Inter Cars

The Market: Speed Wins

The automotive aftermarket is one of the most competitive B2B markets in Europe. Distributors compete not just on price and range, but on how fast a mechanic can find and order the right part.

Workshop technicians typically have two or three supplier websites open at the same time. They search in parallel, comparing availability, delivery time, and price. The supplier who returns the right result first is in the best position to win the order.

Speed matters beyond the search bar. Inter Cars has built one of the most efficient delivery networks in the European aftermarket, designed to get parts to workshops within minutes. That network is essential because every car on a lift blocks the next repair. A car waiting for a part costs the workshop the capacity of this lift.

But even the fastest logistics chain starts with the correct order. If the technician identifies the wrong part, or takes too long to find the right one, the delivery network cannot compensate. Search speed and search accuracy are where the chain begins.

This pressure is growing. Online platforms from Asia and global e-commerce players are entering the European aftermarket with broad catalogues and aggressive pricing. At the same time, vehicle complexity is increasing. More models, more variants, more specialised components.

For a distributor like Inter Cars, search experience is a direct competitive advantage. The faster a mechanic identifies and orders the right part, the sooner the car leaves the lift, and the more likely Inter Cars wins the next order too.

The Challenge: Identifying Parts in the Real World

Inter Cars already had a strong digital platform with advanced filters, VIN decoders, and a well-designed user interface. But even the best filter system hits a limit when a mechanic holds a physical part and needs to match it to a digital catalogue entry.

The limits of keyword search. Some parts are worn, dirty, or missing their label. But even clearly visible parts can be hard to describe with words. Truck rear lights, for example, all share the same name, are not vehicle or trailer specific, but look completely different. Keywords alone cannot distinguish them. One can’t find them by vehicle too. And for parts that do carry printed text, article numbers, or barcodes, visual search reads that information faster than any mechanic can type it. The result: a better, faster search experience across the board, not just for difficult cases.

Time pressure. Every minute spent searching is a minute not spent repairing. Workshops run on tight schedules. A car waiting for the right part blocks the lift, and a blocked lift means the next customer waits too. Ordering the wrong part costs even more: return logistics, a second delivery cycle, and another day of lost workshop capacity.

Inter Cars leads the market by excelling at every step of the chain: catalogue quality, search experience, and delivery speed. Most parts can be found through filters and keyword search. But there are cases where a technician simply cannot find the product, or where searching through old paper catalogues and their modern PDF alternatives is too slow. Visual search was the next step to cover these gaps and give workshop customers yet another edge.

The Solution: nyris Hybrid Search

Inter Cars integrated nyris Visual Search into its commerce platforms via the nyris visual search API. All Inter Cars B2B customers can now take a photo with their smartphone, and the system identifies the part within milliseconds.

But automotive parts are among the hardest product categories for visual search. Here is why, and how nyris solved it.

Hybrid search: vision, text, and context combined

Workshop users do not just photograph bare parts. They photograph oil cans, packaging, shelf labels, items with barcodes, article numbers, and printed text. A pure visual search engine would miss the valuable text information in these images.

nyris uses hybrid search to handle this. The system analyses every image and decides automatically:

• Should it search visually for the shape and appearance of the part?
• Should it extract and search by text, barcodes, or article numbers found in the image?
• Should it combine visual and text results, or rely on one signal alone?

The user does not make this decision. The user takes a photo and expects the right result. Intelligence sits in the system.

This also motivates Inter Cars to further invest in cataloguing their data in ways that support visual and hybrid search patterns even more. Well-structured product data with consistent images, clean article numbers, and accurate metadata makes visual search significantly more effective. In practice, visual search can act as a catalyst for broader catalogue quality initiatives that improve search across the board.

A large reference database with highly variable data quality

The nyris system indexes over several million searchable items backed by millions of reference images. The data quality across this catalogue varies widely: some products have multiple clean studio photos, others only have a single image with a cluttered background. Building and maintaining reliable image coverage across this range required close and ongoing collaboration between Inter Cars and nyris.

VIN-aware search

A photo alone narrows the search. Combined with a Vehicle Identification Number (VIN), the system narrows it further. Search without VIN returns relevant results. Search with VIN returns precise results. This context layer improves accuracy for parts that look similar across vehicle models.

A dedicated recognition model

Standard visual search models struggle with automotive parts. Parts are often metallic, similar in shape, and covered in oil or rust. nyris developed a dedicated embedding model optimised for this domain to improve recognition accuracy across the full range of Inter Cars products.

The Results

Usage Doubles Every Year

After an initial setup and indexing phase in late 2023, visual search adoption at Inter Cars grew rapidly.

Monthly visual searches now surpass 100,000 and keep growing.

This is not a feature that users try once and forget. It is part of how workshops order parts every day. Faster identification leads to faster orders, shorter delivery cycles, and fewer cars blocking the lift overnight.

From the Field

The numbers tell one story. The field tells another.

"The image search has become a true hit in our catalogue. Many of our sales representatives proudly demonstrate it and consistently receive very positive feedback from our B2B customers. Recently, I experienced the true power of this feature myself. During a meeting with the owner of a large truck service centre in Sofia, I overheard a conversation about a specific oil funnel used for filling truck engines. They had a model they liked, wanted to buy more, but couldn't remember where the original had come from. They were about to start searching in household goods stores. I stepped in, took a photo of the funnel, and we found the exact product in our catalogue within seconds. They ordered it immediately for their next regular delivery. The point was not the eight euros for an oil funnel. It was the time saved for our business partner, which is worth far more."

Stefan Deyanov, Sales Director, Inter Cars Bulgaria

Why It Works

What makes this partnership effective is the combination of scale, technical depth, and ongoing collaboration.

Scale. Millions of searchable items and images. This is industrial-grade visual search, not a demo.

Technical depth. Hybrid search combining visual recognition, Text Recognition (OCR), barcode reading, and VIN context. Each signal alone is useful. Combined, they cover the full range of real-world search scenarios in a workshop.

Collaboration. Image coverage, data quality, and model performance require continuous joint effort. As Inter Cars grows its catalogue, nyris ensures search quality keeps pace.

Pushed by real-world demand. The sheer variety of images that mechanics and workshop managers send every day has pushed nyris to build the best possible visual search service. Technicians photograph rusted flanges, oil-stained housings, torn packaging, shelf labels, blurry barcodes, and parts they cannot name. Every one of these images is a test case. Meeting these expectations at scale forced nyris to sharpen every layer of the system: recognition models, OCR accuracy, hybrid decision logic, and image coverage. Inter Cars did not just adopt our technology. Their users raised the bar for what it must deliver.

Whether a technician photographs a rusted bracket in a Polish workshop or scans a barcode on packaging in a Romanian distribution hub, the system returns the right result.

Visual search is now part of how Inter Cars serves its customers every day.

‍