Artificial Intelligence and eCommerce

The Artificial Intelligence (AI) market is booming and constantly evolving. It was worth 428 billion US dollars in 2022¹.

To remain competitive, eCommerce businesses have to invest in this market. 91% of eBusinesses¹ have already embraced it and for over 50% of them¹ it has increased their productivity.

Discover the definition of Artificial Intelligence (AI) and how it can be used in the eCommerce industry.

AI: definition

Artificial Intelligence (AI) is a range of technologies and techniques designed to carry out tasks by imitating human intelligence.

AI improves the eCommerce performance of businesses in many different areas: customization (70%²), internal search engine (54%²), forecasting (52%²), marketing (43%²)… It also saves time by automating employee activities which traditionally occupy them for 60% to 70%³ of their working days.

To take it another step further and get the most from AI's potential in the eCommerce industry, PIM (Product Information Management) solutions represent a true competitive advantage. Some PIM solutions already use AI allowing you to automate numerous tasks related to product information: translation, writing, adaptation to each selling channel, customization…

Using AI in eCommerce businesses

Customized product recommendations

Product recommendations are a « powerful lever » for eCommerce businesses. To be visible to shoppers, a business should place them in different areas of their eCommerce site: home page, product page, onsite results pages…

To generate these recommendations, they can use AI based recommendation engines which analyze large volumes of customer data including browsing and shopping history, preferences, wish lists… and use this data to predict the products most likely to interest shoppers.

By using AI in this way businesses can optimize the conversion rate⁴ of their eCommerce site and increase the average value of orders placed by 26%⁴.

Dynamic pricing

On an eCommerce site, it is essential to present pertinent prices to shoppers. Around 45% of Gen Z shoppers⁵ browse several eCommerce sites before finding « the best online deal ».

For a business, this means adjusting prices in line with the price variations of your competitors: Amazon's prices change 2.5 million times a day⁶ and Walmart's change around 50,000 time per month⁶.

Meeting this challenge can involve the use of AI based dynamic pricing tools which help businesses to automatically adapt their prices according to various factors: competitor prices, shopping behavior, discount strategies…

Dynamic pricing tools are increasingly appealing to businesses. 22% of businesses⁷ are not yet equipped and 17%⁷ have plans to implement a dynamic pricing strategy over the coming years.

Generating product descriptions

ChatGPT: write your product descriptions efficiently

The product description is a key part of any eCommerce site, and is considered as the most important element⁹ in the decision making process. To persuade a shopper to buy your product, the description has to be extremely well written as inaccuracies in product datasheets lead to 20% of lost sales opportunities¹⁰.

Businesses can save time by using a Natural Language Processing (NLP) such as ChatGPT which is capable of interpreting, generating, and translating human language. In an eCommerce context, it can analyze the product information provided by a user (technical characteristics, sales pitches, certifications, and labels…) and then use this information to generate consistent product descriptions.

With this type of tool, businesses can increase the efficiency of their daily tasks, in particular in terms of content writing. 37% of businesses¹¹ are using it for this purpose.

ChatGPT: capacity and limitations

One of ChatGPT's main features is its powerfulness. For example, ChatGPT has been trained on 175 billion parameters¹². When it was launched, the platform attracted 1 million users¹³ in just four days and 100 million users in two months¹³.

To use it in an eCommerce context, ChatGPT should be asked to complete specific tasks: « write a product description », « generate ideas for a sales pitch », « identify technical features » … and you should provide as many details as possible (number of words, information about the product, tone…). After the texts have been generated, it is important to have them checked, corrected, and completed by a human before publishing.

You should be aware that ChatGPT has its limits and cannot produce the same quality as an « expert » in a given topic. ChatGPT also sometimes « hallucinates » and generates rubbish and has even been known to present its own inventions as the truth. OpenAI is currently being charged with « libel¹⁴ » after creating false content.

Translation of product information

On an eCommerce site it is important to present product information in the right language. 65% of shoppers¹⁵ want to find content in their own language and 40%¹⁵ will not buy from a website that is not written in their language.

To translate this information, a business can use translation services based on « Neuronal Machine Translation » (NMT), which is the most powerful automatic translation system. Some of the principal automatic neuronal translation services¹⁶ are DeepL, Google Translate, Microsoft translator…

Using these services can greatly increase the efficiency of the product information translation process: process a large volume of data, guarantee consistency between each translation, large choice of languages… For example, Google translates 140 billion words every day, into 103 different languages¹⁸. Using these services also increases the quality compared to traditional translation services: 50% less errors¹⁸ in the word order, 17% fewer lexical errors¹⁸, 19% of grammatical errors avoided¹⁸…

Despite this increased efficiency and quality, businesses can encounter difficulties in processing certain « subtleties¹⁸ » such as ambiguity in the source text or the presence of certain highly specific or technical terms (legal, medical, insurance topics…). The solution to this is to use professional translators and/or translation agencies to correct misinterpretations, improve sentence structure, proofread and fact check…

Artificial Intelligence has a lot to give eCommerce businesses, in particular in the production, generation and translation of product related information: prices, recommendations, descriptions… However, to get the most from AI you need to fully understand its capabilities and limitations and to view it as an « assistance for humans ».

¹ Liberge, A. (2023, May 21). Intelligence Artificielle : 10 chiffres clés à connaître. Oberlo.
² Statista. (2023, May 15). E-commerce areas where AI is useful according to executives.
³ The economic potential of generative AI : The next productivity frontier. (2023, June 14)
⁴ Rajnerowicz, K. (2023). 99 Customer Experience Statistics & Interesting CX Trends [2023]. Tidio.
⁵ Qu'est-ce qu'un moteur de recommandation de produits pour la vente au détail, et pourquoi en avez-vous besoin ? (2021, June 21). Salesforce.
⁶ Mas, R. (2019, June 21). L'IA, un outil indispensable à la tarification dynamique. Silicon.
⁷ Shpanya, A. Why dynamic pricing is a must for ecommerce retailers.
⁸ Statista. (2023, March 15). E-commerce companies planning on using dynamic pricing 2021.
⁹ E-commerce : 10 chiffres qui montrent l'importance d'une bonne stratégie de contenu. (2017, January 19). Frenchweb.
¹⁰ Writing Product Descriptions : 10 Tips To Persuade in 2023. (2022, Octobre 4). Shopify.
¹¹ IBM Global AI Adoption Index 2022. (2022, May). Shopify.
¹² Alric, J. (2023). ChatGPT va bientôt devenir 570 fois plus puissant, personne n'est prêt. Presse-citron.
¹³ Cromwell, J. R., Harvey, J.-F., Haase, J., & Gardner, H. K. (2023, June 9). Discovering Where ChatGPT Can Create Value for Your Company. Harvard Business Review.
¹⁴ OpenAI poursuivie en diffamation en raison des hallucinations de ChatGPT. (2023, June 16). Programmez !
¹⁵ Keenan, M. (2022, November 24). Global Ecommerce : Stats and Trends to Watch (2023). Shopify Plus.
¹⁶ Définissez la qualité de vos traductions avec Weglot.
¹⁷ Demichelis, R. (2018, May 29). Comment l'IA révolutionne la traduction. Les Echos.
¹⁸ Zameo, S. (2019, May 15). Traduction Automatique Neuronale : conseils et bénéfices pour vos traductions digitales. Textmaster.