Product · Skills
Cortex ships knowing how to do LCA work — dataset lookup, compliance guidance, EPD audit, calculation, format validation, and document export. Skills cover the scenarios LCA practitioners encounter repeatedly, and the library grows as you use it.
What skills cover
Each skill covers a workflow LCA practitioners encounter often enough that re-explaining it session by session wastes time. The catalog is organized by what you are trying to accomplish.
Find carbon data for any material — single lookup in Chat or full BOM export to Excel. Search across 14 databases simultaneously, compare GWP values side by side, and get proxies with their bias direction documented when no direct match exists.
Generate a process-flow diagram for any LCA product system — unit processes, material and energy flows, system boundary, cut-off points. Outputs a self-contained file ready to drop into a report.
Score any dataset across the five Pedigree Matrix dimensions — reliability, completeness, temporal, geographic, and technical correlation — with evidence-based justification grounded in the source record, not estimates.
Navigate the EU Carbon Border Adjustment Mechanism: identify whether your products are in scope, understand which embedded emissions to calculate and report, apply the right default values, and check reporting requirements before a filing deadline.
Work through all 15 GHG Protocol Scope 3 categories with boundary scoping, calculation methods, and data collection templates for each. Clarifies which categories apply to your business and what counts as a material emission source.
Reference guide for construction product EPDs: the full EN 15804+A2 and PCR 2019:14 indicator set — GWP family, PERM/PERE/PENRM, waste and output-flow indicators — lifecycle module definitions (A1–D), and explanations of common scope confusions.
Check an EPD document for compliance before registration. Reads PDF or ILCD+EPD XML, runs checks against ISO 14025, EN 15804+A2, PCR 2019:14, and EU CPR 2024/3110. Every finding cites the normative clause and quotes the evidence from the document. Outputs a structured audit report with risk-graded findings.
Cortex drives your local openLCA instance directly — build product systems, connect foreground to background processes, run LCIA calculations across ReCiPe, CML, or EF 3.0, and read results back into the session. The model stays in your openLCA database. Cortex doesn't hold a copy.
Create and publish LCA background datasets in HiQ Editor. Covers the full authoring sequence: metadata, flow entry, mass-balance validation, DQI scoring, pre-submission checklist, and the covering note for the reviewer. Stops if mass balance fails or a result magnitude looks wrong.
Validate ILCD data packages against schema requirements and convert between ILCD and JSON-LD. Reports every Error and Warning by file location and fix suggestion. Documents every mapping choice before any field is changed — no silent drop.
Read, create, and edit spreadsheets — data cleaning, formula calculation, chart generation, and structured export. Used in BOM matching results, sensitivity analysis tables, and multi-scenario comparisons.
Read and extract content from PDF reports, merge or split documents, fill forms, and produce new PDF deliverables from session results.
Create and edit Word documents with tables of contents, headings, headers, and tracked changes. Useful for modeling scope documentation, assumption records, and audit-ready report sections.
Build and edit slide decks from session results — for stakeholder presentations, EPD summaries, and internal methodology reviews.
Where skills come from
Core LCA workflows ship with every installation and are active from the first session — dataset lookup, CBAM guidance, Scope 3, EN 15804 indicators, openLCA integration, data quality assessment, ILCD validation, and all document export formats. No setup required.
Skills contributed by practitioners, reviewed by HiQ, and available to every user. Cortex checks for updates on launch. Specialist skills — like EPD audit for Programme Operators or dataset authoring for HiQ Editor — are available to install when the workflow applies to your work.
Import a skill you built for a specific workflow, or let Cortex write one based on sessions you have already run. A workflow you have done twice is a pattern worth carrying forward — it becomes a starting point the next time the same task comes up.
Gets better as you use it
Most tools look the same on day 365 as they did on day one. The institutional knowledge a practitioner builds — which databases to check first for a particular material type, how to handle a specific allocation edge case, what a reasonable GWP range looks like for a given product category — stays in the practitioner's head, not in the platform.
Cortex takes a different position. When the same workflow shows up across multiple sessions, Cortex identifies the pattern and adds it to your skill library. The next time that workflow comes up, the session starts knowing what you've already worked out — the step sequence, the databases, the common decision points.
The threshold is intentional: one session could be a one-off. Two sessions establish a pattern worth carrying forward. The skill that gets written reflects your actual decisions, not a generic template.
Sharing with the community
Any skill you have built or refined — whether it handles a specific regional database combination, a particular allocation approach for co-product scenarios, or a compliance check for a jurisdiction not covered in the defaults — can be shared with the broader community.
Skills go through a review before appearing in the marketplace. Once approved, they are available to every Cortex user on next launch. A practitioner who develops a better approach to a recurring problem on Thursday can have it in front of a colleague's session by the following week — not via a forum post or a support ticket, but through the platform itself.
Before any skill leaves your machine, it is checked for credentials or internal URLs that should not travel with it. You decide what to change if anything is flagged. No automatic redaction.
The point
Most LCA software ships a fixed set of capabilities. The practitioner who has spent three years developing judgment about regional grid factors, allocation edge cases, and cross-database consistency — that judgment lives in their head. It does not accumulate in the platform.
The skill library is how Cortex captures it. Not as a black box the vendor controls, but as a readable, shareable, editable record of how experienced practitioners handle recurring problems.
The platform's LCA expertise is not owned by the vendor. It accumulates from actual practitioner work.
Cortex Cowork is available for macOS, Windows, and Linux. Pre-installed skills are active from the first session. Marketplace and custom skills install in one step.