Market Structure Analysis · North America
Every service a property developer buys, who owns it, who signs for it, what it costs — and, crucially, which pocket the money comes out of at that exact moment.
If you read nothing else, read this page. Everything after it is the evidence.
Everything spent before financing close is the sponsor’s own money, at risk, unrecoverable if the deal dies — and roughly 60–70% of pursued deals do die. Everything after close is committed capital drawn against a lender-controlled schedule. Same company, same project, opposite buying behaviour: days-to-yes versus quarters-to-yes, $20k cheques versus $750k cheques, card payment versus lien-waiver-gated draws.
The stages where a developer is most desperate — underwriting in 72 hours, due diligence against a 45-day clock, an investor asking for a revised deck by Friday — carry the smallest and most fragile budgets in the entire lifecycle (Stages 02–05, roughly $200k combined). The stages with real money (Stages 09 and 12, $750k and $23M+) are slow, procurement-gated and price-competitive. Any AI-enabled services business has to choose which side of that trade it plays — or deliberately use the cheap urgent work as a wedge into the expensive work on the same deal.
Through Stage 07 the buyer is the Principal, spending personal capital on conviction, and the product is certainty. From Stage 08 the buyer is usually the architect or the general contractor, spending project capital on throughput, and the product is capacity. At Stage 13 it is a marketing lead spending a pre-approved reserve on lease-up velocity. Three buyers, three pains, three sales motions — routinely collapsed into the single word “developer.”
Soft costs a developer pays out of pocket before close are rolled into the construction budget and credited back as sponsor equity. A vendor who defers, contingently prices, or success-fees Stage 02–05 work is not being generous; they are removing the only genuinely unrecoverable-cash objection in the lifecycle and buying an incumbent position that carries into Stages 06–13 on the same deal, where the money is 20× larger.
Surveying, geotechnical and structural engineering are protected by professional stamps and physical fieldwork. Renderings, pro formas, marketing collateral and investor materials have no licensure moat at all and are near-pure production. But the interesting quadrant is CAD/BIM production and construction documentation: 70–95% repeatable, with liability that attaches to the signing architect, not the producer. That is the largest pool of automatable spend in the industry and it is legally outsourceable today — which is precisely why it is already the most-offshored function in AEC.
All figures in this document are anchored to one worked transaction so the numbers stay comparable. Scale roughly linearly to $10M; sub-linearly above $100M, where fee percentages compress.
A $40M ground-up multifamily project, about 150 units, in a US secondary market, developed by a sponsor with a 3–10 person team. This is the modal North American development company: a $500M pipeline is often run by fewer than 25 people, because everything except capital, coordination and decision-making is bought from outside firms.
Total development cost decomposes as follows. The band that matters for a services business is soft costs — 13% here, and 15–30% across commercial development generally, rising to 25–35% on institutionally complex assets.
Figure 1
Total development cost composition. Hard construction dominates, but it is bought once from one counterparty; soft costs are bought repeatedly from twenty.
This is the addressable pool. Architecture and engineering is the single largest line item in soft costs on virtually every project — typically 6–9% of construction cost for commercial work, and 8–12% where design complexity, sustainability requirements or code intricacy are high.
Figure 2
Architecture and engineering is 40% of soft costs on its own — and is itself roughly half production labour.
The single most useful line to draw across the development lifecycle is not a phase boundary. It is financing close — the moment the money stops being personal.
Regime A · Stages 01–07
Regime B · Stages 08–15
The mechanic almost every vendor misses
Pre-close spend is reimbursable at closing. Soft costs the sponsor has already paid get rolled into the construction budget and credited toward their equity contribution — some sponsors deliberately front-load soft cost spend to demonstrate a larger equity position and improve their loan-to-cost. So a developer refusing a $30k engagement at Stage 02 is not short of $30k. They are refusing because it is the only truly unrecoverable money in the entire project. Price against that, not against the number.
Who is actually behind the money, what each layer costs, and what each layer demands before it funds.
2026 construction lenders are underwriting materially tighter than pre-2022 — 50–65% loan-to-cost rather than 75%. That gap has to be filled, and filling it is what generates most of the document production in Stages 05 and 11. Every additional layer in the stack is another counterparty with its own reporting format, its own model, and its own diligence list.
Figure 3
Ordered by repayment priority. Risk, cost of capital and documentation burden all rise as you move up.
Two payment chains. If you sell into this industry, your cash conversion cycle is decided entirely by which one you are in.
| Step | Regime A — before financing close | Regime B — after financing close |
|---|---|---|
| 1. Trigger | Principal decides a deal is worth chasing | Vendor completes work against an approved budget line |
| 2. Authorisation | Verbal or a one-page proposal. Principal signs. | Invoice enters the monthly pay application; must map to a line in the lender-approved budget |
| 3. Verification | None | Owner’s representative reviews percentage complete |
| 4. Third-party check | None | Lender’s inspector or fund-control agent verifies progress on site |
| 5. Title / legal | None | Title company runs a date-down endorsement; conditional and unconditional lien waivers collected from every tier |
| 6. Release | Wire or card, same week | Draw funds released to the borrower, then disbursed down the chain |
| 7. Cash in hand | 3–30 days | 45–90 days |
| 8. Retention | None | 5–10% held back until substantial completion on construction-tier work |
| Implication: in Regime B every vendor is, in effect, extending 60–90 days of unsecured credit to the developer’s schedule. Small AI-native firms routinely underestimate this and price as though they were in Regime A. | ||
Who pays whom
An important structural detail for anyone selling production services: after Stage 07 the developer usually is not your customer. The developer contracts the architect on a phased fee; the architect then buys CAD, BIM, drafting and visualisation out of that fee. Your invoice goes to the architecture firm, whose margin you are directly compressing or expanding. Same for the GC and shop drawings, takeoffs and coordination.
That flips the value proposition from “we make your project better” to “we make your fee more profitable.” It is a much easier sale, at a much lower price point, to a much more sophisticated buyer.
Each sheet records the process owner, the person who signs, the objective, the services bought with 2025–26 costs, the source of funds, and the dominant friction.
Find a site worth underwriting before a competitor ties it up.
| Service bought | Provider | Cost |
|---|---|---|
| Data subscriptions (CoStar, CREXi, GIS) | Data vendors | $5k–40k / yr |
| Broker opinion of value | Brokerage | Free |
| Market & demographic study | CBRE, JLL, boutiques | $5k–30k |
| Zoning / parcel screening | Consultant or in-house | $2k–10k |
Source of funds
Firm overhead, paid from developer fees and promote earned on prior projects. Frequently a corporate card. This is a subscription budget, not a project budget — renewed annually, benchmarked against competitors, and defended by whoever owns acquisitions.
Dominant friction
Signal-to-noise. Sourcing teams screen hundreds of parcels to underwrite ten and close one.
Kill the deal cheaply — or justify spending real money on it.
| Service bought | Provider | Cost |
|---|---|---|
| Financial pro forma / underwriting model | Analyst or consultant | $3k–15k |
| Yield test / test-fit massing | Architect | $2k–10k |
| Highest & best use study | Advisory firm | $5k–20k |
| Preliminary cost per SF | General contractor | Free (bid-chasing) |
Source of funds
100% sponsor cash. Reimbursed at close only if the deal survives — and most do not. This is the most price-sensitive money in the entire lifecycle, and simultaneously the most urgent.
Dominant friction
They need a number they would defend to a lender, in 72 hours, on a deal they will probably abandon. Speed and credibility, at a price that survives a 70% write-off rate.
Discover deal-killers before earnest money goes hard.
| Service bought | Provider | Cost |
|---|---|---|
| ALTA / NSPS land title survey | Licensed surveyor | $3k–8k |
| Phase I environmental site assessment | Environmental consultant | $2.2k–4.5k |
| Phase II ESA, if triggered | Environmental consultant | $15k–60k |
| Geotechnical investigation | Geotech engineer | $3k–20k |
| Utility / will-serve study | Civil engineer | $3k–15k |
| Title review & legal | Real estate counsel | $10k–50k |
Source of funds
Sponsor cash. Reimbursable at close. Phase I turnaround is typically 2–3 weeks; hillside or contaminated sites push geotech past $20k and can exceed $50k on complex commercial ground.
Dominant friction
Sequencing six or more independent vendors inside a 30–60 day DD window. One late report means an extension fee, a renegotiated price, or a forfeited deposit. Nobody owns the critical path across the consultants.
Tie up the site with the minimum capital exposure and the maximum optionality.
| Service bought | Provider | Cost |
|---|---|---|
| PSA drafting & negotiation | Real estate counsel | $15k–75k |
| Entity formation & JV documents | Corporate counsel | $10k–40k |
| Appraisal | MAI appraiser | $3k–10k |
| Earnest money deposit | Escrow | 1–3% of price |
Source of funds
Sponsor cash plus escrow. On a $6M land basis the deposit is $60k–180k, and it typically goes hard at the end of DD — the point at which the deal becomes financially irreversible and the developer’s risk appetite changes character completely.
Dominant friction
Low. This stage is a legal moat: relationship-driven, liability-bearing, and not meaningfully compressible by a third party.
Close the equity gap and secure a term sheet before land control lapses.
| Service bought | Provider | Cost |
|---|---|---|
| Investment memorandum | Consultant / in-house | $3k–15k |
| Pitch deck & investor collateral | Design studio | $2k–10k |
| Lender-format model rebuilds | Analyst | $5k–20k |
| Hero renderings (2–4 images) | Visualisation studio | $2k–12k |
| Drone & site photography | Local operator | $1k–5k |
| Project website & branding | Agency | $5k–20k |
| Mortgage brokerage | Debt broker | 0.5–1% of loan |
Source of funds
Sponsor cash out of pocket; the broker fee alone is $120k–240k on a $24M loan but is paid from the loan at closing, not upfront. Everything else is reimbursed at close.
Dominant friction
Every lender and every LP wants a different format, and wants it by Friday. Revision velocity beats production polish here by a wide margin — and the developer’s vendors are almost never structured to deliver same-week turns.
Produce a scheme that satisfies the pro forma, the zoning code and the investor deck simultaneously.
| Service bought | Provider | Cost |
|---|---|---|
| Concept / schematic architecture | Architecture firm | $250k–450k* |
| Standalone concept-only engagement | Boutique / freelance | $25k–150k |
| Exterior rendering, per image | Visualisation studio | $800–4,000 |
| Interior rendering, per image | Visualisation studio | $600–2,500 |
| Aerial / context view | Visualisation studio | ~$800 avg |
| High-end photoreal, per image | Premium studio | $2,000–8,000 |
| Massing / BIM concept model | BIM firm | $2k–20k |
Source of funds
*Schematic design is roughly 15–20% of the total A/E fee. With architecture at 6–9% of a $23.4M construction cost, the total A/E fee is $1.6M–2.1M, so concept alone runs $250k–450k. Architects invoice monthly against a phased fee schedule. Note that under AIA B101-2017 the fee is tied to the owner’s budget at contract signing, not final construction cost.
Dominant friction
The scheme has to serve three incompatible audiences. Every pro forma revision triggers a design revision, which triggers a rendering revision, which triggers an investor-deck revision — with three vendors and no shared source of truth.
Convert raw land into approved density. This is where the value is actually created.
| Service bought | Provider | Cost |
|---|---|---|
| Land use / planning consulting | Planning consultancy | $15k–100k |
| Civil engineering | Civil firm | $30k–250k |
| Traffic impact study | Traffic engineer | $10k–75k |
| Environmental review (CEQA / NEPA / state) | Environmental consultant | $15k–250k |
| Landscape architecture | Landscape firm | $10k–100k |
| Land use counsel | Specialist attorney | $25k–200k |
| Hearing visualisation & public engagement | Studio / consultant | $5k–50k |
| Application & municipal fees | Jurisdiction | $20k–400k+ |
Source of funds
Sponsor cash or predevelopment equity, burned over 8–12 months for a typical multifamily entitlement, while land carry costs run in parallel. Jurisdiction dominates the number: Los Angeles entitlement processes commonly run $400k–1.1M all-in, and New York’s ULURP costs $150k–400k in application fees alone before $200k–500k of legal and consulting.
Dominant friction
Unbounded revision cycles driven by parties the developer cannot control. Every agency comment regenerates drawings, studies and visualisations. This is the largest, longest, least predictable at-risk spend in the lifecycle — and the one most likely to kill a deal after the money is committed.
Financing close — the regime boundary
Above this line: personal money, fast decisions, small cheques, total loss if the deal dies. Below it: committed capital, budget lines, lender control, 60–90 day payment. All Stage 01–07 sponsor spend is reimbursed here and credited toward the sponsor’s equity contribution.
Turn an approved scheme into a coordinated, buildable, priceable design.
| Service bought | Provider | Cost |
|---|---|---|
| Design development phase (A/E) | Architecture firm | $350k–550k* |
| Revit / BIM modelling, offshore | BIM outsourcing firm | $15–40 / hr |
| Senior BIM modeller, offshore | India / Vietnam / Philippines | $20–35 / hr |
| CAD drafting, offshore vs North America | Drafting firm | $8–15 vs $80–150 / hr |
| BIM coordination | Specialist | $40–100 / hr |
| Structural / MEP engineering | Engineering consultancies | Within A/E fee |
| Interior design & FF&E | Interior firm | $60k–250k |
Source of funds
*Design development is roughly 20–25% of the A/E fee. Critically, the architect pays the outsourced production firms out of that fee, not the developer. Under a LIFO draw structure the sponsor’s equity is fully deployed before the lender releases a dollar, so early-stage vendors are paid from equity.
Dominant friction
Capacity elasticity. Firms are structurally understaffed for peaks and will not hire permanently for a surge. Offshore rates look attractive until you price in revision cycles, time-zone lag, code literacy gaps and rework — which is why BIM is bought by the deliverable, not the hour.
Produce the permit set and the build set. The largest pure-production spend in the industry.
| Service bought | Provider | Cost |
|---|---|---|
| CD phase (A/E) | Architecture firm | $600k–900k* |
| Sheet production & detailing | Outsourced production | Per sheet / retainer |
| Revit family creation & standards | BIM firm | $15–40 / hr |
| Discipline coordination & clash | BIM coordinator | $40–100 / hr |
| Code compliance review | Consultant | $15k–60k |
| Permit expediting | Expediter | $10k–80k |
Source of funds
*CD is 35–40% of the A/E fee — the single largest phase. At the firm level, annual outsourcing budgets run $50k–200k for boutiques, $250k–1M for mid-size firms, and $2M–20M+ for large practices. Minimum project charges of $150–250 mean small one-off requests never get outsourced.
Dominant friction
This is where the labour actually is. Documentation is roughly half of what an architecture practice does, it is the least differentiated half, and it is the half clients are least willing to pay a premium for. Every firm knows this; almost none have solved it.
Convert drawings into a guaranteed maximum price the lender will underwrite.
| Service bought | Provider | Cost |
|---|---|---|
| Quantity takeoff & estimating | QS / estimating firm | $20k–150k |
| Owner’s representative | PM consultancy | 0.5–1.5% of hard cost |
| Bid package preparation | GC preconstruction | Within GC fee |
| Value engineering exercise | GC + architect | $15k–60k |
| GC fee (overhead & profit) | General contractor | 3–4% of hard cost |
Source of funds
Project budget. Owner’s rep at 0.5–1.5% of a $23.4M hard cost is $120k–350k. The GC fee of 3–4% is $700k–940k and sits inside hard costs, separate from and additional to the developer fee.
Dominant friction
Value engineering is a euphemism for redesign under time pressure. A failed GMP sends the project back to Stage 08 with the loan commitment ticking.
Fund the loan, fund the equity, and reimburse everything spent to date.
| Service bought | Provider | Cost |
|---|---|---|
| MAI appraisal | Appraisal firm | $8k–25k |
| Plan & cost review | Lender’s consultant | $5k–15k |
| Updated pro forma & lender package | In-house / consultant | $10k–40k |
| Loan origination fee | Lender | 0.75–1.5% of loan |
| Closing legal | Counsel, both sides | $50k–150k |
| Title insurance & escrow | Title company | 0.3–0.7% of loan |
Source of funds
Everything here is netted out of loan proceeds at closing — the developer never writes a cheque. Origination on a $24M loan is $180k–360k. This is the moment all Stage 01–07 sponsor spend comes back, and the moment the GP’s $2.0M equity is formally committed.
Dominant friction
Low friction, high stakes. The work is procedural but the calendar is not: a missed close means re-trading the term sheet in a moved market.
Build on programme, on budget and lien-free, against a monthly draw cycle.
| Service bought | Provider | Cost |
|---|---|---|
| Hard construction | GC & trades | $23.4M |
| Construction administration (A/E) | Architect | $320k–460k* |
| Shop drawing production & review | Trades / outsourced | Within trade cost |
| BIM clash detection & coordination | VDC consultant | $40k–150k |
| Reality capture, drone progress, laser scan | Monitoring firm | $15k–80k |
| Third-party inspection & testing | Testing lab | $60k–180k |
Source of funds
*Construction administration is roughly 20% of the A/E fee, and CA scope has expanded 20–30% since 2020 on coordination complexity alone. Payment flows through the full draw chain — pay app, owner’s rep review, lender inspection, title date-down, lien waivers — taking 45–90 days, with 5–10% retention held to substantial completion.
Dominant friction
Coordination volume. RFIs, submittals, change orders and schedule updates are high-frequency, low-judgment, document-heavy work — and the delay cost of doing it slowly is measured in days of interest carry on $24M.
Hit the absorption assumptions in the pro forma before the construction loan converts.
| Service bought | Provider | Cost |
|---|---|---|
| Marketing CGI package | Visualisation studio | $10k–100k |
| Animation / film | Studio | $10k–100k |
| VR walkthrough / interactive sales centre | Interactive studio | $15k–100k |
| Website & leasing funnel | Agency | $5k–50k |
| Branding, print, signage, hoarding | Agency + fabricator | $15k–70k |
| Physical scale model | Model maker | $5k–100k+ |
| Campaign media spend | Media buyer | $50k–250k |
Source of funds
This is the one large creative-services budget that is already approved and sitting in the capital stack — typically 2–5% of total development cost, established at financing close and released monthly. It is defensible, predictable, and rarely cut, because slow lease-up threatens the refinance.
Dominant friction
Timing. Marketing assets are needed 12–18 months before there is a building to photograph, so everything is CGI — and every design change during construction invalidates the imagery already produced.
Stabilise net operating income to the level the exit assumption requires.
| Service bought | Provider | Cost |
|---|---|---|
| Property management | PM company | 3–5% of gross revenue |
| Property management platform | Yardi, MRI, RealPage | Per door / month |
| Leasing & renewal marketing | Agency / in-house | Ongoing |
| Maintenance & building systems | Vendors | Ongoing |
| Investor reporting | In-house / fund admin | Quarterly |
Source of funds
Operating revenue. Unlike every stage above it, this budget renews forever and does not depend on a new deal closing — which makes it the only genuinely recurring revenue pool in property development, and the reason software companies have concentrated here.
Dominant friction
Incumbency. Yardi, MRI and RealPage own the system of record and the switching cost is very high.
Repay debt, return LP capital and preferred return, crystallise the promote, seed the next deal.
| Service bought | Provider | Cost |
|---|---|---|
| Investment sales brokerage | Capital markets team | 1–2% of price |
| Sale legal & closing | Counsel | $50k–200k |
| Refinance package | Debt broker | 0.5–1% of loan |
| Disposition marketing (OM, imagery) | Brokerage / studio | $20k–80k |
Source of funds
Sale or refinance proceeds. Waterfall order: senior debt, mezzanine, LP capital, LP preferred return of 6–10%, then the sponsor promote. The promote becomes Stage 01–07 pursuit capital for the next project — which is exactly why pursuit budgets stay tight even at highly profitable firms. It is the most expensive money in the company.
Dominant friction
Cash is recycled, not banked. A developer with a $10M profit typically distributes a fraction, reserves some, pays tax, and re-deploys the rest as GP equity in the next deal.
The chart that explains developer buying behaviour better than any persona document.
Figure 4
Pursuit cash is unrecoverable and peaks just before close. Committed GP equity is larger but returns at exit. The two are psychologically opposite, and vendors are priced against the wrong one.
Professional and creative services only, by stage, colour-coded by money regime. Hard construction is excluded and noted separately.
Figure 5
Two-thirds of purchasable service spend sits after financing close — but two-thirds of the decisions sit before it.
Every purchasable service in the lifecycle, assessed on the two variables that determine whether software can take it: how much of the work is repeatable production, and what protects it.
| Service | Stage | Who buys it | Who delivers it | Typical engagement | Repeatable production |
Liability barrier |
Verdict |
|---|---|---|---|---|---|---|---|
| Financial modelling & pro formas | 02, 05, 11 | Developer | Analysts, consultants | $3k–25k | 86% | 1 / 10 | Standardised logic, project-specific inputs. Nothing protects it. |
| Market research & feasibility | 01, 02 | Developer | CBRE, JLL, boutiques | $5k–30k | 82% | 1 / 10 | Inputs are already public and digital. Data moat, not labour moat. |
| Investor & capital-raise collateral | 05 | Developer | Design studios, in-house | $5k–35k | 90% | 2 / 10 | Pure document production under extreme time pressure. |
| Architectural visualisation (CGI) | 05, 06, 07, 13 | Developer & architect | Visualisation studios | $600–8,000 / image | 92% | 1 / 10 | Already collapsing on price. Differentiation moving to art direction. |
| Animation, VR & interactive | 13 | Developer | Studios | $10k–100k | 85% | 1 / 10 | Same trajectory as stills, 18–24 months behind. |
| Marketing collateral & campaigns | 13 | Developer | Agencies | $20k–390k | 90% | 1 / 10 | Pre-approved reserve. Repeatable across every project. |
| CAD production drafting | 08, 09 | Architect | Offshore drafting firms | $8–150 / hr | 94% | 3 / 10 | Already offshored. The next compression is software, not geography. |
| BIM / Revit modelling | 08, 09 | Architect | BIM outsourcing firms | $15–40 / hr | 82% | 4 / 10 | Billed by deliverable, not hour. Rework rate is the real cost driver. |
| Construction documentation | 09 | Architect | Architecture firms | $600k–900k | 72% | 5 / 10 | Largest automatable pool in AEC. Liability attaches to the signing architect, not the producer. |
| Quantity takeoff & estimating | 10 | Developer & GC | QS firms | $20k–150k | 78% | 2 / 10 | Measurement automates; pricing strategy and scope interpretation do not. |
| Project coordination, RFIs, reporting | 12 | GC | PM consultancies | 0.5–1.5% hard cost | 70% | 3 / 10 | High-frequency, low-judgment, document-heavy. Prime agent territory. |
| Code & compliance review | 09 | Architect | Consultants | $15k–60k | 64% | 6 / 10 | Rules-based and automatable, but jurisdictional fragmentation is the moat. |
| Concept design & massing | 06 | Developer | Architecture firms | $25k–450k | 44% | 3 / 10 | Generative tools accelerate options; selection stays human. |
| Physical scale models | 13 | Developer | Model makers | $5k–100k+ | 58% | 1 / 10 | CNC already automated the fabrication. Physical constraint caps the upside. |
| MEP engineering | 08, 09 | Architect | Engineering consultancies | $35–100 / hr | 56% | 7 / 10 | Layout is repetitive; the stamp is not transferable. |
| Legal & contract drafting | 04, 11, 15 | Developer | Real estate counsel | $10k–200k | 52% | 8 / 10 | Drafting accelerates; negotiation and liability do not. |
| Structural engineering | 08, 09 | Architect | Structural firms | $30–90 / hr | 46% | 9 / 10 | Calculation is trivial; professional responsibility is the entire product. |
| Planning & entitlement consulting | 07 | Developer | Planning consultancies | $15k–100k | 28% | 6 / 10 | Relationship and political capital. Least disruptable high-value service. |
| Environmental consulting | 03, 07 | Developer | Environmental firms | $2.2k–250k | 26% | 8 / 10 | Reporting drafts; fieldwork and sampling do not. |
| Geotechnical investigation | 03 | Developer | Geotech engineers | $3k–50k | 22% | 9 / 10 | You cannot automate a borehole. |
| Land surveying | 03 | Developer | Licensed surveyors | $3k–15k | 18% | 10 / 10 | Statutory monopoly on certifying boundaries. Structurally protected. |
| Bold buyer indicates the developer is not your customer at that stage — the architect or GC purchases the service out of their own fee. “Repeatable production” is the estimated share of billable effort that is document generation rather than judgment, coordination or fieldwork. | |||||||
Plotting the two variables together shows why the obvious targets are crowded and the valuable target is not.
Figure 6
The lower-right quadrant — high repeatability, low liability — is where software wins outright. The upper-right is the contested zone, and it is where the money is.
The structural point
Liability in AEC production attaches to the professional who signs, not to whoever generated the drawing. That is why offshore drafting has been legal and normal for two decades, and it is why AI-generated documentation is not blocked by professional regulation in the way people assume. The architect’s stamp is the moat — and the architect is a customer for production capacity, not a defender against it.
The trade every services business in this market has to make consciously, because making it accidentally is how firms end up competing on price in the wrong quadrant.
Fastest yes in the market. Principal signs personally. Terrible unit economics on their own — but the only reliable route to an incumbent position before Stage 06.
Real money, real deadlines, and a buyer sophisticated enough to value throughput. Hard to enter cold; straightforward to enter as an incumbent from the wedge.
Either incumbent-locked, relationship-protected, or too small to build a business on. Ignore.
Good revenue, long sales cycles, entrenched competitors. Viable as a second act, punishing as a first one — except Stage 13, where the budget already exists and the incumbent is usually a generalist agency.
The strategically interesting move is the diagonal: enter at the wedge, monetise at the prize. A firm that underwrites a developer’s deal in 48 hours for $8k at Stage 02 has, by Stage 09, a relationship with the Principal, the architect and the GC on a project where $750k of documentation is about to be produced. Nobody entering cold at Stage 09 has that.
Collapsing these into “property developer” is the most common and most expensive positioning error in this market.
Stages 01–07
Founder / Head of Acquisitions
Stages 08–12
Architect of Record / GC Project Manager
Stages 13–14
Marketing Director / Asset Manager
What already exists at each layer, who owns the budget for it, and where the gap sits.
| Capability | Stage | Incumbent stack | AI-native entrants | Budget owner | Gap |
|---|---|---|---|---|---|
| Land & parcel intelligence | 01 | CoStar, CREXi, ESRI, county GIS | Parcel-scoring and zoning-search tools | Developer — overhead | Nothing connects a scored parcel to an underwriting model |
| Underwriting & pro formas | 02, 05, 11 | Excel, Argus | AI underwriting and scenario generation | Developer — at-risk | Lender-format output; nobody solves the “every lender wants it differently” problem |
| Capital raising | 05 | PowerPoint, DocSend, CRM | Deck generation, investor matching | Developer — at-risk | Model, deck and renderings never share a source of truth |
| Due diligence orchestration | 03 | Email, Bluebeam, spreadsheets | Document review, title analysis | Developer — at-risk | No one owns the critical path across six consultants |
| Generative & concept design | 06 | Rhino, SketchUp, Revit | Autodesk Forma; Skema for catalogued design reuse | Architect | Design options are not linked to the pro forma that constrains them |
| Visualisation | 05–07, 13 | Lumion, Enscape, D5, Twinmotion | Text-to-render and AI style transfer; per-image cost from $0.04 | Developer & architect | Consistency across a campaign, not one-off images |
| Entitlement & code compliance | 07, 09 | PDFs, municipal portals, GIS | Zoning copilots, code-review point solutions | Developer — at-risk | Jurisdictional fragmentation; no multi-market coverage |
| BIM & documentation automation | 08, 09 | Revit, Navisworks, ACC | SWAPP (Revit-native CD generation, 35M+ sq ft delivered); firm-standard learning | Architect | Largest pool, most credible entrants, still early on firm-standard fidelity |
| Estimating & takeoff | 10 | STACK, ProEst, Excel | Togal.AI and similar PDF-native takeoff | Developer & GC | Takeoff automates; pricing judgment and scope gaps do not |
| Construction coordination | 12 | Procore, Autodesk Build, Oracle | RFI agents, AI scheduling, reporting | GC | Incumbent-owned system of record; entrants must integrate, not replace |
| Site monitoring | 12 | OpenSpace, Buildots, drone firms | Computer-vision progress tracking | GC | Progress data does not feed the draw request it should justify |
| Marketing & leasing | 13 | Adobe, HubSpot, Webflow | AI creative, copy, campaign generation | Developer — reserve | Generalist agencies hold the budget; nobody is vertical-specific at quality |
| Asset operations | 14 | Yardi, MRI, RealPage | AI leasing, predictive maintenance | Owner — operating | Deeply entrenched; only recurring-revenue pool in the lifecycle |
| The consistent pattern among entrants gaining traction: purpose-built AI for one hard, well-defined task — not a general-purpose platform. Every attempt at an end-to-end “development OS” runs into the fact that no single party owns the whole lifecycle’s budget. | |||||
The gap worth naming
Every participant is assembling a private AI stack — the architect has one copilot, the estimator another, the developer a third, the agency a fourth. No one owns the project intelligence layer that sits above all of them, and the reason is structural rather than technical: the budget for each stage belongs to a different party, and none of them will fund a system that primarily benefits the others.
That does not make the layer unbuildable. It means whoever builds it has to be paid stage by stage, by whoever owns that stage’s budget, while quietly accumulating the cross-project data asset. Which is a services business that becomes a software business — not the reverse.
Derived from the map rather than assumed in advance. Scored on friction intensity, budget access, speed of decision, and how repeatable the work is across their portfolio.
$5M–40M deals · 2–8 staff · Principal decides everything
Maximum friction, minimum budget. No in-house analyst, no established consultant bench, and every pursuit dollar comes out of their own pocket. They will pay for speed and for the credibility that gets them past a lender. They will not pay much.
Best entry: Stage 02–05, priced as a productised package. Risk: churn when they lose deals, which is often.
8–25 staff · Development Manager executes, Principal approves
The sweet spot on paper: enough volume that workflow improvements compound across deals, small enough that they still outsource nearly everything, and large enough to have both at-risk and budgeted spend. Buys at every stage of the map.
Best entry: Stage 02 wedge, expand to 05, 06 and 13. Risk: they already have incumbent consultants and switching requires a failure.
15–150 staff · $50k–1M annual outsourcing budget · Principal or Project Architect decides
Not a developer at all, and the most overlooked buyer on the map. They already outsource, already have a budget line for it, already understand per-deliverable pricing, and are structurally unable to staff for peaks. Their pain is throughput and rework, and both are measurable.
Best entry: Stage 08–09 production, on a paid pilot. Risk: they are price-disciplined and will benchmark against $15–40/hr offshore.
Fee-based, 0.5–1.5% of hard cost · sits between every party on the map
Structurally the most interesting buyer: paid to coordinate exactly the friction this document catalogues, on multiple concurrent projects, with no design liability of their own. They are the only party whose economics improve directly and immediately from better coordination.
Best entry: Stage 03 and 12 coordination. Risk: small, fragmented market; they buy tools rather than services.
50–300 staff · procurement, vendor panels, legal review
Biggest budgets, longest cycles, most competition, and an existing panel of preferred consultants. Every engagement requires insurance, indemnities, and a security review. Worth pursuing only once there is a reference base and the delivery model is proven.
Best entry: not yet. Risk: a 9–18 month sales cycle will consume a small firm.
Our read
The map points at a two-ICP strategy, not one. Sell certainty to Mid-market developers (B) at Stages 02–05 — low price, high speed, high frequency, and the fastest route to a Principal relationship. Sell capacity to Boutique and mid-size architecture firms (C) at Stages 08–09 — where the budget already exists, the buyer already outsources, and the work is the most repeatable in the industry.
They are the same lifecycle and often the same projects, which means one delivery capability can serve both. But they are different sales motions and should be resourced as such. The open question worth resolving next is whether the Stage 02–05 wedge can be delivered at a gross margin that survives a 70% deal-mortality rate — and if not, whether it should be run deliberately at cost as customer acquisition for Stage 08–09.
Cost ranges are published 2025–2026 market figures. The reference-deal allocation, the automation and liability scores, and the ICP scoring are our own analysis.
Sourced: all service cost ranges, A/E fee percentages and phase splits, capital stack percentages and return targets, entitlement costs by jurisdiction, offshore production rates, developer and GC fee percentages, soft cost share of total development cost.
Our analysis: the $40M reference-deal allocation, the two-regime framing, cash-at-risk modelling in Figure 4, per-stage service spend in Figure 5, repeatable-production and liability-barrier scores in Section 08 and Figure 6, the urgency×budget matrix, buyer segmentation, and all ICP scoring.
Figures should be treated as planning-grade, not transaction-grade. Costs vary by a factor of three or more between jurisdictions, and entitlement in particular is dominated by which city you are in rather than what you are building.