For lighting manufacturers

Specifier discovery is shifting from PDF to AI. The manufacturers whose data is machine-readable are the ones AI surfaces, design tools consume, and designers ask for by name.

On this page

Where product discovery is going

Specifier discovery is moving from the designer opening your PDF to the designer asking an AI which luminaire fits. Machine-readable data is what surfaces in the answer; the rest depends on whatever the AI guessed off the PDF. The bottleneck isn't the AI. It's the data layer underneath.

Today's discovery loop, from the manufacturer's seat. Loss accumulates at the extraction and aggregator-interpretation hops.

What ULC is

ULC is the open, machine-readable standard for luminaire product data: the layer AI tools, design tools, and spec databases read alongside the PDF, IES, and LDT you already publish.

Before and after, from the manufacturer's seat. The PDF stays beautiful for human readers. The ULC record stays accurate for machine readers.

How to publish

A .ulc file is JSON, validated against an open schema, that references your existing PDF, IES, and LDT by SHA-256 hash instead of embedding them. Producing one is a two-step path: gather the documents you have, then run a converter. A record is graded on what it carries, never on a metric the fixture's form has no reason to include. That extends to product classes: exit signs and emergency luminaires grade against their own profiles, so a sign is never asked for the LM-79 photometry its spec sheet does not carry, and a battery-driven emergency luminaire is never asked for luminaire efficacy.

Gather the source documents. Collect the ones that apply to the product:

  • the marketing cutsheet (identity, mechanical, environmental, and the headline rated specs),
  • the LED-driver cutsheet, when the driver is on its own sheet (electrical detail, dimming protocol and method),
  • the IES (LM-63) photometric file (the candela distribution and the electrical anchors),
  • the EULUMDAT (LDT) file, if you publish one,
  • the installation instructions (the mounting, wiring, and recess geometry the cutsheet only summarizes),
  • the accredited test reports (LM-79, LM-80 with TM-21, LM-82, TM-30), which carry the measurement uncertainty, corrections, and depth that lift a record to the higher conformance grades,
  • the compliance documents (safety listings, sustainability and origin declarations).

Then choose a route.

From your PIM. If your data lives in Salsify, Akeneo, SAP, or a custom PIM, the structure already exists. The mapping guides at /docs translate your fields onto ULC's, usually a one-time integration that then runs inside your existing publishing pipeline.

From a spreadsheet. If your data lives in spreadsheets rather than a PIM, the ulc from-sheet converter turns it into validated records. The converter is part of the ULC command-line tool: download it for your platform from /downloads, fill in the open-source workbook template (one row per photometric scenario, which is one row per record, each pointing at its source files), then run ulc from-sheet from your terminal. It computes the schema structure, SHA-256 hashes, and dual units, builds the generated index (grading and stamping the conformance level and the Product Achievements view), then validates each record. Deterministic and offline.

The validator catches schema errors before you publish, and the reference CLI shows each record's roadmap: ulc validate prints the grade, the per-field list of what each grade up to Full needs (each entry naming the source document and the governing standard), and a one-line achievements summary; --verbose adds the non-gating enrichment roadmap and the per-theme achievements detail. You do not need every document to begin: a record carrying only its identity grades Incomplete and comes back with the roadmap to Core. The schema, examples, and mapping guides all live at /docs; how the grades work is at /docs/conformance, and what the achievements report is at /docs/achievements.

What changes

  • Found in AI-mediated discovery, with your attributes. When a designer asks an AI for a 2700K downlight under 40W, your products are in the answer with the CCT, CRI, optics, and IP rating you published, not the ones an extraction layer guessed.
  • A direct channel to the designer's tools. Their AI and design tools fetch your authoritative record from your infrastructure and parse it natively. The aggregator's interpretation stops being the data layer. Reps, distributors, and aggregator relationships stay exactly where they are.
  • Early-mover position in the standard your peers will adopt. Open, MIT-licensed, neutrally stewarded, in dialogue with DIAL, the IES, and the LIA. The first to publish are the ones the industry remembers as having defined the layer.
  • Your certifications become a machine-readable profile. The DLC, ENERGY STAR, Declare, DarkSky, and UL 924 qualifications you already hold are computed into a per-theme Product Achievements view and stamped into the record. Attach the certificate file and the theme reads documented, evidence included, instead of a claim a tool has to take on faith. Your RoHS, REACH, and Prop 65 declarations surface beside the themes as a machine-readable restricted-substances flag.

Take action

Pilot it. Publish one to three of your most-specified SKUs as .ulc files; one product line is enough to start. The example records show the shape, the validator catches the easy mistakes, and the conversations that follow shape what the standard becomes. The schema surface is additive-only across minor releases, and a breaking change requires a new major version, so a record you publish today stays valid.

Start at /docs, browse the schema and examples on GitHub, or read the designer track that drives the demand.

Learn more