
Give each layer a job
Ambient light provides a usable background for the room. Task light supports an activity such as reading, preparing food or repairing something small. Accent light gives emphasis to a surface or object. These labels describe functions, not fixed product categories: one wall light might support movement while also drawing attention to a textured wall.
An AI recommendation becomes easier to assess when every proposed source has an explicit job. Ask which activity would become harder if a fitting were removed. If the answer is unclear, the suggestion may be decorative repetition rather than a necessary layer. Good design can be simple, especially in a small room where daylight and one well-placed adjustable lamp already do much of the work.
Distinguish output from light on a surface
Lumens describe luminous flux: broadly, how much visible light a source emits. Lux describes illuminance, or light arriving at a surface. Neither tells the whole comfort story. A fitting with high output can leave a task poorly lit if its distribution, position or shielding is unsuitable. Conversely, a modest local source can be effective when placed correctly.
Ask the model whether a quoted number is a lamp’s output, a luminaire’s output or an estimated task-plane value. The distinction matters. Real calculations also depend on geometry, reflectance, maintenance assumptions and the product’s photometric data. Dividing room area into a catalogue lumen figure is only a rough starting exercise, not proof that a room will perform well.
Use white-light appearance deliberately
Correlated colour temperature, usually expressed in kelvin, describes the appearance of white light. Lower values generally look warmer; higher values generally look cooler. It is not a direct measure of brightness, quality or comfort. A warm lamp can be glaring, and a cooler lamp can be comfortable when its output, shielding and position suit the task.
The International Commission on Illumination, CIE, provides the technical vocabulary behind many lighting measurements. Colour rendering describes another issue: how objects appear under a source compared with a reference. CRI is a familiar summary, while the Illuminating Engineering Society’s ANSI/IES TM-30 method provides a more detailed framework. Neither is a promise that two lamps will look identical in your kitchen.
Protect the view, not just the worktop
Glare can arise from a bright source in the field of view or from a reflected image on a glossy surface. Shielding, source size, viewing angle and surrounding brightness all matter. Ask for likely glare locations from the sofa, the desk and the kitchen work position. A ceiling plan alone cannot answer those questions because people look in different directions.
The Unified Glare Rating, UGR, is used in appropriate indoor lighting assessment contexts; it is not a universal comfort score printed onto every room. A catalogue reference to UGR does not automatically validate a domestic arrangement. Likewise, IEC TR 61547-1 and the IEC TR 63158 measurement framework address aspects of temporal light effects, not a blanket guarantee that an LED product will suit every person.
Plan changes through the day
Make an observation sequence: daylight-supported work, overcast afternoon, evening activity and quiet circulation. The sources needed in each condition may differ. A dining pendant might remain low while a local reading lamp provides focused support. Separate switching or dimming can make this possible without lighting the whole room to its busiest task level.
Control compatibility needs verification. A dimmable LED and a dimmer are not automatically a good pair. DALI, developed within the IEC 62386 standards series, is a lighting-control protocol; the DALI Alliance maintains certification resources. Matter, maintained by the Connectivity Standards Alliance, is a separate interoperability framework. A product’s support for either does not by itself establish compatibility with your existing installation or every desired feature.
Make the evening scene testable
Rather than requesting “a cosy atmosphere,” describe what should remain visible and what should feel quieter. You might want to read comfortably, see another person’s face and walk to the door without switching on a bright central fitting. Ask the model to explain which source serves each need and how the settings would change.
Test adjustable positions with safe, existing portable lights before committing to an installation. Observe seated and standing views, shadows and reflections. Do not improvise mains wiring. Treat the trial as a way to refine the brief, then ask a qualified designer or installer to resolve the product, controls and installation details that an illustrative AI arrangement cannot verify.