A letter of intent is the most repetitive document a commercial real estate firm produces and the one it most often drafts from scratch. The same rent structure, the same free-rent and TI language, the same exclusivity and confidentiality carve-outs — retyped deal after deal, then edited by hand through three rounds of counters. Automation removes that retyping. The question is not whether to automate LOI drafting; it is which kind of automation earns its cost. An off-the-shelf generator turns a prompt or a questionnaire into a letter in minutes. A custom template engine encodes your own clause library and deal logic so every LOI comes out consistent, on-brand, and ready to negotiate. One is cheaper to start; the other is cheaper to run at volume. This guide draws the line between them for a 4–20 person firm.
The short answer
Use an off-the-shelf tool when your LOI volume is low, your terms vary widely deal to deal, and a person reviews every letter anyway. Commission a custom template engine when LOIs are a recurring, high-volume workflow, most of the language repeats, and consistency across a team matters enough that hand-editing has become a source of error and delay.
The dividing line is not sophistication. A generic AI generator can write a perfectly good single LOI. The line is repeatability: how much of your LOI language is the same every time, how many letters you send a month, and how much control you need over exact wording. A template engine wins when the answer to all three is “a lot.” Below that, it is an expensive solution to a problem a subscription already solves.
What LOI drafting automation actually means
Two different technologies hide under the same phrase, and conflating them is the most common mistake buyers make.
Generative drafting starts from a blank page each time. You describe the deal — property, price, term, concessions — and a language model composes a letter. Tools like Genie AI, Legitt AI, and the AI chat built into brokerage platforms work this way. The output is fluent and fast, but it is regenerated from scratch on every run, so two LOIs for near-identical deals can come out worded differently.
Template-based assembly starts from a fixed document with your language already in it. You answer a structured questionnaire — deal type, binding vs non-binding provisions, which concessions apply — and the system slots your pre-written clauses into a pre-approved skeleton. Legal document-automation tools such as Gavel and HotDocs work this way, and it is the model a custom template engine follows. The output is deterministic: the same inputs always produce the same letter.
The distinction matters because an LOI is a negotiating instrument, not an essay. You want the exclusivity window and the confidentiality carve-out to read exactly the same every time, because your counsel already blessed that wording. Generative tools optimize for producing text; assembly tools optimize for reproducing your text — the same structural split we map across the document stack in our commercial real estate document intelligence guide.
The off-the-shelf landscape
“Off-the-shelf” is not one product. It splits into three tiers, each aimed at a different buyer, and knowing which is which keeps you from overpaying.
Consumer and standalone AI generators. ChatGPT, Claude, and dedicated web generators (Genie AI, AILawyer, Legitt AI) produce an LOI from a prompt for little or no cost. Best for a firm sending the occasional letter, with a person checking every line.
Legal document-automation platforms. Gavel, HotDocs, and similar tools convert a template into a smart questionnaire that assembles an error-free document. These are the closest off-the-shelf analog to a template engine — you supply the template once, then generate from it — and they suit firms that want assembly without a custom build.
CRE deal-management platforms. Buildout, Dealpath, and Crexi bundle LOI drafting into a broader workflow. Buildout’s platform includes AI chat that drafts LOIs and PSAs and syncs to live deal data (Buildout); Dealpath and Buildout sit on opposite sides of the transaction, buy-side acquisitions versus sell-side brokerage (NextAutomation). If you already run one, its LOI feature may be all you need.
The trap is buying a platform for its LOI generator. That feature is a rounding error inside a five-figure annual contract you would only justify for the CRM and pipeline tooling around it. If deal management is not otherwise on your roadmap, the platform is the wrong door. We work through that broader buy decision in our off-the-shelf versus custom document AI framework.
What a custom template engine is
A custom template engine is your firm’s LOI language, codified into a system that assembles a finished letter from a short set of inputs. It has three parts, and it is far more modest than the phrase “custom build” suggests.
A clause library. Every provision you use — base rent and escalations, free rent, tenant-improvement allowance, options, exclusivity, confidentiality, expense treatment — written once, in the exact wording your principal or counsel approves, and stored as reusable blocks.
Conditional logic. Rules that decide which clauses appear. A ground lease pulls different language than a retail deal; a non-binding LOI includes the standard disclaimer; a buyer-side letter reads differently than a landlord’s. The logic handles those branches so the drafter does not have to remember them.
An input layer. A short form — deal type, parties, price or rent, term, concessions, key dates — that a broker or analyst fills in. The engine does the rest: correct clauses, correct order, your letterhead, ready for a final human read.
Built well, it produces the same reviewed language every time, cuts drafting from an hour of copy-paste to a few minutes of data entry, and leaves an audit trail of which template version produced which letter. The build is a scoped project, not an open-ended one — the kind we size in our breakdown of what a custom document-automation project costs a CRE firm. The point is ownership: the clause library is yours, portable, and does not live behind a vendor’s login.
The five variables that decide it
Score your firm on five variables. Three or more pointing right, and a custom template engine will pay for itself. Three or more pointing left, and an off-the-shelf tool is the right call.
| Variable | Points to off-the-shelf | Points to custom engine |
|---|---|---|
| Volume | A few LOIs a month | Dozens a month, across a team |
| Language reuse | Terms vary widely deal to deal | Most clause language repeats |
| Control | Generic wording is fine | Counsel-approved wording is non-negotiable |
| Consistency need | One person drafts everything | Several people draft; output must match |
| Integration | Letter lives in a doc or email | Must feed a deal-management or CRM system |
The variable that surprises firms most is consistency. A solo broker who writes every LOI carries the standard language in their head, so a generator is enough. A ten-person shop where five people draft LOIs has five slightly different versions of the exclusivity clause in circulation — and that drift is exactly what a template engine exists to stop. The larger the drafting team relative to the firm, the stronger the case to build.
Why negotiation rounds change the math
The hidden value of LOI automation is not the first draft. It is the fourth.
An LOI rarely goes out once. It comes back countered — a lower TI allowance, a shorter term, a changed commencement date — and you redraft, sometimes three or four times before the parties align. Each round is another chance to introduce an inconsistency: a number updated in one paragraph but not the summary, a concession dropped, a date that no longer matches the rent schedule.
A generative tool treats every redraft as a new composition, so each round reintroduces variation you have to proofread out. A template engine changes only the inputs that changed and regenerates deterministically — untouched clauses stay byte-for-byte identical, so the only thing to check is the term you actually moved. Automated LOI workflows can compress negotiation cycles by roughly half, and consistent redrafting is a large part of why (Ironclad). Any comparison that measures only first-draft speed is measuring the wrong thing.
What each option costs
Costs sit in different units, which is what makes the comparison confusing. Here is the honest shape of each, in market ranges rather than any one vendor’s list price.
| Option | Typical cost | What you pay for |
|---|---|---|
| Consumer AI generator | Free to ~$25/user/month | Fast drafting, generic wording, consumer or business data terms |
| Legal document-automation tool | ~$100–350+/month by tier and templates | Questionnaire-driven assembly from templates you supply |
| CRE deal-management platform | Low-to-mid four figures/year and up | LOI drafting bundled with CRM, marketing, and pipeline tooling |
| Custom template engine | Project-based build (market range for small custom automation runs roughly $25–150K depending on scope) | Your clause library and logic, owned and portable |
Document-automation platforms commonly tier from around $99 a month for a handful of templates up toward $350 a month and beyond for larger libraries and multiple seats (Knackly). A custom engine is a larger upfront commitment that removes the per-seat subscription and gives you an asset you own. The crossover is a volume-and-longevity question: subscriptions win when volume is modest or the need is temporary; a build wins when LOI drafting is a permanent, high-volume part of how the firm operates. The same subscription-versus-build tradeoff, applied to lease abstraction, is worked through in our comparison of ChatGPT and purpose-built document AI for lease review.
The confidential-terms question
An LOI contains exactly the information you least want leaking: price, concessions, the identity of a not-yet-public deal. That makes the tool’s data terms as important as its output.
On consumer AI tiers, prompts can be used to train the model by default unless you find and toggle the opt-out. Pasting an LOI’s terms into a free or personal account hands live deal pricing to a general training process. The fix is not to avoid AI — it is to use business-tier accounts (ChatGPT Business, Claude Team, or the API), which do not train on your inputs by default, or a platform with explicit data-handling terms.
A custom template engine sidesteps this cleanly: the clause library and the assembled letter live in your own environment, and no deal terms leave it. For a firm that handles confidential acquisition or leasing data routinely, that containment is a real part of the value. The rule holds across every tool: read the data terms before you type a live deal into anything.
The threshold: when to build
Put the pieces together and the decision resolves cleanly.
Start off-the-shelf when LOI drafting is occasional, your terms genuinely vary, and one person owns every letter. A business-tier AI generator or a document-automation subscription covers that firm for the price of a few coffees a week, and building anything custom would be solving a problem you do not have.
Build a custom template engine when three things are true at once: LOIs are a recurring, high-volume workflow; most of the language repeats and has been blessed by counsel; and several people draft, so consistency has become a real cost. At that point the subscription tools leave errors and rekeying on the table that the build removes, and the owned clause library becomes an operating asset rather than a monthly fee.
Whichever side you land on, one rule does not change: automation drafts, a human owns the binding provisions. The exclusivity window, the confidentiality terms, the enforceable pieces of an otherwise non-binding letter — those get a final read from a principal or counsel every time, because a mis-worded carve-out is a legal exposure a template cannot catch (Texas A&M TRERC). The tool removes the typing, not the judgment.
The industry backdrop rewards firms that pick the tool that fits over the one with the longest feature list. Deloitte’s 2026 Commercial Real Estate Outlook, drawn from more than 850 executives across 13 countries, found only 7 percent reporting transformative AI impact — up from 1 percent a year earlier — while 27 percent are still experimenting (Deloitte). The firms pulling ahead match a specific tool to a specific workflow — the structural speed a lean shop can turn into an advantage, as we argue in the small CRE firm AI manifesto.
FAQ
What is LOI drafting automation?
LOI drafting automation produces a letter of intent from structured inputs instead of manual typing. It comes in two forms: generative tools that compose a fresh letter from a prompt, and template-based assembly that slots your pre-written clauses into a fixed skeleton from a questionnaire. Both cut drafting time; assembly also guarantees consistent, reviewed wording every time — which matters most when an LOI goes through several rounds of counters.
Should a small CRE firm buy an off-the-shelf tool or build a custom template engine?
It depends on volume, language reuse, and how many people draft. Buy off-the-shelf when LOIs are occasional, terms vary widely, and one person reviews every letter. Build a custom engine when LOIs are high-volume, most language repeats, and several people draft, so consistency has become a source of error. Firm size alone does not decide it — a busy two-person shop can justify a build a sleepy ten-person one cannot.
Is an automated LOI legally binding?
An LOI is generally non-binding on the core deal terms, but specific provisions — confidentiality, exclusivity, sometimes a break-up fee — are usually intended to bind. Automation reproduces whatever language you encoded; it does not judge enforceability. That is exactly why a person reviews the binding provisions on every letter: only counsel or a principal can confirm the enforceable clauses say what the deal requires.
Is it safe to draft an LOI in ChatGPT?
Not on a free or personal account, where inputs can be used for model training by default — and an LOI carries price and concession terms you do not want in a training set. Use a business-tier account (ChatGPT Business, Claude Team, or the API), which does not train on your data by default, or a document-automation platform with explicit data terms. A custom engine avoids the question by keeping deal terms in your own environment.
How much does LOI or document automation cost?
Consumer AI generators run free to about $25 per user a month. Legal document-automation platforms tier from roughly $99 a month for a few templates up past $350 a month for larger libraries and multiple seats. CRE deal-management platforms that include LOI drafting run into the low-to-mid four figures a year and up, because you are buying the whole platform. A custom template engine is a project-based build, with small custom-automation engagements in a market range of roughly $25–150K depending on scope.
Does Buildout or Dealpath draft LOIs?
Buildout’s platform includes AI chat that drafts LOIs and PSAs and syncs to live deal data. Dealpath focuses on the buy-side acquisitions pipeline rather than sell-side document generation. If you already run one of these platforms, its drafting feature may be all you need. If you do not, buying the platform solely for LOI generation is rarely worth it — the feature is a small part of a large contract.
What is a custom template engine for LOIs?
It is your firm’s LOI language codified into a system that assembles a finished letter from a short set of inputs. It has three parts: a clause library of your counsel-approved provisions, conditional logic that picks the right clauses per deal type, and an input form a broker fills in. It produces the same reviewed wording every time, cuts drafting to minutes, and — unlike a subscription — the clause library is an asset your firm owns.
Does an automated LOI still need a lawyer?
Yes, for the provisions meant to bind. Automation reliably reproduces language and updates figures, but it cannot judge whether a confidentiality carve-out or exclusivity window fits the deal or exposes the firm. The pattern is the same regardless of tool: the system drafts, and a principal or counsel reviews the binding clauses before the letter goes out.
How much time does LOI automation actually save?
The biggest savings come across negotiation rounds, not the first draft. Automated LOI workflows can compress negotiation cycles by roughly half, largely because consistent, deterministic redrafting removes the proofreading manual counters create. A first draft might drop from an hour to a few minutes; the compounding value is that every later round changes only the inputs that changed, leaving the rest of the letter untouched and correct.
When does a firm outgrow an off-the-shelf LOI generator?
When three conditions arrive together: LOI volume becomes a recurring, high-volume workflow; most clause language repeats and has been approved by counsel; and several people draft, so inconsistency has become a real cost. Below that threshold, a subscription tool is enough. Above it, the manual rekeying and version drift cost more than a custom engine would, and ownership of the clause library starts to matter.
Key takeaways
- Off-the-shelf AI generators compose a fresh LOI from a prompt; a custom template engine assembles your counsel-approved clauses deterministically — the second guarantees consistency the first cannot.
- Score the decision on five variables: volume, language reuse, control over wording, consistency across drafters, and integration. Three or more pointing to “custom” and a build pays off.
- The real value of automation is consistent redrafting across negotiation rounds, not first-draft speed — measure the fourth draft, not the first.
- Never draft a live LOI in a consumer AI account, where deal terms can feed model training; use business-tier accounts or keep terms in an owned engine.
- Automation drafts; a human always owns the binding provisions — confidentiality, exclusivity, and any enforceable term get a final read every time.
Not sure which side of the threshold your firm sits on? A short, free AI-readiness assessment will map your LOI volume, clause reuse, and data-handling needs and tell you exactly which approach earns its cost. Book your free AI-readiness assessment → and we will size it for your firm.
Arthur Wandzel
