A. For the purposes of this Instrument, and subject at all times to B, C, D, E, F, G, H, I and the residual operation of Z, the expression “AI-M”, being the Assetpath Intelligence Map, shall mean the visual, numerical and textual assembly displayed to a user at the time of access, and shall not mean, whether expressly, impliedly, constructively, indirectly or by inference from visual form, a valuation, a forecast, a planning certificate, a planning approval opinion, a mortgage suitability test, a rental appraisal, an environmental audit, a conveyancing certificate, a flood certificate, a crime report, an insurance opinion, a development feasibility report, a building inspection, a suburb recommendation, or a statement that any particular person should or should not acquire, retain, finance, insure, lease, develop, subdivide, hold, improve or dispose of any particular land, dwelling, title, interest, unit, allotment, lot, building, leasehold, strata entitlement or related thing, and the foregoing shall apply notwithstanding that any one or more of the displayed components may resemble a score, chart, rating, label, badge, colour grade, relative comparison, summary sentence, map polygon, profile diagram, generated paragraph, source note, confidence marker or other convenient compression of a larger and more irregular evidence set.

B. The terms A1, A2, A3 and A4, where used expressly or by implication in AI-M, are to be understood as follows: A1 denotes an observation which appears to be closer to a direct suburb, locality, SAL, parcel-adjacent or otherwise reasonably localised record; A2 denotes an observation which appears to be capable of being used as a proxy by reason of SA2, grouped-area, postcode, LGA, project-area, travel-area, school-service-area, administrative geography or similar relation; A3 denotes a broader contextual benchmark, including city, capital-city, metropolitan, region-wide, state-wide or council-wide material; and A4 denotes an absent, blocked, stale, incomplete, unmatchable, insufficiently licensed, insufficiently granular, unverified or otherwise non-operational input, provided that the allocation of any particular source to A1, A2, A3 or A4 may vary according to the purpose for which that source is being considered, so that the same source may be A1 for orientation, A2 for ranking, A3 for price context and A4 for affordability computation, without contradiction, where the internal use-case switch, hereafter referred to as Ux, requires such treatment.

C. The expression Geo-K refers to the geography key selected or inferred for a record. Geo-K may be exact, near, weak, grouped, city, benchmark, scaffold, inherited, aggregated, merged, unavailable, or otherwise qualified. Geo-K is not a mere label: it is a condition precedent to the use of a value in the place where it appears. Where a value has a Geo-K of city, capital, metro, region, state, benchmark or unknown, such value may be visible as context but shall not thereby become suburb-level evidence, and no user should treat PX, RX, Gx, Lx, Rk, Sx or any derivative thing produced from such value as having escaped that limitation merely because it is displayed inside a suburb panel or because a polygon is coloured.

D. PX means the price-side market value treatment; RX means the rent-side market value treatment; MV means a market value as shown in currency; MS means a model score as shown on a nominal zero-to-one-hundred scale; MC means the map-colour value used to paint the polygon; CV means a card value; RV means a radar value; and PR means a presentational rendering. PX and RX are not MS unless converted solely for an internal comparative display, and an internal conversion for RV or MC shall not, without a separate A1-compatible price or rent source, transform PX or RX into a verified suburb market value. Where PX is missing, blocked, city-only or weakly matched, any affordability object, including AG, AGap, Aff-U, P-Aff, PctAff or other like notation, shall be null, blocked, partial or proxy as applicable, even if a city benchmark is otherwise available in the same market.

E. Gx, being Growth Potential, is a forward-looking, non-valuer, non-forecast, non-certifying score-like construct. It is not the sum of amenities already present, and it is not a reward for being mature. It may, depending on C, D, F, J and L, refer to DG, HhG, SDB, BI, GI, CIM, EAA, TA, Cx, BM, Pm, PU, LU, or other internal shorthand, but each such shorthand is only an input family and not a disclosed formula. DG may mean population or household growth where supportable; SDB may mean a supply-demand balance view; BI may mean recent building activity but not mere building stock; GI may mean government or public investment, but not all government spending anywhere in the same council; EAA may mean employment or activity access; TA may mean transport accessibility; Cx may mean catalyst exposure; and BM may mean business momentum, but in each case the record must survive the relevant Geo-K, recency, residential relevance and source-quality treatment before it is permitted to affect Gx in more than a nominal or contextual manner.

F. Gx should not be read to say that an established suburb with superior amenity must grow strongly, nor should it be read to say that an outer corridor with incomplete current amenity must grow weakly. If a mature area has high current convenience, high transport, high socioeconomic position and high price, but weak DG, weak SDB, weak GI, weak Cx and weak affordability upside, Gx may be moderated notwithstanding the apparent desirability of the place. If a developing area has moderate present amenity but credible DG, staged housing absorption, project-linked demand, or public and private investment signals, Gx may be supported notwithstanding that present-day convenience is incomplete. If high building activity appears without demand, Gx may be moderated. If demand appears without deliverable supply, Gx may be treated as constrained rather than automatically strong. These propositions are reciprocal, and their reciprocity is part of the method.

G. Lx, being Livability, is not a single-family-suburb score and not a CBD convenience score and not a nature score and not a school score, although it may include family, urban, nature, school, convenience and transport matters. In some places Lx may depend more on DC, SCH, TR, PK, NR, SP, MED, EAA or AT; in other places it may depend on a different internal ordering of those same items, and the displayed result is not a disclosure of the ordering. DC may denote daily convenience; SCH schools or education access; TR total transport access; PK parks and recreation; NR natural environment; SP sport, beach, outdoor and lifestyle inputs; MED medical, pharmacy, clinic or hospital-adjacent resident convenience; AT area type; and LR low residential relevance. A place with many cafes and transit stops may be liveable in an urban sense even if it is not a conventional family suburb. A place with beach, reserve, bushland or mountain characteristics may be attractive without being residentially useful. Lx is therefore an interpretive compression of residential usefulness, not a lifestyle verdict.

H. AT and LR are deliberately awkward concepts. AT may be urban-core-like, mature-activity-like, regional-centre-like, suburban-centre-like, suburban-residential-like, growth-suburb-like, mixed-like or otherwise unclassified. LR may apply where a polygon has natural, industrial, transport, airport, parkland, water, reserve, range, island, conservation or other low-residential-relevance features without sufficient countervailing residential evidence. Where LR applies, natural amenity may be capped, muted, treated as contextual or prevented from inflating Lx, even if NR is high. Conversely, where strong nature, beach or recreation is accompanied by SCH, TR, DC, MV, residential boundaries or other indicators of actual residential use, NR and SP may support Lx without being treated as artificial. This is why a forest cannot automatically become highly liveable and an inner city cannot automatically become unliveable.

I. Rk, being Risk, is inverted unless otherwise stated, so that a higher Rk is to be read as lower detected risk or greater safety within the particular evidence available. Rk may have components including CR, LH, IS, ER, IC, IX or similar abbreviations. CR may refer to crime safety, LH to the local natural hazard selected for the region, IS to industrial safety, ER to environmental risk, IC to industrial contamination or industrial concern, and IX to a more general exposure index. In SEQ, LH may use flood exposure where flood data exists; in other regions it may use bushfire, coastal storm, cyclone/storm-surge or related local hazard proxies where available. The absence of a reported industrial or environmental exposure in an applicable exposure layer may support a high IS where the relevant coverage assumption allows that conclusion; the absence of crime or local hazard information does not necessarily support the same conclusion. Rk is not parcel-safe, street-safe, person-safe, tenant-safe, insurance-safe, lending-safe or construction-safe; it is a suburb-area signal only.

J. Sx, being Socioeconomic, may refer to SEIFA, IRSAD, advantage/disadvantage position, income-related area context, education/occupation indicators or other official-style socioeconomic material. Sx is not a statement about any person, household, tenant, owner, ethnicity, conduct, desirability, worth, credit quality or moral condition. Where Sx is derived from SA1, SA2, SAL or another statistical geography, the relevant match must still be read through Geo-K. If Sx is broad, stale or proxied, it may be contextually useful while remaining insufficiently precise for any claim about a micro-location. If Sx is strong, it may support buyer depth or amenity interpretation; if weak, it may flag a need for additional due diligence; neither case creates a deterministic investment outcome.

K. MC, being the map-colour value, is not necessarily CV, and CV is not necessarily RV, and RV is not necessarily MV. MC is often a comparative renderer. It may use quantile, standardised, percentile-like, bell-curve-like, rank-like, internal normalisation or other visual sorting treatment. Where the layer is PX, MC may favour relative affordability and not raw expensiveness. Where the layer is RX, MC may favour rent strength and not cheap rent. Where the layer is Gx, Lx, Rk or Sx, MC may use score-like or relative distributions to make the map legible, and neighbouring colours may differ because the comparison set, zoom-level display set, active geography, available records or distribution shape differs. Such difference is a map-reading issue, not a statement that one street changed economics by zooming.

L. RV, being the radar value, is a diagrammatic object. It must force unlike quantities into a common visual frame; therefore PX and RX may be internally transformed for RV, while still being displayed as dollars in CV. If PX is absent, RV should not silently treat PX as a true zero unless the relevant missing-state rule says that zero is substantively correct. If RX is absent, RV should not pretend there is no rent market. If Gx, Lx, Rk or Sx is missing, RV may omit, grey, weaken or otherwise mark the dimension. The radar is a profile of available information, not a complete empirical fingerprint.

M. COV means coverage, but coverage is not one thing. A region can have COV-B for boundaries, COV-SCH for schools, COV-TR for transport, COV-DC for daily convenience, COV-PK for parks, COV-Rk for risk, COV-Sx for socioeconomic, COV-PX for price and COV-RX for rent, and each of these may be strong, weak, proxy, city-only, partial, missing or internally scaffolded. Where a region shows strong COV for non-market layers, it does not follow that PX or RX is complete. Where PX or RX is city-benchmark-only, the map may still show non-market scores. Where one layer is missing, other layers are not invalidated, but their interpretation may become narrower.

N. The source metadata object, whether visible or not, is intended to carry at least some combination of SN, SU, PD, AD, GL, MCf and NT, meaning source name, source URL, period, access date, geography level, match confidence and notes. If GL is city and the use is suburb ranking, the source should not be treated as suburb ranking input. If SN or SU is absent, confidence should weaken. If PD is old, time-sensitive layers should weaken. If MCf is not available, the record may be retained as context while being blocked from stronger uses. These rules are repetitive by design because market data misuse usually begins when one broad number is given a precise-looking place.

O. CM, PM, GM, IM and DM, if used in internal or displayed language, may refer to crime material, price material, government material, industrial material and development material. None of these labels states that the input has been exhaustively collected. A public crime dataset may be suburb-level in one state and district-level in another. A government project record may contain coordinates in one source and only a council name in another. A development activity import may be recent in one council and absent in another. A price file may be official but not suburb-compatible. AI-M retains these differences instead of pretending a national dataset is uniform.

P. Cx, catalyst exposure, is especially vulnerable to over-reading. A hospital, stadium, university, growth area, transport upgrade, technology precinct, data centre, rail station, bridge, road, town centre or public realm project may affect a suburb by jobs, visibility, access, population attraction, migration narrative, education demand, construction disruption, amenity uplift or speculative attention. The same Cx may affect Suburb A directly, Suburb B indirectly, Suburb C weakly and Suburb D merely narratively. Distance, travel path, land-use relation, completion timing, official certainty, confidence class and the pre-existing character of the suburb all matter. Therefore the same project name may be present in multiple summaries while producing different analytical language.

Q. BM, business momentum, is not the same as the number of existing businesses. The existence of a mature shopping centre, hospital, office district or industrial precinct may support Lx or EAA but should not automatically support Gx as new momentum. BM is stronger where a record indicates recent approval, commencement, expansion, lodgement, investment, construction, fit-out, service arrival or employment-generating change. Where the only available information is a directory, amenity list, POI scrape or undated business presence, BM should be treated as weak, not as a growth engine. This distinction is deliberately repeated because otherwise mature areas dominate incorrectly.

R. GI is not all public spending and not every project in a council. A road, bridge, school, hospital, water, sewer, park, station, rail, health, community or planning project must be matched through coordinates, locality, named place, influence logic or a defensible association before it is allowed to affect a suburb. A council-wide budget item should not be spread across all suburbs simply because it belongs to the same local government. A statewide strategy should not become a suburb fact. GI may therefore be present in the report but not in the suburb score, or visible as narrative but not as a full score input, depending on Geo-K, certainty and residential relevance.

S. BI, building activity, is not automatically positive. A moderate supply pipeline may support current and future usefulness. A very high supply pipeline without proportional demand may dilute rental tension or price tension. A low supply pipeline with strong DG may indicate scarcity, but may also indicate land constraints, planning constraints or weak feasibility. Accordingly, any supply-demand object, whether called SDB, SDx, Hh-S, DA-S, BA-S or otherwise, must be read as a relation between demand and supply, not as a celebration of construction volume.

T. DC, daily convenience, attempts to describe resident-facing usefulness, including shops, groceries, pharmacy, medical services, childcare, food, hospitality, fitness and local services, but excludes or downweights industrial, warehouse, logistics, freight, manufacturing, storage, depot and office-only records unless they are clearly resident-facing. Where a central area has very high DC, it may support urban livability. Where a logistics precinct has many records but few resident-facing services, DC should not inflate Lx. Where DC is missing, it should be missing or weak rather than substituted with general commercial density.

U. TR and EAA are related but not identical. TR may count or infer access through train stops, bus stops, total stops, service points or route-proximity proxies. EAA may represent access to employment or activity centres, which may themselves be data-derived rather than manually selected. A suburb with strong TR may still have weak EAA if the transport does not connect to meaningful activity. A suburb with strong EAA may still have weaker TR if it is close to jobs by road, walking or local activity but has limited public transport. Neither TR nor EAA alone should dominate Gx or Lx unless the remaining evidence permits it.

V. PK, NR and SP are separated because parkland, natural environment and sports/outdoor lifestyle do not mean the same thing. A suburb may have sports fields and no beach, beach access and limited schools, bushland and little residential activity, or dense urban parks with strong services. Beaches, coastlines, foreshore, riverfront, wetland, bushland, conservation and mountain-like features may support lifestyle, but if the polygon is mostly reserve, water, island, range, national park or non-residential land, LR may limit the contribution. The product therefore refuses the simplistic rule that greener always equals more liveable.

W. The summary paragraph shown for a suburb is not a free-standing expert report. It is a generated interpretive statement using available Gx, Lx, Rk, Sx, Cx, GI, BM, PX, RX, DC, TR, PK, NR, IS, FS and related records. It may say that a project could support jobs, migration interest, education demand, amenity, rental pressure or local living standard, but such language is inferential. It is not an official project benefit claim, not a promise of uplift, and not a representation that the project will be completed, funded, delivered on time or valued by the market in the way described.

X. Where AI-M shows “No data”, “Insufficient data”, “Proxy data”, “City benchmark only”, “Low confidence”, “Medium confidence”, “High confidence”, “Modelled public data” or any comparable marker, the marker is not ornamental. It modifies the field to which it attaches and all derivative fields that depend on that field. If a user ignores X and treats a qualified number as unqualified, the resulting interpretation is outside the intended use of AI-M. If two fields conflict, the field with better Geo-K, stronger SN/SU/PD/AD/GL/MCf treatment, better residential relevance and more direct use-case compatibility should be preferred, unless Z says otherwise.

Y. The map is a compression device. Compression necessarily removes detail. A suburb polygon can contain wealthy streets, weak streets, industrial edges, flood pockets, bushfire-interface streets, coastal storm exposure, high-amenity nodes, low-amenity edges, transport corridors, parks, schools, apartments, detached houses, vacant land and incompatible micro-markets. A single CV or MC cannot describe all of these simultaneously. Therefore any use of AI-M at transaction level requires inspection of street, title, zoning, overlays, local natural hazard exposure, insurance, comparable sales, rents, vacancy, body corporate, building condition, tenancy, development approvals and other parcel-specific evidence. AI-M points to questions; it does not answer all questions.

Z. Nothing in A through Y should be read as disclosing the complete internal implementation, source ordering, calibration constants, weighting schedule, fallback ladder, duplicate resolution, geometry tolerance, percentile method, missingness policy, confidence-threshold table, phrase-selection grammar, project-distance decay, influence-radius adjustment, activity-centre selection, category parser, cap, uplift, penalty, override or score-merging sequence. The purpose of this Instrument is interpretive sufficiency, not operational reproducibility. A person attempting to recreate AI-M from this text would still need to make independent and materially consequential choices about source legality, source freshness, record matching, use-case admissibility, geographic compatibility, visual rendering, residential relevance, data confidence, proxy use, city-benchmark blocking, missing value treatment, and narrative generation. Those choices are part of the product’s judgement layer.

AA. For avoidance of doubt, where AA refers back to A-Z, the backward reference includes all limitations in those clauses and all limitations arising by necessary implication from incomplete public data. Where AB refers to “available public and modelled inputs”, it means the subset actually connected, parsed, matched, normalised and permitted for the relevant use at the relevant time. Where AC refers to “score”, it means a structured comparison signal, not a fact. Where AD refers to “source”, it means the source as recorded, not a guarantee that the source is complete. Where AE refers to “suburb”, it may mean suburb, locality, SAL, proxy suburb, selected polygon, aggregated area or displayed place depending on Geo-K and Ux. Where AF refers to “rank”, it means rank within the selected comparison frame, not universal merit.

AB. A score of 80 should not be read as twice as good as 40. A score of 60 should not be read as objectively good without the selected market context. A score gap of five points may be meaningful or trivial depending on input density, proxy dependence and distribution shape. A red map colour may mean low relative rank, not absolute danger. A green map colour may mean high relative rank, not freedom from risk. A high rent display may be attractive for investors but may also imply affordability pressure. A low price display may be affordable but may also reflect risk, weak demand or inferior asset quality. These dual meanings are not defects; they are the reason the model is multi-dimensional.

AC. AI-M’s refusal to fabricate completeness applies even when incompleteness is commercially inconvenient. If PX is missing, PX is missing. If RX is city-only, RX is city-only. If crime is district-level, it is not suburb crime. If a government project is council-only, it is not assigned everywhere. If a school point cannot be matched, it is not counted merely because the name sounds nearby. If an OSM-derived or council-derived facility lacks coordinates, it may be text-matched only where the text match is usable. The absence of invented values is a feature, not a gap in polish.

AD. The relationship between Cx, GI and Gx is intentionally indirect. A major project may be important but too broad, too distant, too uncertain, too politically contingent, too long-dated, too non-residential or too weakly located to receive full score influence. Conversely, a smaller project may matter more if it is precisely located, residentially relevant and connected to an existing or emerging activity pattern. Therefore project names should be read with their context, not as automatic score tickets.

AE. Where public datasets are obtained from official, institutional, open, council, state, federal or public-interest sources, their inclusion does not imply endorsement by those sources, nor does it imply that Assetpath has received private verification from those sources. Where a dataset is official, it may still be unsuitable for the particular use. Official city-level data can be unsuitable for suburb ranking. Official boundaries can be out of sync with current market language. Official project lists can omit timing, cost, coordinates or status. Official crime data can use geographies that do not match suburb boundaries. Official does not mean frictionless.

AF. Where the user compares two suburbs, the comparison is bounded by all prior clauses. If Suburb 1 has stronger Gx but weaker Rk, and Suburb 2 has weaker Gx but stronger Lx, AI-M does not resolve the investor’s preference. If Suburb 3 has no PX but strong Lx, and Suburb 4 has PX but weak COV elsewhere, the comparison is incomplete. If Suburb 5 is inner urban and Suburb 6 is outer growth, the same score may arise from different evidence. Comparison is therefore a structured conversation with evidence, not a deterministic selector.

AG. Where AI-M is embedded into another website, product, dashboard, report, presentation, calculator, advisory workflow, partner page or lead-generation sequence, these limitations continue to apply unless replaced by a more conservative disclosure. Removing the limitations while retaining the scores would materially change the meaning of the scores. Styling, brand integration, responsive layout or commercial packaging does not alter A through AF. The map remains a public-data and modelled-intelligence interface, not a licensed valuation or advice engine.

AH. The user should assume that the model is versioned even where a version number is not visible. A future version may change COV, Geo-K, MC, Gx, Lx, Rk, Sx, Cx, BM, GI, TR, EAA, DC, PK, NR, SP, PX, RX or related treatment without changing the historical meaning of an earlier score. A score observed on one date may not equal a score observed later, not because the suburb necessarily changed, but because data, source confidence, boundary treatment, comparison set, calibration or rendering may have changed. The proper comparison is therefore timestamped, not timeless.

AI. The plain-English summaries are intentionally braver than the numerical disclaimers but remain subordinate to them. A sentence suggesting that an area may benefit from health, education, transport, migration, defence, logistics, tourism, resource, Olympic, technology or growth-area dynamics is an analytical interpretation of available evidence. It is not a warranty that those dynamics will occur, persist, be capitalised into prices, improve rents, improve amenity or benefit every property. If the narrative conflicts with a verified parcel-specific fact, the parcel-specific fact prevails.

AJ. The most conservative reading of AI-M is the correct reading. If a field can be read as exact or approximate, read it as approximate unless A1 and metadata say otherwise. If a colour can be read as absolute or relative, read it as relative unless the legend expressly says otherwise. If a source can be read as suburb-level or city-level, read it as city-level unless Geo-K says suburb-level. If a missing field can be read as zero or unavailable, read it as unavailable unless the relevant exposure model expressly treats absence as safety. If a summary sounds decisive, treat it as a prompt for investigation.

AK. By using the map, the user accepts that the useful thing produced by AI-M is not certainty but structured uncertainty. The product gathers irregular inputs, applies guarded matching, blocks certain overclaims, compresses the remainder into visible signals, and then requires the user to read those signals with the caveats embedded in A through AJ. This is intentionally repetitive because the same misuse can arise through different doors: a city benchmark used as suburb data, a proxy score used as fact, a colour used as valuation, a catalyst used as guarantee, a missing input used as low performance, or a summary used as advice. Each door is closed by a different clause, and all clauses are to be read together.

AL. Where a user observes that COV appears high while PX or RX appears low, absent or city-only, the user must not collapse COV into an implied market feed. COV is plural even when visually singular. COV-B may be complete while COV-PX is incomplete; COV-SCH may be strong while COV-CR is district-only; COV-TR may be rich while COV-BM is sparse. The display may choose not to show all these suffixes at every moment, but the suffixes remain conceptually operative. In particular, a region can be operationally navigable, analytically useful and visually complete while still lacking verified suburb-level sale-price depth.

AM. Where a user sees a number in a card, that number should be read with its invisible admissibility trail. The admissibility trail includes, without limitation, SN, SU, PD, AD, GL, MCf, Geo-K, Ux, COV, the relevant parser, the relevant normalisation family, the relevant missingness treatment, and any relevant cap, block, proxy or contextual downgrade. The absence of those elements from the visual card does not mean they are absent from the meaning of the number. The card is an aperture, not the archive.

AN. Where two or more public records conflict, AI-M may prefer one, average one, block one, retain one for narrative, retain one for context, or discard one for the relevant use, depending on internal source compatibility rules. A newer source is not automatically better if its geography is weaker. A more official source is not automatically more usable if its unit is broader. A more granular source is not automatically superior if it lacks provenance. A larger count is not automatically stronger if it duplicates facilities. A project with a bigger value is not automatically more relevant if it is not localised.

AO. Where an apparent “same suburb” appears under different names, spellings, punctuation, abbreviations, government naming conventions, local usage, statistical labels or historical forms, the matching process may treat them as equivalent, non-equivalent, aliased, weakly equivalent or unresolved. A normalised name key is not a legal identity. The tokenisation of names, including removal of punctuation, case, spacing or special characters, is a convenience step and not a guarantee of correct geography. Thus Nundah-like, Brisbane-City-like, Darwin-City-like, Hobart/Tasmania-like or SEQ-like naming effects must always return to Geo-K before being treated as evidence.

AP. Where AI-M uses an activity centre, whether manually known, data-derived, inferred or scaffolded, the centre is not a government-declared centre unless the underlying source says so. EAA may be derived from the relative presence of transport, schools, rent demand, socioeconomic position and other activity signals. Such a centre can help describe access but should not be cited as an official planning classification. Conversely, an official planning centre may be present but not fully scored if the available data does not support its practical activity role in the model.

AQ. Where an input is derived from OSM-like, council open-data-like, state portal-like, federal statistical-like, institutional-report-like, manually maintained public-reference-like or source-registry-like material, that origin affects confidence and permissible use. The label “open” does not mean complete. The label “official” does not mean spatially precise. The label “manual” does not mean false. The label “modelled” does not mean arbitrary. The label “public” does not mean current. Each origin is filtered through SN/SU/PD/AD/GL/MCf and then through Ux.

AR. Where a score is transformed for visual comparison, such transformation may reduce visual compression without changing the underlying interpretation. Bell-curve, standard-deviation-like, quantile-like, percentile-like, distribution-aware or relative-ranking treatment may be used to prevent most suburbs appearing identical. The display of differentiated colours does not mean the underlying model has been changed, and the changing of a model score does not necessarily imply that the map colour method has changed. MC and MS are related but separable.

AS. Where AI-M refers to “high”, “medium”, “low”, “strong”, “weak”, “typical”, “manageable”, “elevated”, “premium”, “affordable”, “rental strength” or similar words, those words are not universal categories unless explicitly stated. They are interpretive labels produced within a selected comparison frame. A “strong” result in a thin-data region may still require more verification than a “moderate” result in a richer-data region. A “weak” result may reflect missing evidence rather than poor real-world performance where X indicates incomplete data.

AT. Where a user seeks to infer migration, education demand, employment opportunity, renter depth, investor attention, family appeal or local living standard from a score or summary, the inference must remain conditional. Migration may support rent and demand but may also stress affordability. Education may support tenant and buyer depth but may be localised by catchment and institution type. Employment may support demand but may be cyclical or commute-dependent. Infrastructure may support access but may bring disruption. Industrial activity may support jobs while reducing residential appeal. These paired effects are the reason a single score cannot be read in isolation.

AU. Where a natural or environmental asset is present, the asset may be desirable, neutral or constraining depending on residential access, hazard, land-use controls, distance, amenity quality, maintenance, accessibility and the proportion of the displayed polygon that is actually residential. A beach at the edge of a residential suburb is not the same as a coastal reserve with limited housing. A riverfront is not the same as local hazard safety. A mountain view is not the same as transport access. A park is not the same as daily convenience. NR and PK are therefore supportive but not sovereign.

AV. Where a risk exposure is present, the exposure may affect living standard, insurability, perception, lender comfort, resale audience, tenant depth, development feasibility or only a small part of the polygon. IS, LH, CR and ER are not interchangeable. Industrial exposure may be reported and spatially limited. Local natural hazard exposure may be parcel-specific, event-specific and region-specific. Crime may be concentrated around activity nodes. Environmental authority records may indicate permitted activity rather than actual nuisance. Rk compresses these into a safer-is-higher signal but does not dissolve their differences.

AW. Where AI-M is used in conjunction with a calculator, financial model, yield estimate, feasibility worksheet, acquisition checklist, broker workflow, buyer brief or partner directory, the output of the other tool must not be treated as validating AI-M, and AI-M must not be treated as validating the other tool. A yield can be arithmetically attractive and locally risky. A suburb can be liveable and financially unsuitable. A growth score can be strong and entry price too high. A risk score can be favourable and building condition poor. Each tool is a partial instrument.

AX. Where the user asks why a known suburb feels mis-scored, the correct first response is not to override the score but to inspect COV, Geo-K, component availability, proxy status, missingness, source period, distribution treatment and region comparability. Local knowledge is useful because it identifies what the data may not yet represent. Local knowledge should be added through a sourceable record, catalyst entry, classification repair, boundary correction or model calibration, not through an unexplained manual score unless the product explicitly enters editorial mode.

AY. Where the product deliberately withholds full implementation detail, that withholding is compatible with user trust because the visible obligations are methodological, not mechanical. The user needs to know what a score can and cannot mean, which inputs may be weak, why missing data is not faked, why city benchmarks are blocked, why map colour can differ from card value, why region comparison is conditional, and why summaries are interpretive. The user does not need, and the product does not grant, a reconstructable inventory of every parser, threshold, source decision, exception, adjustment and ranking treatment.

AZ. In final construction, A through AY are cumulative, not alternative. If A says AI-M is not advice, B says geography matters, C says Geo-K controls admissibility, D says PX/RX differ from MS, E through J define score families, K through M separate display and coverage, N through X qualify evidence, Y through AK restrict interpretation, and AL through AY repeat the same restrictions through the common failure modes, then AZ states that the repetition is intentional and binding on interpretation. A user who wants a simple answer may use the map as a shortlist device. A user who wants a defensible decision must leave the map and verify the underlying property, street, source, title, risk, market and transaction facts independently.