Short answer: Use AR to show options that are difficult to keep on the floor—a color, accessory location, trim combination or guided feature explanation—when the digital content is accurate and clearly labeled. Use a real vehicle for questions about seat fit, touch, sightlines and physical controls; use an appropriate drive for questions about driving behavior. Ask each buyer question first, then choose the medium that can answer it.
An Eventagrate article about AR car showrooms presents virtual expansion and customization as benefits. Its numerical claims about costs, satisfaction and sales are not evidence for an individual dealership decision unless the underlying data, methods and applicability are checked. A 2017 SAE technical-paper abstract describes a tabletop prototype with a tablet and AR glasses for a self-guided feature tour, including a rendered accessory. It establishes a demonstrated design approach, not measured sales lift or a replacement for inspecting a vehicle. The full paid paper was not reviewed for this article.
Allocate each question to the medium that can answer it
| Buyer asks… | Good first medium | What must be checked |
|---|---|---|
| “Where would this roof rack sit?” | AR overlay on a recognized vehicle or accurate scale model. | The accessory, mount points and model year match the offer; tracking stays aligned as the viewer moves. |
| “How do these two trims look?” | Side-by-side digital views, optionally overlaid near a physical car. | Images, labels, availability and price come from current approved product data. |
| “Will this seat support me?” | Sit in the actual relevant seat. | The physical seat and adjustment range match the trim being discussed. |
| “Can I see the road from this position?” | Physical seat and, where appropriate, a drive. | The buyer can adjust the real position and inspect true sightlines. |
| “How does it accelerate or brake?” | A safe, authorized test drive or verified technical information. | An animation is not represented as a felt driving result. |
This matrix is a proposed decision aid, not an experimental finding. The medium may change if the specific AR system is inaccurate or the physical vehicle is a different variant. The salesperson should be able to explain that distinction without interrupting the buyer's exploration.
Prototype one difficult choice
Do not start by rendering the whole catalogue. Choose one question that the floor inventory cannot answer well—for example, the position of two accessory options on a car that is physically present. Create a small AR flow with the exact model and option codes. Keep a non-AR route, such as a labeled comparison image, so the team can test whether the overlay adds understanding rather than just novelty.
Ask prospective buyers to find and explain the difference, then ask what they still need to see in person. Observe tracking breaks, glare, device handling, whether people confuse an illustrative rendering with a sold configuration, and whether staff can recover from a failed launch. Capture which question was answered before asking whether the experience was enjoyable. No result from a prototype should be reported as proof of purchase intent or sales conversion without an appropriate study.
Make the data handoff as carefully as the visual
AR content must name the model, year, market, trim and option set it depicts. If availability or price is dynamic, the approved source and refresh process matter as much as the image. A stale overlay can be persuasive and wrong at the same time. The team needs an owner for changes, a date on the content package, and an operator route to remove an outdated option. If the digital image exaggerates a finish, mark it as illustrative and bring the visitor back to a physical sample.
Archv's general guide to using AR at events asks whether AR earns its place in an activation. In a car showroom, the narrower question is how to route each vehicle decision to evidence the buyer can actually inspect. AR can help a buyer navigate options; the final handoff should make clear what is depicted, what is confirmed, and what still needs a real-world check.