Superior enterprise and customer experiences driven by signal capture and relevance boosting using machine learning.

Experience Management

Simple adjustments to the data experience shouldn’t require engineers to get involved. Fusion AI’s Query Workbench gives you point-and-click configuration to shape, enhance, and sharpen the user experience. Evaluate configuration changes before they are committed to live to production.

  • Redirects and landing pages
  • Facets and filters
  • Security trimming

Signal Capture

Fusion AI captures and aggregates signals like queries, clicks, views, purchases and 
other user behavior to create a customized, optimized experience for each user.

Recommenders

Powerful recommenders for collaborative filtering for several varieties of recommendations customized for each user:

Collaborative recommendations that re-rank results based on clicks including models like user-item, item-item, query-item, and query-query models.

Personalized recommendations that aggregate all of a user’s history and actions, and highlights items for further exploration and discovery.

Predictive search that suggests items and documents before a user even enters query.

Relevancy Tuning with Preview

Full control over relevancy with simulated preview before committing changes live to production—and of course rules for boosts and blocks as well.

Machine Learning

Add intelligence to your apps with machine learning models that are pre-tuned and ready to use immediately – and create and train new models on existing data sets.

Classifiers for Understanding Intent

Classifying with Models

Classifier technology analyzes incoming queries to understand intent and fill in the blanks leading to broader, more diverse, or more precise answers as desired. Fusion ships with the following popular classification models:

  • Random Forest
  • Logistic Regression
  • Bisecting k-Means
  • Word2Vec

Enriching Data

To train an effective classifier you need rich data. Server ships with several techniques that take raw, unstructured data and define more features to train the data on. Including:

  • Term Frequency (TF)
  • Term Frequency - Inverse Document Frequency (TF-IDF)
  • Word2Vec

Increasing Understanding

There are many ways to get information about your incoming data. Server features several jobs that help you better understand the data you are working with. Including:

  • Common Misspellings
  • Statistically Interesting Phrases
  • Head Tail Analysis
  • Multi-armed Bandit Experiments

What can Fusion do for you?

Ready for enterprise search without the enterprise headache? Download Fusion and start building your first search app in minutes - or contact us today to learn more.

Stay up to date on Fusion