Stratifyd 4.0 New Features
AI Modules 🤖
Stratifyd 4.0 introduces new ways to incorporate machine learning into your analysis and business workflow.
A number of pre-trained, industry-specific models will be included and ready to deploy into your data. These include the Customer Loyalty model and the Agent Scoring model.
In addition to pre-trained models, 4.0 introduces Auto-learning, a predictive classification engine easily trained with labeled data.
General Auto Learning
Auto-learning allows users to train their own AI models to fit their unique use cases directly through the Stratifyd platform without the need for a data science team to build out models. Auto-learning trains on your labeled data using various combinations of cascading models and evaluates the performance of each combination to identify the highest performing model and automatically selects this model for you to deploy on any other data sources.
Customer LoyaltyThe Customer Loyalty model allows users to train on a sample of customer verbatims and corresponding customer ratings, such as NPS, review ratings, recommendation ratings etc., to predict loyalty scores for all customer feedbacks.
Agent ScoringThe Agent Scoring model allows users to train on existing agent performance scores to expand agent scoring coverage to all agent-customer interactions.
Semi-automatic TaxonomyThe Semi-automatic Taxonomy model trains on an existing taxonomy model and generates a machine learning based classifier using the labels from your taxonomy. This allows analysts to use reinforcement learning on the model rather than the guess-and-check method of updating taxonomy keyword logic.
Case Management 🗃️
Case Management is a task tracking solution that is deeply integrated with the new AI workflows. Users can create and assign cases based on insights discovered in the dashboard or set up rules to automatically create cases from any data elements including predictive model results or other AI modules.
Alerts can be set on any combination of dimensions to alert a user or group of users via in-platform notification, email, or text message. Alerts can be based on static or dynamic thresholds.
Stratifyd New Improvements
Dynamic Topic Modeling
Topics can now be applied to new data without retraining, meaning the topics will grow and shrink dynamically as the unstructured content shifts.
Taxonomy Co-occurrence introduces added flexibility when breaking down data by taxonomy labels. Users can now apply filters from multiple taxonomies or multiple labels within the same taxonomy concurrently.
New Calculated Metrics
In addition to simple aggregations like average, sum, max, and min, Stratifyd 4.0 now offers the NPS calculation, CSAT (CSAT10 and CSAT25) and DSAT.
Filters can now use nested AND and OR logic to create more complex slices of data.
Administrative functions have been added for designated users in an organization to allow provisioning/deprovisioning and controlling access to content all from one console.