinely does well

  • Reading a location. Assembling what trades around an address, what the category mix is, what the block lacks, and what the neighbourhood rates highly. This is patient, boring work across many sources, and it is the kind of thing software has always been better at than people.
  • Searching wide and ranking. A person can consider perhaps fifty businesses for a unit. A search can consider thousands and return the forty worth your attention, which changes what a shortlist even means.
  • Finding who signs. The licence, the registration, the profile, the verification: about an hour per business by hand, and the reason most outbound stops at the company rather than the person. This is the single largest time saving available, and the argument is in find the franchisee, not the brand.
  • Drafting the first message from real facts. Not generic templates, which read as generic, but a first line built from what is actually true about that business and that unit.
The common thread: every one of those is research, and research is the part of leasing that scales badly with human effort and well with machines. None of them is the part where the deal is won.

What it does not do

It does not know that the landlord two doors down is about to lose their anchor, or that the operator you are emailing fell out with a previous landlord in this town, or that the council is about to change the parking on that street. It does not read a room on a walkthrough, and it does not negotiate.

More practically, it cannot tell you whether a tenant will still be trading in year four. That judgement comes from the questions in how to qualify a commercial tenant and from having watched businesses like theirs succeed and fail. Anyone selling you a model that underwrites covenant strength is selling confidence rather than information.

The adoption gap is the interesting part

Professional adoption of these tools is broad and the reported benefit is much narrower. In NAR’s 2025 Technology Survey, 68 percent of respondents reported using AI tools while only 17 percent said it made a significant difference to their business, and we looked at what that implies in what the data says about prospecting.

That gap is not evidence the technology fails. It is evidence that most of it is applied to tasks that were never the bottleneck. Summarising a document you were going to skim anyway saves minutes. Removing a week of candidate research saves a week.

How to evaluate a tool in one question

Ask what hours it removes, and make the answer specific. Not “it uses AI to improve leasing outcomes”, but “it replaces the two days you spend building a candidate list and the hour per business you spend finding who signs”. The first is unfalsifiable. The second you can check against your own calendar after a month.

Then check the output the way you would check a junior’s work: does each recommendation come with a reason you can evaluate, or only a score? A ranked list without stated reasoning cannot be argued with, which means it cannot be trusted or improved. The comparison across categories of tool, including where a marketplace or a data platform is still the better buy, is in tenant search tools compared.

Where we sit

Atlas is built on the narrow claim above and nothing wider. It reads your address and the streets around it, searches thousands of businesses four ways, and returns matched tenants with the verified decision-maker at each, with the reason every match surfaced written out so you can disagree with it. Drafted outreach is in development. The negotiation, the walkthrough and the judgement about who to trust remain yours, which is the correct division of labour.

The prospecting platform page sets out the workflow end to end, and the plan is $200 a month. If a week of research is not what stands between you and your next tenant, no tool in this category will help you, and it is worth knowing that before buying one.

Common questions

Can AI find tenants for a commercial property?
It can do the research that precedes finding one: reading a location, searching thousands of businesses against it, ranking which plausibly fit, and identifying who signs at each. The lease is still won by a person who knows the street, the building and the negotiation.
Is AI worth paying for in commercial real estate?
It depends entirely on whether it removes hours you currently spend. Tools that compress candidate research and decision-maker lookup replace work with a measurable cost. Tools that summarise what you already know usually do not, however impressive the demonstration.

Find prospective tenants for your space

Start with your property and leasing criteria. Atlas helps you discover prospective tenants, review the reasons they could fit, and research contacts for outreach.

Run your first search