AI Analytics

The backbone of everything Sift does is built off the intelligence we add on top of the unstructured conversations in your social ecosystem community. That's how we decide what things need action, what things are just useful information, and more.

Within the Analytics section you're able to dig deeper to understand the topics people are discussing, trends, sentiment, and the effectiveness on your support team.

On the first tab ("AI Analytics") you'll see different ways to understand the content of interactions on the Sift platform.

This is organized by:

• Clusters

• Content Insights

• Relevancy

• Conversation Topics

Clusters

Clusters allow you to understand the types of conversations taking place across your social channels with summarizations and groupings generated by AI.

Sift also provides a sentiment analysis for conversations within a given cluster, shown on the right.

Along the top you are able to select filtering by time period and sources.

Click a cluster to see conversations that belong to the cluster.

Content Insights

To provide a deeper understanding of sentiment around individual topics and across social channels as a whole, Sift provides a breakdown on the "Content Insights" tab.

These graphs show the distribution of sentiment and trends to show how it changes over time. Use filters, across the top of the page, to focus on a particular source or timeframe.

Relevancy

Sift uses AI to analyze each message and evaluate it for relevance to your brand and if it requires action. These evaluations help triage interactions for your support team, typically eliminating non non-relevant and non-actionable items before they enter a support queue.

While graphs provide a quick overview, to see details of how messages are being evaluation, click "Table View" on the top right corner.

You'll be presented with a table showing a summary of the interaction, Relevancy scores, Actionability scores, and the channel the message came through.

Conversation Topics

To help understand trends around individual topics, Sift provides a "Conversation Topics" page which shows high-level groupings of individual topics. Next to each topic is shown how many interactions related to the topic have occured in the selected timeframe (selectable from the top of the page) and change in volume.