As a practising lighting and software designer, Foad Shafighi has built a new, AI-driven tool aimed at making luminaire specification easier. Here, he tells us all about the Universal Luminaire Cutsheet.

Lighting specification has always relied on trust.

A designer finds a product, studies the cutsheet, checks the photometry, reviews the mounting detail, confirms the driver and dimming options, and tries to make sure that all of this information survives the long journey from design intent to procurement, submittal review, installation, and commissioning.

When this works well, it is often because of a lot of manual effort, professional memory, and a fair amount of patience.

The issue is not that manufacturers fail to provide information. In many cases, they provide a great deal of it: PDF cutsheets, IES files, installation instructions, finish charts, driver data, wiring diagrams, BIM content, compliance documents, and product configurators.

The issue is that most of this information is still packaged for human reading, not for structured digital workflows. A PDF cutsheet is useful to a designer; it is not always useful to software.

That distinction is becoming more important as lighting specification moves into a more connected and AI-assisted future. Designers are being asked to compare products faster, build schedules more efficiently, review substitutions more carefully, and coordinate information across more platforms.

Manufacturers are trying to keep product data accurate across websites, rep channels, specification platforms, internal databases, and marketing material.

AI adds another layer to this conversation. There is real excitement around what AI can do for design and specification, but its usefulness depends heavily on the quality of the information underneath it. If the source data is incomplete, inconsistent, or difficult to verify, then the AI layer built on top of it will also be unreliable.

That is the problem ULC, the Universal Luminaire Cutsheet, sets out to address.

I came to this problem first as a practicing lighting designer who also designs software. Over the past year, while building better tools for designers, I kept running into the same obstacle: the product data those tools depend on was scattered, inconsistent, and hard to verify. So, I decided the data had to come first.

ULC is an open, machine-readable schema and proposed standard for translating lighting product information into a structured format while preserving traceability back to the original manufacturer documents. It does not replace the traditional cutsheet. The PDF still has value. It presents the product visually, communicates brand and application intent, and gives designers a familiar document to review. The role of ULC is different. It acts as a reliable, machine-readable data layer alongside the cutsheet.

A ULC record describes a luminaire’s product identity, electrical characteristics, optical and photometric information, physical properties, mounting conditions, controls, accessories, environmental ratings, and other specification-critical attributes. Just as importantly, each value is linked back to the source document it came from, with a content hash anyone can verify.

That traceability matters. If a schedule lists a wattage, a dimming protocol, an IP rating, or a CCT range, a designer should be able to understand where that information came from. Was it taken from the cutsheet? A driver document? A photometric file? A compliance sheet? A structured data layer should not only carry the answer but make it reviewable.

This is especially important in an AI-assisted workflow. We should give designers fast ways to verify a summary, not ask them to trust it blindly.

The larger question is not only “What can AI do?”

The better question may be, “What kind of information structure does the lighting industry need in order for AI to be useful, safe, and accountable?”

Today, lighting product data is often fragmented across multiple documents and formats. One manufacturer may communicate dimming behaviour in a matrix. Another may bury it in a catalogue code. Another may only clarify it in a separate driver sheet. Mounting conditions, optical accessories, emergency options, photometric performance, and controls compatibility can all be expressed differently from one product line to another.

Human beings can navigate this, but it takes time. Software can navigate it only if the information has been structured with enough consistency.

A shared standard would not eliminate the need for judgment. It would not tell a designer which luminaire is beautiful, appropriate, comfortable, or right for the atmosphere of a space. Those decisions remain human decisions.

But it can reduce the time spent hunting for basic information, make product comparison more transparent, and keep schedule-building cleaner. It can also flag missing or uncertain data before it becomes a coordination problem.

For manufacturers, it is also a visibility issue: product information that is easier to understand, compare, and verify makes the product itself easier to specify with confidence. For specifiers, the benefit is not speed alone but better confidence in the information behind a decision. For reps, it can mean fewer repetitive questions during product selection, value engineering, and submittal review. For software platforms, it offers a more reliable foundation than scraping PDFs, which is fragile and inconsistent.

Of course, a standard like this has to be practical. The lighting industry includes large global manufacturers with sophisticated data teams, but also smaller specialist manufacturers with limited digital infrastructure. Any useful standard needs to allow for different levels of completeness. It should help manufacturers begin with essential product information and improve over time, rather than making participation feel impossible from day one.

This is the direction ULC has taken. Rather than treating an incomplete record as a failure, it treats partial data as a valid floor and computes the grade it has reached: Core, Standard, or Full. Every record carries a roadmap of exactly what it still needs to reach the next grade, so a manufacturer can start with essential data and deepen it over time.

It also needs to be developed through dialogue. No single designer, manufacturer, rep, or software provider can define this alone. The value of a structured product data layer depends on whether it reflects the real needs of the people who create, distribute, specify, and review lighting products.

That is why I think of ULC as a working standard that will keep evolving through dialogue, not a finished, top-down answer. AI has made this more urgent, but the problem is not new; it simply exposes the weakness of our current information infrastructure.

If we want AI-assisted tools to support better lighting decisions, the data beneath those tools needs to be structured, traceable, and reviewable.

The future of lighting specification should not be a black box. It should be transparent enough for designers to trust, flexible enough for manufacturers to adopt, and useful enough to serve the many tools and workflows that already shape our industry.

The goal is not to make lighting less human, but to give the industry a better foundation, so human judgment can be applied where it matters most.

www.lightingagent.ai

www.ulcspec.org