Not every sealed document is ready
A licensed engineer in Ontario might sign a dozen different document types across a career - field review reports, condition assessments, reserve fund studies, letters of structural adequacy, Phase I Environmental Site Assessments, building envelope reports, shop drawing reviews, and others. Each document carries the same professional weight: the engineer's seal attests to their judgment, and PEO Regulation 941 defines the standard of care regardless of document type.
But the documents are not equally ready for AI-assisted drafting. Some are narrative-heavy, template-consistent, and produced in high volume - ideal candidates for a drafting tool. Others require computational primitives, regulatory integrations, or document structures that narrative AI cannot handle without significant product engineering.
Three factors determine readiness:
Structural standardisation. How consistent is the document structure across firms and projects? A field review report follows a predictable skeleton - header, people table, observations, findings, photos, sign-off - that varies in detail but not in shape. A Phase I Environmental Site Assessment follows a structure prescribed by CSA Z768 but includes records-search components that vary dramatically by site. The more standardised the skeleton, the less prompt engineering required to produce a reliable first draft.
Narrative-to-computation ratio. How much of the document is prose versus calculation? A field review report is almost entirely narrative - the engineer describes what they observed, classifies findings, and writes recommendations. A reserve fund study is approximately half narrative and half computation - the component inventory is descriptive, but the 30-year cashflow projection is spreadsheet-grade amortisation math. AI excels at narrative generation. It does not excel at financial modelling. A document that requires both needs a product that handles both, and the computational half is a different engineering problem.
Buyer concentration. How many firms produce this document type, and are they the same firms already using the tool? A document type produced by the same structural engineering firms that already use a drafting tool for field reviews is a natural expansion - same buyer, same sales motion, same trust relationship. A document type produced by a different firm type - environmental consultants, reserve fund specialists, building envelope engineers - requires a separate go-to-market.
Architectural general review reports - the nearest target
Every permitted building in Ontario requires general review under OBC 1.2.2.2 - not just structural review, but architectural and mechanical/electrical review as well. The structural engineer, the architect, and the MEP engineer each conduct independent field reviews of their respective scope and each produce independent sealed reports.
Architectural general review reports are produced under the same OBC clause, follow the same workflow (site visit → observation → draft → review → seal), and use the same document structure as structural field review reports. The signer is different - an OAA-licensed architect rather than a PEO-licensed engineer - but the template is approximately 95% identical. The observation vocabulary differs (the architect observes cladding, fenestration, and interior finishes rather than rebar and formwork), but the finding classifications, photo documentation practices, and regulatory citation structure are the same.
This is as close to a "template swap" as sealed-document expansion gets. The generation prompt needs a vocabulary update and a template variant, not new product primitives. The engineering effort is measured in weeks, not months.
The strategic value is multiplicative. On any given project site in Ontario, a structural firm conducting a field review visits alongside an architectural firm doing the same. If the structural firm already uses an AI drafting tool, the architect working the same project sees the tool's output in the form of the structural firm's consistently formatted, quickly produced reports. The referral is built into the workflow.
Condition assessments - the natural extension
Structural condition assessments are narrative-heavy documents that evaluate the current state of an existing building - typically for property transactions, insurance, capital planning, or litigation support. The engineer inspects the structure, documents deterioration with photographs, maps the extent and severity of damage, estimates remaining service life, and recommends remedial work.
The document structure is similar to a field review report but longer and more detailed. Where a field review documents a single visit to an active construction site, a condition assessment documents a comprehensive evaluation of an existing structure - often requiring multiple visits and 20 to 40 pages of narrative with extensive photo documentation.
The buyer persona is the same: structural engineering firms in Ontario. Many firms that produce field review reports also produce condition assessments. The observation vocabulary overlaps significantly - concrete deterioration, corrosion, membrane failure, water infiltration. The generation prompt requires extension rather than replacement.
The readiness factors are favourable: high structural standardisation (the document follows a predictable outline), high narrative-to-computation ratio (the assessment is almost entirely descriptive prose and photographs), and identical buyer persona. The engineering effort is approximately two weeks of template work - significant, but not a new product build.
Reserve fund studies - the most attractive and the hardest
Reserve fund studies occupy a unique position in the Ontario sealed-document landscape. The Ontario Condominium Act, Section 94, legally mandates that every condominium corporation in the province commission a reserve fund study every three years. Ontario Regulation 48/01 specifies the content requirements. The study must be prepared by a qualified professional - a P.Eng, architect, certified engineering technologist, or Certified Reserve Planner, as prescribed by O. Reg. 48/01.
The regulatory hook is the strongest in the entire landscape. Unlike field review reports (which exist because the building code requires construction observation) or condition assessments (which exist because a client requests an evaluation), reserve fund studies exist because a statute with a fixed cadence demands them. Every Ontario condo corporation is a buyer. The market is not discretionary - it is legally mandated.
The challenge: reserve fund studies require product primitives that a narrative drafting tool does not currently possess.
Component inventory. The study must catalogue every major building component - roof, windows, balconies, parking garage, elevators, HVAC systems, common area finishes - with age, condition grade, expected useful life, and replacement cost. This is a structured data problem, not a narrative drafting problem. The inventory is a database with engineering assumptions, not a paragraph.
30-year cashflow projection. The study must project the condominium's reserve fund balance over a 30-year horizon, modelling component replacement timing, inflation assumptions, contribution rates, and deficit scenarios. This is financial modelling - spreadsheet-grade amortisation math that requires computational accuracy, not narrative fluency. An AI that generates plausible-sounding cashflow numbers is worse than useless - it is dangerous.
Different buyer persona. Reserve fund study work is concentrated in a small number of specialist firms - firms whose entire practice is RFS preparation, not structural engineering. These firms use different software, attend different conferences, and have different procurement cycles than the structural firms that produce field reviews. Reaching them requires a separate go-to-market.
The sequencing implication is clear: reserve fund studies are the most commercially attractive expansion target but the hardest to build. The right time to invest in them is after the narrative-drafting product has been validated across multiple document types within the structural engineering buyer persona - proving that the "expand by document type within the same buyer" playbook works before applying it to a new buyer.
Phase I Environmental Site Assessments - the volume play that is not ready
Phase I Environmental Site Assessments are the highest-volume sealed document type in Ontario - an estimated 30,000 to 50,000 per year, driven by commercial real estate transactions. Ontario Regulation 153/04 and CSA Z768 define the content requirements. The document must be prepared by a Qualified Person (typically a P.Eng or P.Geo with environmental experience).
The volume is attractive. The readiness is not.
Phase I ESAs require a records-search integration that does not exist in a narrative drafting tool. The assessment involves querying environmental databases (ERIS, historical aerial photographs, fire insurance plan archives, Ministry of Environment records) for evidence of contamination at or near the subject property. The records search is a significant portion of the professional effort - and it is a data-integration problem, not a drafting problem.
The narrative portions of a Phase I ESA - site description, findings, conclusions, recommendations - are structurally similar to other sealed documents and would benefit from AI-assisted drafting. But the records-search component is the bottleneck that defines the document type, and a drafting tool that handles only the narrative half solves less than half the problem.
Phase I ESAs are option value on the expansion map - worth revisiting after the narrative-drafting product has absorbed several document types and the question becomes whether to invest in data-integration primitives. The minimum build is estimated at six months.
Why the order matters
The cross-vertical research on AI drafting tools reveals a consistent pattern in how successful products expand: they expand by document type within the same buyer before they expand to a new buyer.
Ambience Healthcare expanded across 200+ clinical specialties and subspecialties - cardiology, orthopedics, dermatology, psychiatry, and dozens more. Each specialty is a template pack and prompt pack, not a new product. The physician buyer is the same. The EHR integration is the same. The attestation workflow is the same. Only the clinical vocabulary changes.
The same pattern applies to sealed engineering documents. Field review reports, architectural general reviews, condition assessments, letters of structural adequacy, and shop drawing reviews are all produced by the same buyer - Ontario structural and building-science firms - using the same workflow - site visit, observation, draft, review, seal. Each document type is a template pack and prompt pack on the existing product. The firm profile, attestation workflow, DOCX export, and observation library are shared.
Reserve fund studies and Phase I ESAs break the pattern because they introduce new buyers and new product primitives. These expansions are valuable but riskier - each one is a partial rebuild, not a template swap. The right sequence is to prove the template-swap expansion works (architectural GR, condition assessments) before committing to the product-rebuild expansion (reserve fund studies, ESAs).
The order is the strategy. A tool that tries to serve five document types at once across three buyer personas will do none of them well. A tool that masters one, expands to the adjacent ones that share the same buyer and the same product primitives, and then - with a validated playbook and a stable revenue base - commits to the harder expansions, will build depth that no competitor can replicate without doing the same work in the same order.