Skip to content
WoodIntel™Intelligence Platform
Sign in

AI specification generator

From project brief to tender clause.

Describe the project. The engine returns species, treatment, service life, maintenance plan, formal specification text and a BIM-ready description — with sources.

  • CrossrefResearchDOI registry
  • OpenAlexResearchOpen scholarly index
  • Europe PMCResearchFull-text abstracts
  • Semantic ScholarResearchAI2 corpus
  • DOAJResearchVetted open-access journals
  • OpenAIREResearchEuropean open science graph
  • HALResearchFrench national open archive
  • ZenodoResearchEU project deliverables with DOIs
  • DataCiteResearchResearch datasets
  • ERA5 / CopernicusClimateClimate reanalysis
  • Open-MeteoClimateStation normals
  • EN 350:2016StandardNatural durability
  • EN 335StandardUse classes
  • EN 351-1StandardPreservative penetration
  • EN 16449StandardBiogenic carbon
  • EN 1995-1-2StandardCharring rates
  • EN 1990StandardDesign working life
  • EN 13183-1StandardMoisture measurement
  • EN 599-1StandardPreservative efficacy
  • EN 16755StandardFire-retardant durability
  • EN 14080StandardGlulam requirements
  • EN 16351StandardCLT requirements
  • EN 927-6StandardArtificial weathering
  • EN 15026StandardHygrothermal simulation
  • ISO 15686-8StandardService-life prediction
  • ISO 21887StandardGlobal use classes
  • EN 15804+A2LCADeclared environmental impacts
  • Scheffer indexModelDecay hazard
  • Lacy driving-rain indexModelFaçade exposure

Start with your conditions

Anything you already told the assistant is carried over — location, orientation and exposure are filled in for you. Change only what is still missing, then open the evidence behind any number.

Live decision engine

Describe the exposure

Deterministic building-physics model. No waiting, no guesswork — the ranking updates as you move a control.

Climate

Orientation

South — peak UV and thermal cycling

Exposure

Surface treatment

What matters most

Expert mode

Show raw scores, trace values and the calculation inputs.

Model inputs: Köppen zone Cfb, driving-rain index 92, salt load 88, freeze-thaw 46.

Recommended · Coastal Scotland

Siberian Larch

In plain terms

In this climate, expect roughly 29 years of service — likely between 9 and 53 — with maintenance about every 10 years. Full silvering takes around 34 months.

Modelled estimate under method WI-SL v1.2.0, not a measurement. The range is the modelled spread: half of the simulated cases fall on either side of 29 years.

Workable with detailing discipline. Salt governs and needs an explicit response.

Behind this numberPercentiles, uncertainty per parameter and confidence · 1.2.0

Monte Carlo propagation · 1,200 draws

P90−P10 = 152.4% of median
OutputP10P50P90
Service life(yrs)8.928.852.8
Maintenance interval(yrs)8.610.412.2
Time to silvering(mo)263341
Performance score88.196.5100

Empirical percentiles from a seeded run of the full model (E11–E13). Every draw perturbs the climate drivers, orientation and exposure modifiers, species resistance profile, treatment credit, reference life and model residual at their declared 1σ.

Where the uncertainty comes from

Temperate (C) · σ ×1
  • Model residual88.8% · ±15.22 y
  • Reference service life10.9% · ±5.35 y
  • Climate drivers (ERA5)0.1% · ±0.61 y
  • Orientation & exposure0.1% · ±0.45 y
  • Species resistance profile0.1% · ±0.44 y
  • Climate-zone reliability0% · ±0 y
  • Surface-treatment credit0% · ±0.18 y

First-order variance shares of modelled service life, one parameter group varied at a time. Climate-zone reliability for Temperate (C): densest observation network; reference case for the model. Effective climate σ — uv ±8%, rain ±14%, humidity ±7%, salt ±22%, freezeThaw ±12%. Interaction terms are not captured.

Confidence · Moderate

74%

Basis: EN 350 class DC 3 (moderately durable) · EN 335 use class 3.1 · ERA5-derived climate load · 1200-draw Monte Carlo, Temperate (C) zone σ ×1 · WI-CAL v1.1.0 zone factor ×0.70 · assumed EN 927-1 thin-film system (not declared). Ranges are modelled, not measured — confirm with project-specific exposure data before tender.

Turn this into a specificationCompare the top two

The report is a self-contained, print-ready document for the project file: brief, ranking, load–resistance balance, uncertainty breakdown and detailing flags. The JSON export carries every input, intermediate value, equation and parameter — enough to recompute this result independently. Method: WI-SL v1.2.0.

Deterministic clause generator

Tender-ready clause text assembled directly from the calculation — same numbers as the calculation record, no generated prose. Export as text or as a BIM property set.

years

EN 335 UC3.1–3.2 · deterministic, no AI

Target 30 years at P10Does not meet requirement

Modelled service life (P50 28.8 years, P10 8.9 years) falls short of the declared 30-year design life. Change species or treatment strategy, improve exposure protection, or reduce the required design life.

1 · Scope

  1. 1.1External timber cladding to the project, Coastal Scotland (Köppen Cfb — temperate oceanic).
  2. 1.2Elevation orientation: South — peak UV and thermal cycling. Exposure: moderate.
  3. 1.3Service-life expectations in this section are modelled values from WI-SL v1.2.0, calibrated per WI-CAL v1.1.0. They are design guidance, not a warranty.

2 · Species and grade

  1. 2.1Specified species: Siberian Larch (Larix sibirica), Siberia, Northern Russia.
  2. 2.2Natural durability of heartwood: EN 350 durability class DC 3 (moderately durable). Sapwood shall be excluded from all exposed faces.
  3. 2.3Mean density 590 kg/m³ at 12 % moisture content. Dimensional stability: moderate.
  4. 2.4Permitted alternatives, subject to approval and equal detailing: Thermally Modified Pine (EN 350 class DC 2–3 after modification (process dependent)); Thermally Modified Ash (EN 350 class DC 1–2 after modification).

3 · Use class and durability requirement

  1. 3.1Use class to EN 335: UC3.1–3.2 — Above ground, intermittent wetting depending on detailing.
  2. 3.2Declared design life: 30 years, to be met at the P10 confidence level.
  3. 3.3Assessment against the requirement: Does not meet requirement. Modelled service life (P50 28.8 years, P10 8.9 years) falls short of the declared 30-year design life. Change species or treatment strategy, improve exposure protection, or reduce the required design life.
  4. 3.4Replacement strategy: Board-by-board replacement of failed elements.
  5. 3.5Modelled service life: P50 28.8 years, P10–P90 8.9–52.8 years (1200 Monte Carlo draws, model confidence moderate, 74/100).
  6. 3.6No design flags raised by the model for this combination.

4 · Surface treatment

  1. 4.1Timber shall receive a maintained surface treatment. Coating system to be tested and declared to EN 927-1/-2 for exterior wood coatings in the stated exposure category.
  2. 4.2First maintenance intervention is modelled at 10.4 years (P10–P90 8.6–12.2 years) for this elevation.
  3. 4.3Where an even silver-grey appearance and a low-preparation maintenance cycle are required, a non-film-forming mineralising system class (waterborne silicate/silane surface treatment) is documented in the evidence register: the treated surface cannot peel or flake, so re-application is inspection-led and does not require sanding back, and verified EPDs to ISO 14025 / EN 15804+A2 are available for this class. Reaction to fire up to B-s1,d0 is documented for such systems applied over fire-impregnated timber; the class shall always be quoted with the tested substrate. Decay protection shall not be attributed to this class — durability shall be achieved by species, use class or impregnation.
  4. 4.4Treatment selection is performance-based. No proprietary product is named in this specification; any system meeting the declared performance and maintenance interval is acceptable.

5 · Moisture control and detailing

  1. 5.1Detailing shall follow EN 1995-1-1 principles for protection against moisture: ventilated cavity behind the cladding, drained and back-ventilated joints, end-grain sealed or shielded, and no horizontal upward-facing surfaces without a fall.
  2. 5.2Local climate loads used in the assessment: driving rain 92/100, relative humidity 86/100, airborne chloride 88/100, UV 28/100, freeze-thaw 46/100.
  3. 5.3Wind-driven rain and chloride load dominate. Biological growth on north elevations within 24 months.
  4. 5.4Minimum 150 mm clearance from finished ground level and from any snow-retaining or splash surface.

6 · Inspection and maintenance

  1. 6.1Inspect annually for the first three years, then at not more than five-year intervals.
  2. 6.2Record moisture content at the plinth and at the most exposed joint; sustained readings above 20 % indicate a detailing failure and shall be reported before any recoating.
  3. 6.3Plan the first maintenance cycle within the modelled 8.6–12.2 year window rather than on visual failure.

7 · References and traceability

  1. 7.1EN 350:2016 — Durability classes of heartwood against wood-destroying fungi (Annex B / Table B.1)
  2. 7.2EN 335:2013 — Use classes for biological exposure (Clause 5)
  3. 7.3EN 350:2016 — Performance-based durability and modified wood (Clause 4 / Annex C)
  4. 7.4EN 927-1:2013 — Classification of exterior wood coating systems (Clause 4)
  5. 7.5EN 927-3:2019 — Natural weathering test for coating systems (Clause 6–8)
  6. 7.6EN 1995-1-1 (Eurocode 5):2004 — Service classes and modification factor kmod (Clause 2.3.1.3 / Table 3.1)
  7. 7.7EN 351-1:2007 — Penetration and retention classes for preservative-treated timber (Clause 5 / Annex A)
  8. 7.8EN 14915:2013 — Solid wood panelling and cladding — harmonised product standard (Clause 4–5)
  9. 7.9Assessment engine: WI-SL v1.2.0 with WI-CAL v1.1.0 bias correction. Method, equations and validation are published at woodintel.io/method and woodintel.io/validation.
  10. 7.10WoodIntel is funded by SiOO:X Wood Protection and is brand-neutral: no supplier influences the ranking or this clause text. This document names no product and recommends no supplier, including the funder — treatments are specified by generic system class and performance criteria only.

AI project briefOpen +

For free-form project descriptions where the clause generator inputs are not enough.

Typically 30–60 seconds of reasoning.