For a small commercial real estate firm buying a custom AI automation build, a fixed price on a tightly scoped single workflow is the safer default — but only after a paid discovery step makes that price honest. Time and materials fits genuinely open-ended work, yet it quietly assumes you have someone to supervise the meter, and a 4–20 person firm with no IT department does not. The right structure is rarely a pure version of either. It is almost always phased: a small fixed-price discovery to pin down the scope, then a fixed-price build against the resulting spec — or a capped, milestone-gated engagement when the work is real research. This is a decision about who carries the estimating risk, and it turns on how well the work can be defined before anyone commits.
Most advice on this question is written for companies with a technical buyer and an engineering team to watch the work. That reader can run a time-and-materials engagement well, because someone on staff can judge whether the hours are real. A commercial real estate firm has no such person, which changes the math entirely. Before you decide how to build, the buy-versus-build playbook covers whether a custom build is even the right spend; this piece assumes you have cleared that bar and are now choosing how to contract for it.
The Short Answer for a Firm With No IT Department
Pick the structure that protects a buyer who cannot watch the work.
A fixed price puts the estimating risk on the builder. If the job takes twice as long as they thought, that is their problem, not your invoice. For a firm with no one on staff to supervise a developer, that transfer of risk is the whole point: you agree on a deliverable and a number, and you are insulated from how many hours it takes. The catch is that a builder will only quote a fixed price they can stand behind, which means the work has to be defined well enough to estimate — and a lot of small-firm automation is not, on day one.
Time and materials puts the estimating risk on you. You pay for hours worked at an agreed rate, so if the work runs long, your bill grows. That is fair and transparent when both sides can see the work clearly and the buyer can tell a productive week from a slow one. It becomes a trap when the buyer cannot, because the only control on a T&M engagement is an informed client saying “why did that take forty hours.” Remove the informed client and you have removed the brake.
So the default for a 4–20 person firm leans fixed price — with one condition attached that most proposals skip, covered below.
What Each Structure Actually Means
The two models price the same work on opposite principles.
Fixed price is a single agreed number for a defined deliverable. The proposal names the workflow, the systems it touches, the acceptance test, and the price. Change the scope and you renegotiate — that is the change order, the mechanism buyers dread and builders rely on. A fixed price is a promise about an outcome.
Time and materials is an agreed rate — hourly or daily — against actual hours worked, usually billed on a regular cycle. There is no promised total; there is a promised rate and a running tally. Scope can flex week to week without a formal renegotiation, which is the model’s real advantage. A T&M contract is a promise about a rate.
A third arrangement sits between them and matters more than either in its pure form: capped time and materials, sometimes written as a “not-to-exceed.” You bill hourly, but the contract names a ceiling the total cannot cross without your written approval. It keeps T&M’s flexibility while restoring the budget certainty a small firm needs.
Where the Risk Sits in Each Model
Every contract structure is a decision about who absorbs the gap between the estimate and reality. Naming where that risk lands is the fastest way to read a proposal.
| Dimension | Fixed price | Time and materials |
|---|---|---|
| Who carries estimating risk | The builder | You |
| Budget certainty | High — you know the number | Low — you know the rate, not the total |
| Scope flexibility | Low — changes need a change order | High — scope flexes week to week |
| Oversight required from you | Low — you judge the outcome | High — you judge the hours |
| Best when | Scope is well defined | Scope is genuinely unknown |
| Failure mode | Padded bid, or change-order friction | Runaway meter with no informed brake |
The row that decides it for a small CRE firm is oversight. Fixed price asks you to judge one thing at the end: did the automation do what the acceptance test said. Time and materials asks you to judge the work continuously, every invoice cycle, with enough technical fluency to know whether the hours were earned. If you have that fluency in-house, T&M can be the more economical choice because you are not paying a risk premium baked into a fixed number. If you do not — and most firms this size do not — the fixed price buys you out of a supervision job you cannot staff.
Why CRE Data Breaks a Naive Fixed Price
The reason you cannot simply demand a fixed price on day one is the same reason CRE automation projects overrun: the condition of your own data.
An automation that reads leases is only as predictable as your leases are consistent. If fifteen years of documents live as a mix of scanned images, emailed attachments, and re-saved copies across a shared drive three assistants organized three different ways, no honest builder can quote a firm fixed price to process them sight unseen. They either pad the number heavily to cover the unknown, or quote low and recover the difference through change orders when the mess surfaces in week three.
This is why the condition of your data is the largest cost line most quotes underprice — a dynamic explored in the guide to what a custom automation actually costs. The contract lesson is narrower: unknown scope and a fixed price do not belong in the same document. When you see them together, someone is absorbing risk they cannot see, and it is usually priced back to you.
The Phased Structure That Beats Both
The structure that protects a small firm is not fixed price or T&M. It is a sequence that converts unknown scope into known scope before the big number is committed.
Phase one — a paid discovery, fixed price. A short, bounded engagement at a small fixed fee, in which the builder examines your actual documents, maps the workflow, and inspects the systems the automation must touch. The deliverable is not code; it is a specification and a firm build quote. You are buying a de-risked estimate — and if the number that comes back is wrong for your firm, you walk having spent a few thousand dollars instead of committing to a six-figure guess.
Phase two — the build, fixed price against the spec. With discovery having replaced unknowns with a documented scope, a fixed price is honest, and you get the risk transfer you wanted from the start. The builder can stand behind the number because they have seen the data; you get budget certainty because the scope is written down.
When the work is genuine research — an accuracy target on messy documents that no one can promise up front — use capped time and materials with milestone acceptance instead of a fixed build. You pay for hours, but a not-to-exceed ceiling caps your exposure and each milestone carries an acceptance test you sign off before the next tranche begins.
Either way, the principle holds: never commit the large number until discovery has made it real. Sequencing the spend this way is the same logic that lets lean firms out-operate larger ones, described across the small CRE firm AI approach — prove the ground before you build on it.
A Decision Matrix by Project Shape
The right structure follows the shape of the work more than any pricing preference.
| Project shape | Recommended structure | Why |
|---|---|---|
| Single, well-understood workflow (e.g. lease term extraction into a spreadsheet) | Fixed price, after a short discovery | Scope is knowable; you want the risk transfer and the budget certainty |
| Multi-system build with defined logic | Fixed-price discovery, then fixed-price build | Discovery converts the integration unknowns into a spec you can price |
| Accuracy research on messy documents | Capped T&M with milestone acceptance | No one can honestly promise the target up front; the cap and checkpoints control exposure |
| Small, exploratory prototype to test an idea | T&M, tightly time-boxed | The point is to learn cheaply; a fixed price on a two-week probe is overhead |
| Ongoing maintenance and model updates | Monthly retainer | Recurring, unpredictable, low-volume work fits a flat monthly fee, not a project price |
Two firms asking for “lease automation” can correctly land in different rows because one has clean data and a single workflow while the other has a shared-drive mess and three systems that must talk. The matrix is a starting point; the discovery step is what confirms which row you are actually in.
Red Flags in a Fixed-Price Bid
A fixed price is only as good as the definition behind it. Watch for these:
- No discovery, no breakdown, one flat number. A firm price on unseen data is either padded or a change-order trap. A real fixed price cites the scope discovery it rests on.
- No acceptance test. If the contract does not define what “done” means in measurable terms — reads X documents at Y accuracy, writes to Z — you have no way to hold the builder to the price you paid.
- Change orders priced to punish. Some builders quote a low headline number and load the change-order rate, betting on scope creep. Ask how changes are priced before you sign, not after.
- The number seems low for the scope. A fixed price well under the market range for the work described is not a bargain; it is a bid that plans to recover margin through changes. The cost guide gives the ranges to sanity-check against.
Red Flags in a Time-and-Materials Arrangement
T&M’s transparency is real, but only if the arrangement is built to be watched. Warn signs:
- No cap. Open-ended T&M with no not-to-exceed hands a blank check to a firm you cannot supervise. Insist on a ceiling.
- No milestones or acceptance gates. Hours with no checkpoints mean you find out the work went wrong at the end, having paid for all of it. Gate the money to accepted milestones.
- Vague or absent reporting. You should get a regular, legible account of what was done and how long it took. “Trust us, it’s progressing” is not a report.
- A rate with no seniority mix. Know who is doing the work and at what rate. Junior hours billed at senior rates is the quiet way a meter runs hot.
Frequently Asked Questions
Is fixed price or time and materials better for a CRE automation project?
For most small commercial real estate firms, a fixed price on a tightly scoped workflow is the safer default, because it transfers the estimating risk to the builder and requires no ongoing supervision — which a firm with no IT department cannot provide. Time and materials is more economical only when you have the in-house fluency to judge whether hours are justified. The important nuance is that a fixed price is only honest after a discovery step defines the scope; demanding a firm number on unseen data produces either a padded bid or a change-order fight.
What is time and materials in a software contract?
Time and materials means you pay an agreed hourly or daily rate for the actual hours worked, billed on a regular cycle, with no promised total. It gives the builder room to adjust scope week to week without a formal renegotiation, which suits genuinely uncertain work. The trade-off is that your budget is open-ended — you know the rate, not the final number — and the only real control is a client who can tell a productive week from a slow one. Small firms usually pair it with a not-to-exceed cap to restore budget certainty.
Why would a builder refuse to quote a fixed price?
Usually because the scope is not defined well enough to estimate honestly, and the most common culprit is the condition of your data. If your leases or deal files are an inconsistent mix of scans and re-saved copies, a builder cannot know how many hours the cleanup will take, so a responsible one proposes a paid discovery first rather than guess. A refusal to quote fixed on unseen data is often a good sign — the builder will not pad a number or set up a change-order recovery later.
What is a not-to-exceed cap and why does it matter?
A not-to-exceed cap is a ceiling written into a time-and-materials contract stating the total cannot cross a set figure without your written approval. It keeps the flexibility of hourly billing while restoring the budget certainty a small firm needs, so you get the best of both structures. It matters most for a buyer who cannot supervise the work continuously, because the cap enforces a limit that an informed client would otherwise have to enforce by watching every invoice.
How do I stop scope creep from blowing up a fixed-price build?
Define the scope precisely before signing, insist on a written acceptance test, and agree in advance how change orders are priced. Scope creep hurts when the original scope was vague, because everything becomes a judgment call about whether it was “included.” A build that starts from a discovery-produced specification has far less room for that argument. Treat any new request as a deliberate change with its own small quote rather than an assumed extra.
Should I use fixed price for the discovery phase too?
Yes. A discovery engagement is short and bounded, which makes it easy to quote as a small fixed fee — and that fee buys you a specification and a firm build quote you can trust. Paying T&M for discovery reintroduces the open-ended risk the phased approach is meant to remove. The clean pattern is fixed-price discovery, then fixed-price build against the spec discovery produced, with the discovery fee typically a small fraction of the build.
Does my confidential deal data affect the contract structure?
It affects the scope, and therefore the price and the discovery. A firm that requires the automation to run in infrastructure it controls, with tighter access controls and a data-handling review, is describing more work — and that work needs to be named in discovery so it lands in the fixed price rather than surfacing as a change order. Raise your confidentiality and hosting requirements at the discovery stage, in writing, so the structure and the number account for them from the start.
What if the whole build is genuine research with no guaranteed outcome?
Use capped time and materials with milestone acceptance. When the deliverable is an accuracy target on messy documents that no honest builder can promise up front, a fixed price forces them to either pad heavily or overpromise. A not-to-exceed cap limits your exposure, and signing off each milestone before the next tranche begins gives you checkpoints to stop or adjust. That structure gives a builder room to iterate on a hard problem while keeping you from funding an open-ended experiment.
How do I choose a partner who structures the deal fairly?
Look for a builder who proposes discovery before a fixed price, defines acceptance tests in writing, and is transparent about who does the work and how changes are priced. A fair partner structures the engagement to protect a buyer who cannot supervise it, rather than to maximize billable hours or bury recovery in change orders. The guide to the best AI development partners for small CRE firms covers what separates a partner from a vendor in more detail.
Where to Start
The contract structure matters less than the decision upstream of it: which workflow is worth building at all, and whether its scope is defined tightly enough to price. Most firms argue about fixed price versus time and materials before they have done the work that makes either one honest.
A free AI-readiness assessment produces exactly that groundwork — a working session that maps your firm’s workflows, flags the one or two where a custom build pays for itself, and returns a ranked plan with real cost ranges and a scope defined well enough to quote. With that in hand, the contract structure often chooses itself. The economics of building versus renewing subscriptions are worked through in the companion piece on proptech subscriptions versus a custom automation project.
Book a free AI-readiness assessment if you want the scope pinned down before you sign a proposal of either shape. If a build is not the right spend for your firm, the assessment will say so, and you will still leave with the plan.
Arthur Wandzel
