Social Search Made Easy

In our Google-enhanced private lives, we expect immediacy. Thus, we’re disappointed when the applications we use in our business lives fail to be easy and effortless.

An example that hits home, the ‘Social Listening’ space has long required users to write complex boolean queries. 

While we haven’t eliminated the need for boolean – it plays a vital role in dashboard creation – we’re excited to announce functionality that brings us closer to realizing that dream – AI-powered search.

With AI-powered search, Synthesio easily and quickly auto-suggests and recognizes the specific search term of interest (FIFA the organization as opposed to the game). Fast, easy, and no boolean required. 

Let’s dive deeper.

Boolean: Expert Skills Required

Mastering Boolean queries can significantly improve the data and insights you receive in your dataset. However, writing effective Boolean queries requires skill, practice, and, above all, time.

Consider the example when querying for gluten-free soft drinks:

The Trouble With Social Data

While it’s quite easy and even natural for a human to understand the context and the meaning of a social post, it’s an entirely different and complex decision for a machine.

For example:

For us humans, we quickly recognize that the post references FIFA the organization, not the game or event.

For a machine lacking human intuition, it’s not so simple. There is no @FIFA, no supporting images, no mention of football, the World Cup, or any FIFA members. 

The above example is commonplace on the social web. For these reasons, we use boolean operators to include or exclude keywords and, imperfectly, capture and categorize ambiguous terms.

AI-Powered Search

Again, writing boolean searches is complex, time-consuming, and flawed. More importantly, it’s insufficient to meet the demands of today’s businesses who require quick data and insights they can trust.

Users want to search a social topic, evaluate that topic, decide whether to explore further and then move-on.

Fortunately, thanks to Machine Learning, we can build models that will learn from the details mentioned above (called “inputs”) and drastically speed up and simplify the social searching process. Fed with millions of documents (social posts), the models begin to learn over time and precisely predict when a conversation is about FIFA the company vs FIFA the video game.

In other words, by using deep learning techniques, our models have learned to recognize different entities (products, places, organizations, countries, facilities, people, etc.).

Across the Synthesio Experience

We’ve packed this capability as an AI-assisted query feature in Flashdash, our social search engine, and across our dashboard experience as a filter option. AI-powered search:

• Eliminates complexity and reliance on boolean queries
• Saves you time when building reports or searching and validating topics
• Drives trust, confidence, and clarity when assessing social data

Flashdash

AI-powered Flashdash allows users to search social topics at a glance, eliminating the need for complex Boolean operators. Flashdash scans the entire Synthesio corpus. Search terms are suggested as you begin typing.

Dashboard-Level

After dashboard creation, AI-powered search improves your ability to search and browse entities across your data-set (products, places, people, etc.).

Simply leverage the ‘Entity Search’ filter and select the desired entities that you would like to analyze.

Building on top of this functionality, you can incorporate named entities across your Reports widgets. Here, we’ve specifically selected a set of planes (products) to compare, looking at their co-mention frequency with various emotions:

AI-search is available to all Synthesio users at no extra cost. Soon, into the future, these capabilities will dramatically facilitate the dashboard creation process. Stay tuned for more to come!