Your team works hard to produce, curate and polish thousands of articles, videos and infographics, but most of them are buried once they’ve been live for a week or two. Your readers, meanwhile, struggle to research the most authoritative, intelligent or fun content relevant to their unique needs and interests. 

This is where Algolia can step in with a solution that can conveniently work for both you and your customers. 

Our mission to connect media consumers to the most relevant content in milliseconds is now supercharged by the latest in artificial intelligence. With the release of Algolia NeuralSearch, our AI powered Search and Discovery platform, is now available to any website or media app. NeuralSearch is packed with AI capabilities such as dynamic re-ranking, personalization, recommendations, and dynamic synonym suggestions plus a lot more. 

In this article, I describe how some of these AI capabilities work and how they can assist your visitors to find what they want.

Understand your users’ needs better

Connecting a visitor to the right content is a real challenge when dealing with thousands of pieces of content. People search in “messy” ways, using different synonyms to describe the same thing (“romcom” for “romantic comedy”), we make typos (“cooiking” for “cooking”), and sometimes ask questions instead of typing in clear, search terms. Furthermore, every language has its own specificities — such as very long words or words without spaces. All this makes it nearly impossible to manually predict every new way that a user might search for a term or a concept. 

Sometimes readers know exactly what they’re looking for, and sometimes not so much. A search for the “G20 summit” for example is a query for a precise match. Other times, they are searching for ideas like “global economic conference” to discover what’s available. Search has traditionally required people to type one or two words to find answers, then use filters to refine results. While filters (sometimes known as search facets) are still important, AI search allows users to type in natural language queries like, “that conference with global leaders discussing economics” to find what they want.

Algolia Natural Language Processing

With Algolia natural language processing (NLP) , you’ll be able to handle the most complex user queries. 

Users express intent in two ways:

1. Explicitly, through words easily understandable and taggable by software, for example when they are looking for “Season 2 Episode 4 of Fresh Prince of Bel Air”, and
2. Implicitly, through natural language, like when they are looking for “the episode where Will meets his father” 

With Algolia NLP, you can recognize both types of asks from your users, and surface to them the right content, including those hidden gems.

For you, this means getting more out of your existing content, and bringing back the best content for even the most complex requests.

 

Dynamic Synonym Suggestions: let users say it in their own words 

Synonyms are a powerful way to improve the relevance of search results and to avoid giving users “no results”, which makes them think that you don’t have relevant content for their query (and causes them to bounce). But it can also be cumbersome to manage synonyms manually: defining it one by one, and thinking about all the various possible scenarios.

Algolia AI makes synonym management easy. 

Our Dynamic Synonym Suggestions automatically detects and suggests new synonyms that your audience is using to look for content, even those that you wouldn’t suspect!

“eucalyptus” <> “gumtree”
“data” <> “chart”
“two way tape” <> “two sided tape”

“ABC is the largest news platform in Australia, and our audience is very diverse. It makes it difficult to continuously connect our visitors to the right content among our 1M+ articles. Algolia Dynamic Synonym Suggestions helps us get our viewers to the right content without friction by automatically suggesting relevant synonyms for specific events or specific to parts of our audience. Since using it, we’ve seen our search greatly improving on www.abc.net.au”

Peter Buffington, Product Manager

 

Dynamic Synonym Suggestions is available for Algolia Keyword Search customers. 

Context is everything: offer relevant, personalized experiences

Many things affect what the user will find most relevant: seasonality, trends, personal preferences. Think about people looking for Tim Burton movies. The most relevant result approaching Christmas might be The Nightmare Before Christmas, while the most relevant result in the beginning of 2019 might have been the just-released Dumbo. One user might prefer gothic horror movies (offer them Sleepy Hollow), while another might be into musical comedy (a better result is Charlie and the Chocolate Factory). 

Our new AI features enable you to create more personalized browse and navigation experiences by tailoring recommended content based on user behavior.

Dynamic Re-ranking: boost the most performing results for each user query 

Algolia Dynamic Re-ranking leverages AI to analyze users’ actions on your digital properties, then boost the most performing results for each search term.

Algolia already offers a full suite of tools to manage relevance taking into account both individual visitor inputs and your own business rules.

With Algolia AI, we are now leveraging collective intelligence. Dynamic Re-ranking uses AI to analyze each query, and promotes the best performing items related to that query in the search results. This is particularly beneficial for relevance edge cases — as relevance is relative to the end-user.

We apply on top of your existing set of rules a new ranking based on the multitude of signals sent by your users. By offering unprecedented content relevance to each individual visitor, our Dynamic Re-ranking will make your audiences return.

Dynamic Re-ranking creates a flywheel to improve results based on a customer’s KPIs and events data.

Personalization: surface the content your audiences have the most affinity with 

Your visitors are exposed to vast amounts of content from various digital sources. They expect content recommendations to be diverse and personalized, and their bar in this regard is only getting higher.

Building personalized experiences takes time and resources. It’s hard to decide where to start to personalize the experience, what data to act upon, and which systems to connect together. Relying on one size fits all solutions is risky — each business and each audience is different, so you’ll need to adapt the personalization logic to the 1:1 experience. 

Our Personalization feature leverages AI to power tailored experiences to each of your visitors. It takes care of computing a personalization scoring and re-ranks your content automatically, but also comes with a simulator that shows you exactly how it optimizes the results for individual users and each query. 

This enables you to surface the content your audiences have the most affinity with, from most popular content pieces, to the very niche content titles that they would have a hard time finding elsewhere. 

Algolia Personalization is already available on the Algolia Premium plan. Read more about personalization here.

Algolia Recommend

Visitors appreciate when a media site can suggest other relevant articles in a non-invasive timely manner. Consider someone looking for recent articles on “best food for dogs”, reading the article, and then noticing other suggested articles that may include related content. Algolia Recommend can use different attributes to deliver smart results related to the content someone is reading. It can display articles that are similar by keyword, or display results based on “people who read X also liked Y”, or it can use AI combined with personal attributes such as geo, time of day, membership status, past reading history, and more, to present intelligent recommendations. 

Algolia NeuralSearch: a breakthrough in search

Algolia NeuralSearch represents a watershed moment for businesses who want to offer a human-like search experience to their customers. It combines traditional keyword search with advanced AI search. NeuralSearch combines multiple technologies — vector search, keyword search, natural language processing (NLP), reinforcement learning, and more — into a single end user experience. 

It means your visitors can type in just about anything using natural language to find what they’re searching for. NeuralSearch works even better when you can define events — clicks, conversions, “saved for later” pages, etc. Events provide signals to the search engine that will improve results over time.

Empower your editorial teams and improve content curation

Empowering your editorial team with the right tool to search, retrieve and curate content is key. With thousands and thousands of articles to dig through, it’s often hard from them to know what is already produced, what to cross-reference or reuse.

Algolia AI makes it a breeze for your editors to search for relevant content and maximize your catalog exposure in milliseconds, thus offering them the same quality experience as to your audiences. 

What does this mean for your business?

With this set of AI-powered features, your visitors and editors will be served only the most relevant content through seamless and natural interactions within your apps, websites or internal tools. Usage and subscriptions will be boosted, while decreasing bounce and no-results rates. 

But knowing that each media business is unique, we want to take into account your own editorial strategy and business logic. This is why Algolia AI is an open platform, enabling organizations to bring their own AI into Algolia. By sending the output of your AI models into our platform, Algolia allows you to couple what makes your business unique for years to our battle-tested Search and Discovery product. 

More than 17,000 organizations have selected Algolia for Search & Discovery. Read what they say and talk to us today to get a demo of Algolia for your business.