AI tools for commercial real estate fall into seven jobs: general assistants such as Claude and ChatGPT, data and comps platforms, lease abstraction, underwriting, CRE CRMs, public-record connectors, and document drafting. Most commercial real estate teams pair one general AI assistant with one or two specialist tools, and check each tool's data source and license terms before trusting it.
The AI in Real Estate Field Guide puts four applied CRE workflows, an eight-question readiness scorecard and a 30-day pilot plan in one 18-page PDF.
What kinds of AI tools do CRE teams actually use?
Seven jobs cover most of it. The vendors named below are examples verified to exist and offer that kind of product as of September 2026. They are listed alphabetically within each row and are not a ranking or an endorsement.
| Job | What the tool does | Examples | What to check |
|---|---|---|---|
| General assistant | Reads, drafts, analyzes files and connects to other systems. | ChatGPT, Claude, Microsoft 365 Copilot | Plan-level data terms, which systems it can reach. |
| Data, comps and integration | Sales and lease comps, property and owner records, market analytics, or joining those sources into one model. | Cherre, CompStak, CoStar, Crexi Intelligence, Reonomy | Coverage in your market, and license limits on AI use. |
| Lease abstraction | Extracts key terms and dates from lease PDFs into structured records. | Prophia | Links from each field back to the clause, human review. |
| Underwriting and extraction | Spreads rent rolls and T12s into models; runs cash flows. | ARGUS Enterprise, Clik.ai | Whether it fits your model template, citations to source pages. |
| CRM and deal management | Contacts, pipeline, listings and deal workflow. | Buildout, Dealpath, mortr | Brokerage versus investment focus, what it connects to. |
| Public-record connectors | Puts county deeds, parcels and owner records inside an assistant. | Custom MCP connectors | Which counties, how fresh, how it cites the record. |
| Document drafting | OMs, BOVs, memos, owner letters and marketing materials. | General assistants, Buildout, mortr | Whether figures trace to a source, and who signs off. |
Most of these vendors now describe their products as AI-powered. That label says little on its own; the useful questions are in the last column.
Where do general assistants like Claude and ChatGPT fit?
They are the layer everyone on the team touches: reading an OM, drafting an email to an owner, cleaning up a rent roll, writing a first-pass memo. As of September 2026 both create spreadsheets, documents and decks directly, and both connect to other systems over MCP, the open standard for plugging an assistant into a data source.
Their limit is data. A general assistant knows nothing about a specific parcel, a specific lease or last quarter's comps until you give it that material or connect it to a source. That is why the rest of this map exists. For a side-by-side of the two, see Claude vs ChatGPT for business teams, and for the broker-specific version, AI for commercial real estate brokers.
Can AI tools use CoStar data?
Not freely. CoStar's public terms of use restrict sharing credentials with AI agents, scraping, and using its data as input to generative AI tools outside CoStar.
That does not make CoStar less useful to the people who license it. It means an assistant should not be logging in or reading exported screens on your behalf. We wrote the longer answer, with the relevant terms, in can Claude connect to CoStar? The same question is worth asking of every licensed data source: read its terms on AI use before you connect anything to it.
What about lease abstraction and underwriting tools?
These are the most mature specialist categories, because the documents are repetitive and the fields are well defined.
Lease abstraction tools such as Prophia read lease PDFs and produce structured abstracts, with fields linked back to the clause they came from. Underwriting tools split into two groups: cash-flow and valuation software such as ARGUS Enterprise from Altus Group, and extraction tools such as Clik.ai that pull rent rolls, T12s and operating statements into a model.
A general assistant can do a reasonable first pass on a single lease or rent roll, especially with a Skill that encodes your firm's fields. Specialist tools earn their place at portfolio volume, where consistency and review workflows matter more. Our notes on rent roll and lease abstraction and AI underwriting go further.
How do public-record connectors fit?
A public-record connector gives an assistant direct, cited access to county records: parcels, owners, deeds and sales. It answers "who owns this and what did it last sell for" from the source, instead of from a licensed screen.
We built one for the Tampa metro. The county records connector covers Hillsborough, Pinellas, Pasco and Polk inside Claude, including ownership, sale history and comparables, and produces branded valuation documents. Outside Florida the same approach works wherever a county publishes its roll and sales as downloadable files; MCP server for commercial real estate covers how one is built. Public records have limits: they are dated, they rarely carry rents or leases, and each county publishes different fields. See Florida property appraiser data for what the roll does and does not hold.
Where does a CRE CRM like mortr fit?
A CRE CRM holds the relationships and the pipeline. The newer ones add research and document generation beside them.
Brokerage CRMs such as Buildout combine CRM with marketing tools, while platforms such as Dealpath focus on investment teams' deal pipelines. mortr, our own product, is a CRM for commercial real estate teams that looks up county property appraiser records, manages contacts and deals, and generates broker opinions of value in the firm's branding, using dated public records.
How should a CRE team choose?
Start from one job the team repeats every week and pick the smallest tool that does it well.
- Name the job and who does it today.
- List the data it needs, and whether you own or license that data.
- Check each license for restrictions on AI use.
- Try a general assistant on the job first, with a Skill for your firm's steps.
- Add a specialist tool only where volume or consistency justifies it.
- Require every tool to show where a figure came from.
More on the approach is in our commercial real estate hub.
Common questions
What are the main AI tools for commercial real estate?
AI tools for commercial real estate cover seven jobs: general assistants, data and comps, lease abstraction, underwriting, CRMs, public-record connectors and document drafting. Most commercial real estate teams pair one general AI assistant with one or two specialist tools.
Can I use ChatGPT or Claude with CoStar data?
CoStar's public terms of use restrict sharing credentials with AI agents, scraping, and using its data as input to generative AI outside CoStar. Read your license before connecting any AI tool to licensed data, and use sources you own or public records instead.
Do I need a specialist CRE AI tool or is a general assistant enough?
For single documents and drafting, a general assistant with a Skill for your firm's steps is often enough. Specialist tools earn their place at portfolio volume, where consistent fields, review workflows and integrations matter.
Can AI pull county property records?
Yes, through a connector built on the county's published data. It can answer ownership, parcel and sale questions with a citation to the record. Public records are dated and rarely include rents or lease terms.
What should I check before buying a CRE AI tool?
Where its data comes from, its coverage in your market, whether its license allows AI use, how it cites the source of each figure, and whether it connects to the systems your team already uses.