GEOSherpa
PageSherpa
AccountSherpa
ContentSherpa
Self-Nurturing Landing Page
Multi-Variant Landing Page Builder
A/B Website Personalization
Content Hubs
Headless CMS/DAM
Single Field Lead Form
Website Abandonment Capture
Sales Assist
Lead Enrichment
Web Card
Native Advertising & Content Streams
AI Content Topic Tagging
AI Personalization
AI First-Party Intent
Gen AI Content Chat
Gen AI Video Transcription
Gen AI Image Creation
Gen AI SEO
Secure Pages
Language LocalizationFunctionality
AI Personalization
Hushly AI personalization offers personalized content recommendations to every single visitor – known or anonymous
How It Works
Visitor to Advocate
Step 1
All eleven Hushly AI personalization models are trained and tested on
- Content that is processed using Natural Language Processing (NLP)
- Asset content metadata
- First party topics and descriptions, tags, catalog fields, and asset types
- Behavioral data
- Engagement data
Step 2
All eleven AI models feed their top five recommendations into Hushly's AI arbitration engine which makes the ultimate determination for which top 3-5 assets from all the recommendations provided will be offered.
Step 3
The Hushly AI arbitration engine makes the final decision and presents the content offer on Hushly powered pages or experiences
Step 4
The Hushly AI personalization models are refreshed every seven day or a custom interval.
Features
AI Engine Models
Hushly created an AI personalization models each looking at different behaviors, traffic sources, etc., while working engine that consists of eleven unique AI independently from one another to offer the best recommendation for their respective model. Customers can enable/disable models as well as rank them to give priority to one model over other to ensure that the recommendations are accurate and performing well
Content Similarity Model
This Hushly AI personalization model recommends assets that are similar in content to the assets already viewed by a website visitor. Similarity is derived in two ways:
- Using natural language processing (NLP) to compare the actual content (text) of the assets
- Using customer defined correlations, i.e., the tags, catalog values along with topics associated to assets
Collaborative Filtering Model
This Hushly AI personalization model recommends assets that were viewed by other visitors who also engaged with the same asset(s) as the current visitor
- i.e., users who viewed asset X also found assets Y and Z relevant or engaging
Path Similarity Model
Using this model, Hushly AI personalization model, AI recommends assets that the visitor is most likely to engage with based on the asset engagement path taken by other visitors
- i.e. users who viewed asset X followed by asset Y, subsequently engaged with assets P and Q
Trending Assets Model
This Hushly AI personalization model calculates a Z Score (trending) for all assets based on number of views in windows of 7 days over the past six months
- The AI personalization model recommends assets who's current engagement rate is higher than its historical average / trending higher
Most Popular Asset Model
This Hushly AI personalization model recommends assets that are considered most popular
- The model ranks assets based on number of views till date. The higher the number of views, higher the score
New Asset Model
This Hushly AI personlization model recommends assets that are new
- This model prioritizes assets by creation date and modified date, most recent first, and recommends the first ones
Important Asset Model
This Hushly AI personalization model calculates a score for each asset based on three factors
- The average sort order of the asset in various streams that it belongs to
- The number of streams that the asset belongs to
- The number of asset views
Tag Similarity Model
This Hushly AI personalization model recommends assets that have similar tags to the assets viewed by the visitor
- Recommendation is based on calculating cosine similarity between tag values of assets
Catalog Similarity Model
This Hushly AI personalization model recommends assets that have similar catalog field values
- Recommendation is based on calculating cosine similarity between catalog field values of assets
Topic Similarity Model
This Hushly AI personalization model recommends assets that have similar intent topics
- Recommendation is based on calculating cosine similarity between the intent topics of assets
H-Signal Model
This Hushly AI personalization model recommends assets that have first-party intent topics a given visitor or a company is surging on
Related Resources
