Artificial Intelligence is the buzzword of the year with immense anticipation and excitement attached to it, but also often a fear of the unknown. So much so, that tech behemoths Facebook, IBM, Google's parent company Alphabet, Amazon, and Microsoft recently announced a partnership to discuss AI best practices.
While there are the science fiction-driven angles of AI, like robots, self-driving cars, Internet of Things, and augmented reality, there are also more practical applications that affect business owners every day, especially those working in the virtual customer service world of online retail.
Gartner predicts that by 2020, 85% of interaction between customers and retailers will be through artificial intelligence customer service programs. Brands are rushing to build out their customer service approaches leveraging AI to create accurate product catalogs, fine-tuned search capabilities, and truly personalized online experiences.
In a recent post on IBM's Commerce Blog, the company says artificial intelligence can boost online retail sales by analysing customer data, identifying purchasing trends and behaviors, and informing decisions about when to put items on sale and when to order new ones.
Increasingly Refined Search Capabilities
The company best known for search is Google. Google has been using AI to help refine their search algorithms for years, and eCommerce brands have realized the importance of an accurate search feature.
After creating a detailed catalogue of items, brands are learning to enhance their search features. As mentioned in an article by Kissmetrics, the search results page is at the heart of whether a user clicks or leaves. And with a good site search engine, you can (and should) customize it a great deal, depending on what your target audience is looking for.
"Companies integrating deep learning into their eCommerce site will drastically improve user's search capabilities," says the AI expert Akash Bhatia, cofounder and CEO of Infinite Analytics. "For example, a woman could take a picture of a dress that she likes, upload the photo into the search bar of an eCommerce site and, using AI, the site would immediately analyze the image, understand the patterns, fit, style, color, brand, and other attributes to identify the dress. Voila! That consumer is able to convert right away."
Other experts agree with Bhatia. Ryan BeMiller, inbound marketing expert focused on the ecommerce sector, writes, "Photos alone cannot be expected to provide a full understanding of the product. The array of products on display should have distinct and clear product descriptions. Moreover, great product descriptions can also eliminate the overhead of customer inquiries. Anticipate and answer customer questions in the content of your website and you'll reduce the number of support calls and emails received."
Truly Personalized Online Experiences
Machine learning is ideal for finding patterns and using those to either recognize, categorize or predict things. And traditional recommendation engines use business rules and logic and perhaps one or two sources of user data to provide personalization.
However, personalization is a multi-faceted problem and there are many factors that depend on what the user might be shopping for. There are subtle clues that AI can pick up on - are they shopping for a life-changing event? How has their taste in brands, style, color and sizes evolved? This requires a lot of data to be processed if brands want to accurately personalize their shoppers' experience.
"The worst thing you can do with a personalization tool is 'set it and forget it'," says Linda Bustos, cofounder of Edgacent, in a roundup of personalization tips for Strands Retail. "If you invest in a powerful personalization tool, be prepared to invest in the right people to manage it and apply strategic merchandising, and be ready to measure results and refine your strategies based on real data."
From Paper to Digital: Improvement of Product Cataloging
Catalogues have gone from hefty magazines sent monthly to your mailbox, to billions of items available at your fingertips. While managing the constant influx of products that thousands of vendors upload daily, broken and unstructured cataloguing has become one of the biggest problems in online retail today.
"As a result of the misinformation uploaded, the user experience gets hit badly," says Bhatia. "There are companies that try to use brute force to solve these issues, such as employing hundreds of low cost, low quality workers in developing countries to look at each product image and annotate the details. For small to medium businesses, this tactic is expensive, not scalable and not effective."
Alternately, business owners who implement AI and deep learning algorithms that identify key features in a product image have more enriched, accurate online catalogues.
eMarketer estimates eCommerce sales to top $27 trillion in 2020. Whether you're a freelancer creating your own online store via Etsy, a SMB bringing your brick-and-mortar offerings online, or a large brand outsourcing all over the world to meet demand, this opportunity in eCommerce is vast. To effectively reach your target consumer and shield from competition, business owners must learn to implement AI tools properly to survive.