Faster and More Efficient with AI

Summary:

PIM and AI optimize data management and automation, delivering better results faster and more efficiently. Y1 offers our vast experience and best practices to support companies effectively. Examples: Automated product categorization and personalized product recommendations. Our expertise ensures smooth implementation and maximum efficiency.

Images & Documents

Event Pictures

Brands that often organize events, attend trade fairs, or participate in exhibits yield a large number of images. In the past, these were manually sorted and indexed by assistants, meaning that images were only available and searchable weeks later.

Nowadays, images can be imported directly into the PIM system and automatically categorized from there. All images will be readily available come the next morning and no manual work needs to be done.

AI determines the following image properties:

– Are there people in the picture?

– Is the company logo recognizable in the image?

– Keywords for the image

– Subject

– Mood

– Time of year, season

  • Competitor logos or other company logos

Sorting and Indexing Historical Images

AI can be useful to companies with thousands of historical images that need to be sorted and categorized.

AI can be trained to analyze and parameterize what should be recognized in images in about a week’s time. After that, it should take exactly one more week for all images to be indexed, sorted, and grouped suitably.

Retrieving Metadata from Documents

Nowadays, it is no longer necessary to laboriously maintain the number of pages in a document, which products can be seen on it, or which languages are available in the document. AI can automate this whole process.

Correct Images

Different images can be displayed and sorted based on time of year, target group, or other criteria. AI analyzes images to determine which image is more apt to be displayed more during the summer or for women.

COLORS

Identifying Webshop Color

Back then, companies had to cross-check marketing colors to determine which color a product should be displayed as on the store. While this is usually easy, it required manual work for each product.

With AI, marketing color names are sent directly to the AI to reliably retrieve the desired color. If the AI is unsure, it will make suggestions and point out mappings that should be manually reviewed.

Alternatively, colors can also be extracted from assigned images. This is particularly helpful if the product comes with multi-colored components, e.g., sole color and surface color for shoes.

Texts

Generating Texts

In the past, you had to provide manually written text for each product.

Today, AI can receive a reference text for reference and the parameters to generate the text, within just a few seconds.

Shorten Texts

In the past, long descriptions for certain target systems had to be manually shortened. Each change would also had to be manually made for all text variants.

Using AI, you can generate required texts at any length from longer texts.

Optimize Texts for SEO

Back then, long-form texts had to be manually SEO-optimized, and this had to be done differently for various marketplaces. This is no longer necessary today—firstly, because there are certain predefined parameters for each marketplace, and secondly, you can now define which keywords to add to texts for each category. You can even use AI to automatically generate keywords and optimize texts for marketplaces.

Customize Titles

On your own store, the brand name doesn’t have to appear before the product name, but it does on marketplaces. AI can do this automatically for you.

Synchronize Title Spelling

There used to be different spellings for products within a category. Now, there can be a reference product in each category and the titles of the other products are automatically adapted to this reference product.

Standardize Titles

Certain target systems also have technical specifications as to how titles may look (length, special characters, abbreviations). In the past, these rules had to be laboriously programmed. Today, all you have to do is pass the rules and parameters for the target system to the AI and you get the adapted title back immediately.

Technical Data in Texts

In the past, it was important to avoid placing technical data in continuous text, as this was often done manually and resulted in maintenance errors. Today, technical data is automatically integrated into the body text and the text is simply updated automatically if the technical data changes.

TRANSLATION

Translation into Other Languages

With AI technologies such as neural networks and machine learning algorithms, language barriers will soon become a thing of the past. Unlike traditional methods, AI takes a context-aware approach to translation and learns over time

Standardize Titles

In addition to languages, texts can now also be displayed in language variants or dialects such as Swiss German, Flemish, Central American Spanish, etc. at minimal cost and effort.

REVIEWS

Interpreting Negative Reviews

If a product has many negative reviews, e.g. under 3.5 stars, AI can highlight the three most important negative aspects from a certain period. These are then relayed to the product manager in the PIM so they can check whether texts should be adjusted or technical data needs to be corrected. It is also possible to show them exactly which attribute needs to be corrected.

Interpreting Positive Reviews

Positive reviews can also be taken and mapped against the existing product description. If there are no positive comments, the product manager will be notified he can decide to change the texts.

Top Testimonials

It is also possible to automatically interpret the reviews that best reflect the product USPs or the product description from the best reviews during a specific period. These testimonials are then displayed in the store as statements on the product.

TECHNICAL DATA

Class Assignment

In the past, a mapping had to be done manually—from your own products in the product category of the merchandise management system to the product maintenance tree in the PIM. Today, this is automatically determined by the AI and similar solutions. The same applies to purchased products where only the supplier classification is known.

Attribute Mapping

Attribute mapping, previously often solved with scripts or complex transformations, can now also be handled by AI.

DIN Standards

References to DIN standards are often found in product texts. These can be analyzed and the correct DIN standards are maintained in their own attributes or even as a reference to the correct product.

Scanning Labels

In the past, the information on labels often had to be laboriously typed out for merchandise. Today, a photo is taken and the information is automatically read and written to the product.

CROSS-SELL & UP-SELL

Cross-Sell Categories

AI can determine what the best cross-sell categories are for a product and then display the corresponding products.

Matching Products

The criteria for the matching products for cross-selling can also be color, similar quality or price categories.

Audience-Specific Cross-Selling

If you know the target group, you can adapt the texts or display other images as well as re-sort the cross-sell. The AI knows the preferences of different target groups, or you can enter these yourself using parameters.

Upselling

In the past, upsell products had to be manually selected and maintained. Today, AI can add suitable products according to defined parameters.

GOT QUESTIONS? FEEL FREE TO ASK!

Each company has unique requirements, prerequisites, and data, which is why every PIM project must be approached individually. On behalf of Y1, I can offer you a customized solution and personally advise you. Get in touch by arranging an initial consultation.

Marc Kulow berät zu PIM-Lösungen, egal ob PIM-Integration, PIM-Upgrade oder allgemeines PIM-Projektmanagement.