The future of search with Artificial Intelligence
3 March 2020
Nowadays, search engines are no longer a simple index of websites: they are improving at understanding what users are looking for. That’s where AI is helping, says Naguib Toihiri, Head of SEO & Growth at Artefact MENA
User behaviour revolving around seeking and discovering information online has never been so complex. Considering the context of today’s society and its abundance of information, instant gratification is the status quo. A recent study from Sparktoro found that, for the first time, over 50% of Google searches end without a click on an organic or paid search result.
Nowadays, search engines are no longer a simple index of websites: they are improving at understanding what users are looking for. They are repositioning themselves to become “answer engines” rather than just “search engines” by efficiently providing users with what they need. Artificial Intelligence will aid search engines to understand the human’s natural language by identifying meaning associated with every query.
Starting Search with voice
Made popular by features such as Siri, Google Voice Search and Alexa, voice search queries are becoming a significant percentage of web search requests. Google has already reported that 20 per cent of searches via its mobile app and Android devices are now voice queries. The new generation of web users is much more ready to adopt new technological trends and improvements. According to ComScore, this figure is expected to grow at a fast pace: it is predicted that up to 50% of all searches will be voice searches by 2020.
As a digital marketer, the best way to help users find what they need is to better understand their search intent. Voice search queries are becoming longer and more conversational; in order to win the voice search battle, website content needs to be constructed in the form of questions people are likely to ask and take advantage of long-tail keywords.
New technologies such as Artificial Intelligence, particularly Natural Language Processing (NLP), can help us validate both the quality of website content and the ability of search engines to process this content. A recent initiative developed by Pragnakalp Techlabs has demonstrated the capacity of a machine to answer questions based on the nature of content in multiple languages, including Arabic.
As Artefact aimed to pioneer voice search, we developed the first voice-based predictive shopping list for Monoprix in France. We gathered data scientists and creatives to identify a valuable use case. To launch the smart shopping list, users simply tell the Google Assistant: “OK Google. Connect me to Monoprix” and start dictating their shopping list on Google Home and then complete it on the e-commerce site.
The rise of Visual Search
The popular adage “a picture is worth a thousand words” is applicable to the search marketing industry. While still in relative infancy, visual search-led by platforms such as Google Lens and Pinterest Lens-is primed to become one of the biggest search marketing trends in the near future. Users can now take a photo of a physical object, upload it on search engines (also called reverse search) to start their journey online. Using advanced algorithms, search engines are able to retrieve information about the image content. In 2018, Snapchat launched a visual search feature in partnership with Amazon that allows users to search for a product by pointing their app’s camera to a physical product or its barcode.
One of the most interesting ways in which visual search can help digital marketers is in regards to influencing user decision-making. Given that visuals can have a significant impact on the credibility of a site, providing high-quality images with all of the necessary information plays an important role when a user is making a buying decision. Amazon, Google Shopping and eBay all maintain strict minimum requirements in order for a product’s image to qualify as good. Those requirements will help the search engines understand the content of those visuals and better-classifying image subjects by using Artificial Intelligence.
An interesting case in point was a recent experiment conducted with an aim to benchmark good Artificial Intelligence performed at the task of distinguishing a chihuahua from a muffin. More specifically the experiment sought to test and compare the image labelling capabilities of a variety of computer vision API vendors, including Amazon Rekognition, Microsoft Computer Vision, Google Cloud Vision & IBM Watson Visual Recognition. Google was the only API to successfully identify “muffin” as the highest probability label.
Finally, Ben Silbermann, Pinterest CEO, stated: “The future of search will be about pictures rather than keywords”. Visual search is not expected to replace search queries and text-based searching completely in the near future, yet it will inevitably disrupt the search industry.
Amazon is the new Google Product
The battle between search engines has moved to a new stage in the e-commerce industry. Amazon is now the starting point for consumers to search for products online – a title formerly held by Google. High visibility on Amazon search results pages is strategic for brands: 70% of users stay on page one, and 50% of clicks come from the top three products.
To improve its shelf space on main queries, brands can leverage multiple channels: price, Amazon profitability, advertising… and content quality. To manage its catalogue quality, brands currently have access to a list of modules that allow them to customise and enrich the content of their product pages: an example of this is a feature named Page A + that allow brands to add image, video, comparative graph etc. on its product pages. With its enhancements, Page A + is an additional conversion lever that allows merchants to reach 3 to 10% more conversions compared to a classic page, according to Amazon.
At Artefact, we can help you at every stage of your Amazon journey, from launch including automated catalogue creation to the day-to-day management and monitoring through our Retail Suite.
Artificial Intelligence will play a huge role in content creation for e-commerce. NLP can help marketers work efficiently by automatically generating content at scale. Recently, the Allen Institute for AI released a demonstration of a model that generates content automatically. Every 30 minutes, the application will grab the top headlines from aggregator websites and produce articles from scratch.
Think Beyond the Search Bar
The main takeaway is that we are no longer optimising for search engine bots, but rather for a better user experience. It has become essential to understand the online customer journey in order to identify which lever should be activated in a digital strategy. Stepping up search strategies beyond tactical optimisation techniques is a must to capture the current demand by creating a tailor-made search experience using Artificial Intelligence.
Want to know more about how we are intending on keeping ahead of the game based on all these changes? Contact our experts.