A large dresses dispenser uses AI for its electronic commerce operations

• Amarra is a manufacturer and wholesaler based in the New Jersey of special wise dresses.
• Over the past five years, he has used AI to help inventory management and content creation.
• This article is part of “how AI changes everything: small business”, a series exploring how small businesses use AI to succeed.

In the constantly evolving electronic commerce landscape, companies must find a more efficient, competitive and agile way to be. And not surprisingly, some small businesses turn to AI to get help.

Amarra, a company that makes and wholesale of special used dresses and dresses – including for weddings and Quinceañeras – has integrated AI tools in its operations since 2020. Artificial intelligence has helped business managers better understand customer needs and increase efficiency through tasks such as writing products and investment management, Kunal Madan Amarra Amarra, Coominder of Amarra, said Caoder.

“The main motivation was to improve efficiency and raise customer experience,” said Madan in an e-mail interview. “Electronic commerce is a highly competitive area, and AI shooting has allowed us to optimize how we manage stocks, create content and interact with our customers – which has considerably improved our overall performance.”

The company, whose headquarters is in New Jersey and basically of its products to more than 800 independent retailers worldwide, uses AI to optimize its marketing and customer service operations and analyze market trends. Amarra launched in 2018 with only two employees, up to 25 today.

Gather electronic commerce and AI

Amarra uses Chatgpt, a generative-a assistant, to write product descriptions. The team enters information on product features and unique sales arguments in the chatbot, which then produces an initial project of each presentation text. From there, the team refines the text to align themselves with their brand voice.

Its website presents dresses by color, silhouette or used, each with a multissense description designed to bring a retailer to buy them. The terms for a pink neon dress style dress include “fanciful” and “inspired by fairy”. Other kids refer to floral applications and stone accents – all written with the help of chatgpt.

The use of Amarra chatgpt, said Madan, has reduced content creation time by 60%, without loss of quality.

The company also uses a stock management system powered by AI which predicts action needs according to historical data and seasonal trends. The tool has reduced the 40%overtocking, said Madan, and helped Amarra to avoid selling popular items. This allows the company to ensure the availability of products for bridal stores, shops and other independent retailers in North America, Europe, the United Kingdom and Australia which buy its dresses.

Amarra also uses AI to analyze customer exam data on all platforms, which helps it better understand changes in purchase behavior and consumer preferences.

“This helps us to identify emerging trends and product improvement opportunities, keeping us in advance on the market,” said Madan.

AI also directs Amarra customer service operations. The company said that AI chatbots are now managing 70% of customer information requests, answering basic or commonly asked questions and providing faster answers. Madan said it had released the Amarra team to focus on more complex problems, such as product personalization requests and complex customer payment problems.

Use of AI to complete, not to replace

Although IA -centered innovations have enabled Amarra to rationalize its operations, the adoption of technology has not been entirely fluid.

Madan said one of his team’s greatest challenges in the use of AI had “been to achieve the right balance between automation and human touch”. For example, the Amarra chatbot initially provided rigid and impersonal responses to customers.

“Since then, we have refined them to better reflect the personality of our brand, improving the global customer experience,” he said.

The merger of AI systems with the existing tools and processes of Amarra and the management of biases in AI models has been two other challenges. For example, the company’s chatbots did not take into account cultural shades on different markets, leading to occasional erroneous interpretations.

He also experienced biases in product recommendations, such as AI promoting certain elements without considering more recent inventories. Madan said “constant adjustments” helped to alleviate these problems. But it is a continuous process to resolve biases and transparently integrate the technological environment of Amarra while keeping the data secure.

To guarantee responsible use of AI, the company actively implies its employees in the implementation and personalization of tools. Madan said collaboration has helped Amarra to maintain a balance between AI automation and human surveillance. The company also regularly trains its workers to effectively use AI in their daily work, he said.

“We focus on using AI to increase the capacities of our employees, rather than replacing their roles,” he said. “The AI ​​manages repetitive tasks, allowing our team to invest more time in creative and strategic functions that require a human touch.”

Amarra learned a lot because he tested and adapted his use of AI during the last half-receiver. Madan’s advice for other small businesses that embark on a similar course are to focus on good data management practices, because data is the basis of AI models and algorithms.

They should also consider starting with very specific and manageable areas where AI can have the most impact, such as customer support or inventory management, he added. All functions must be flexible enough to adapt to the evolution of commercial and operational requests, while remaining faithful to the ethics of the company.

“Be ready for a learning curve,” he said. “Some initial hiccups are normal, but iterer based on feedback leads to better results.”