How AI is transforming product experiences

This article is provided by BRC Associate Member, Valcon.

__________________________

 

Ten years on from the development of digital product information management (PIM) systems, many retailers are still facing the same challenges when it comes to showcasing high quality, accurate product information. These data anomalies can have wide reaching implications, which can impact customer satisfaction, loyalty and the retailer’s reputation. 

So why are retailers still facing such challenges despite the digitisation of product information? Common causes include ongoing legacy system issues, an over dependence on Excel spreadsheets and the complexity of an eco-system that involves third-party suppliers, factors which conspire to erode the quality of the data that PIM systems depend on. 

Product information challenges retailers face

Problems with product information can be wide reaching. Inconsistent data can result in clothing being inaccurately described, FMCG goods being incorrectly labelled or electronic goods having the wrong specifications. This can have significant consequences for retailers. 

It is a sliding scale of risk. At the base level it will result in customers ordering the wrong product, causing frustration for the customer and costly returns and lost sales for the retailer. At the more serious end, if products for consumption are incorrectly labelled, it can cause medical problems. Natasha’s Law - formally known as the Food Information Regulations 2019 - came into force in 2021 when Pret a Manger customer Natasha Ednan-Laperouse had a fatal allergic reaction to an ingredient in a sandwich. It is a European-wide regulation that aims to improve food safety, empower consumers with the information needed to make informed choices (including traceability of ingredients) and holds food retailers and manufacturers to account for food labelling. 

Whether the risk is major or minor, product data issues can have a significant impact for retailers. Frustrated customers, costly returns and lost sales at a minimum and reputational damage, a hit to revenue and profits, a loss of market share and causing customers physical harm at the other end of the scale. So in a world where customers expect correctly labelled products and a seamless shopping experience, bad product data can be a dealbreaker, damaging the brand and impacting customer loyalty.

Regulatory hurdles
In addition to Natasha’s Law, there is another regulatory regime due to have a big impact on retailers. The digital product passport (DPP) – which will come into effect in 2026 - is one of the key actions under the Circular Economy Action Plan (CEAP) and Ecodesign for Sustainable Products Regulation (ESPR). It contains information on a product’s origin, materials, environmental impact and disposal recommendations and its aim is to improve transparency and sustainability across product value chains. The DPP system relies on advanced data technologies to securely store and verify product data. 

Supercharging PIM
In meeting these regulations, product information management (PIM) is vital in ensuring product data is structured, clean and consistent. It can reconcile data from multiple sources – ERP, e-commerce platforms and supplier feeds – which can all degrade the quality of product data. If product specifications do not match across channels, customers get confused and retailers could be in breach of regulatory regimes.

And when risk mitigation around product data is not just about maintaining customer satisfaction, it’s a regulatory imperative, it’s even more important they get it right. Strong PIM mitigates against customer frustration, reduces return rates and ensures retailers can stay compliant with product regulations. 

But if retailers are managing data reconciliation manually, errors can creep in fast. PIM ultimately ensures product reliability, but what if we can take that a few steps on?

How AI improves the product experience
Obviously, these days it’s all about artificial intelligence (AI). And where PIM is concerned, it’s no different. AI can transform PIM in all sorts of ways, anticipating issues, fixing problems before they even happen, making the customer experience smoother and the retailer compliant.  

Instead of manually cleaning product data, AI automates that process and keeps product data attributes consistent across all channels. Here are some key benefits:

Automated data enrichment 
This fills in missing information, ensuring product listings are complete. It automatically classifies products into categories and assigns attributes based on product descriptions. NLP (natural language processing) extracts relevant details from supplier data ensuring consistency across all uses. For example, if a supplier provides shoe size data in US sizes but the site sells in UK sizes, AI can automatically convert and populate the correct size mapping. It can also generate product descriptions and translate them into different languages, meaning customers get the size they expect, reducing customer frustration and the need for returns.

Image recognition for product accuracy 
AI can analyse product images and auto generate accurate descriptions. This makes sure there is a uniform product catalogue that doesn’t rely on manual entry. Machine learning algorithms analyse product images and compare them to descriptions to prevent mismatches. If an online listing says ‘red jacket’ but the image is pink, the technology flags discrepancies, which reduces incorrect listings and prevents those frustrating ‘this is not what I ordered’ experiences.

AI-driven error detection 
This functionality scans product data to identify inconsistencies, duplicate entries, or identify conflicting information. These tools suggest corrections instantly, ensuring data remains accurate and reducing manual fixes. It essentially spots potential problems and fixes them before they happen. 

Smart product tagging and attribute management
AI is proving extremely useful in smart product tagging and attribute management. It can identify product size, colour and material from images and descriptions, which cuts down on manual intervention. And automated keyword generation can enhance SEO and improve the discoverability of products on e-commerce platforms. 

Automated translation and localisation
If you’re a European or global retailer, using different languages for different markets is par for the course. AI can be useful in this scenario. AI powered translation tools ensure product descriptions and attributes are accurately described for global markets. AI can also adjust product recommendations based on regional preferences, trends and even weather patterns – what a customer might be interested in in chilly Helsinki will be different to a Barcelonan who is basking in 35 degrees. 

PIM AI in action
In AI parlance, it addresses ‘use cases’ within the business, so business challenges a retailer might be experiencing. For example, a global fashion retailer might sell thousands of new products each season. Manually tagging each product with attributes – like slim fit, petite, cotton etc – is really time consuming. Using AI (image recognition and machine learning, for example) results can include faster catalogue updates, improved SEO and discoverability and ultimately better conversion rates. It can also do things like ‘complete the outfit’, suggesting recommendations based on what a customer wants to buy. Or for a retailer selling across multiple marketplaces (like eBay or Amazon), AI can be used for multi-channel synchronisation - AI driven PIM software gives real time updates across all sales channels, and NLP scans product listings and auto corrects inconsistencies or any missing information. 

Bottom line: get product data right with AI-powered PIM
Product data is everything. Get it wrong, and customers lose trust, get hacked off and go and shop elsewhere. In a worst-case scenario, incorrect data can cause them harm. But get it right, and they have a seamless shopping experience – and a retailer will find it much easier to stay compliant. AI-powered PIM takes the guesswork out of managing product information and reduces errors and improves efficiency. If your business still relies on manual processes to manage product data, it is time to let AI do the heavy lifting. It is not just about making life easier for your teams - it is about making sure customers always get what they expect.

Boudewijn Brobbel is a Senior Data Principal at European consultancy, Valcon www.valcon.com