Daisee_logo

Daisee provides automated quality management for customer interactions, powered by artificial intelligence and speech analytics.

Our solution radically improves customer experience by surfacing insights to improve your self-service solutions and ensure high value interactions are directed to the best agents. Daisee has a unique approach to improving quality by identifying the critical areas of friction that result in complaints and the loss of valuable customers. These can result in complaints, or even worse – losing valuable customers.

Daisee accurately filters through all the information in the interactions and quickly locates any issues. Daisee achieves this through a unique, comprehensive and layered process that uses Artificial Intelligence and Natural Language Understanding in a way that no other system can. Daisee can evaluate conversations in depth to uncover real meaning and intent, giving you an edge in delivering superior customer experience and identifying your best agents.

Daisee has developed a solution that leads the world in accuracy and insight.

Unique Solutions 

Semantic Engine

Trained and tailored for your business

Daisee ingests conversational media into our patent-pending Semantic Engine, utilising a comprehensive combination of Artificial Intelligence (AI) and Natural Language Processing (NLP) to derive Natural Language Understanding (NLU). Daisee’s machine-read comprehension of the underlying context of the communication is far superior in focus than traditional speech analytics that use a simple word-match and counting approach. The software groups language based on meaning which is essential to locating desirable features in interactions where similar things are said in different ways.

  • Natural Language Understanding not just Processing
  • Derives MEANING rather than just word-matching

The foundation of the engine are phrases as they provide the meaning lacking in single words and enable the decoding of sentences and longer conversations. The model behind the engine is hierarchical, exploiting the mathematics of word context to learn increasingly generalisable language features. Generalisation is the key property that separates deep learning and AI from traditional machine learning. Critical to information retrieval from transcripts, generalisation means similar phrases can be grouped and isolated, providing accurate phrase-matching and multi-dimensional categorisation.     

image002
image004

Programmable Scorecard™ 

The world's first

The world’s first Programmable Scorecard replaces the traditional paper scorecard with an automated digital scorecard that is displayed and works in tandem with Media Player.  The Programmable Scorecard will automatically evaluate several criteria to rate elements of the interaction to identify issues and flag them for human review.  Critical issues can be easily seen within the media player and users can jump between points of interest or toggle them on/offIn addition to points of interest, the media player displays customer sentiment and instances of over talk and silence.  

  • Digital, automated scorecard replaces traditional paper scorecard 
  • Automates analysis of customer interactions 
  • Customisable within 3 core areas – Communication, Compliance & Conduct 

Scorecard questions are customisable and relate to 3 critical areas – Communication, Compliance & Conduct. Each scorecard question/topic has a positive, neutral, or negative rating as well as a descriptive rationale detailing how the solution came to this conclusion This ensure that the AI is explainable and easy to understand. Scorecard questions can be programmed to filter questions based on the type of call or team the agent is in to ensure the scorecard is evaluating quality appropriate to the customer interaction. When a user listens to a conversation using the media player, the solution follows the transcript of the conversation and semantic search capabilities are enabled. 

Issue Remediation Workflow

Issue Remediation Workflow automatically triages low-scoring, high-risk calls for operational personnel to review within the Programmable Scorecard™. These calls can be flagged if they require further action and remediation – comments can be allocated, and an audit trail details actions and outcomes. Data from the interaction workflow that require further action can be extracted periodically – either manually or automatically based on the customer preference.  

  • Automatically triages low-scoring, high-risk calls 
  • Calls can be flagged for further attention and remediation 
  • Comments made and an audit trail details actions and outcomes  

Call Drivers - Daisee Essence  

Understanding the "Why"

For the first time in the world, Daisee has been able to derive the reason your customers are contacting your business. Daisee Essence captures the principle reason calls are made, providing an instantaneous snapshot of call categorisation, insight into evolving call trends, and revealing new trends when they emerge. Daisee Essence employs the semantic engine to identify the key phrases likely to complete the caller statement, “The reason for my call is _____”.

  • World first ability to automatically and accurately identify why your customers are calling.
  • Allows the presentation of trends in the topics of interaction

Traditionally, key words and phrases with the highest counts, which may simply be repeated language, are classified as "Call Drivers". Daisee Essence, by contrast, are the phrases most meaningful to the organisation and describe the real reason calls are made. Crucially, the phrases are grouped semantically in clusters to provide a true count for each Call Driver and allow the presentation of high-resolution call trends.  The interface can select time periods of interest and show the impact of seasonality and external events on the reasons why customers call.

Verbatim & Non-Verbatim Script Adherence

Script Adherence measures elements of the call to ensure that the Agent has stated required information within the call for Compliance and Regulatory protection of the customer.  These statements can be Verbatim – or an exact match of the words, or Non-Verbatim where they convey the understanding of the issue without having to use the exact words. Daisee leverages its proprietary Semantic Engine to evaluate Non-Verbatim Script Adherence and can determine how well the information conveyed the issue to the customer. 

  • Automatically triages low-scoring, high-risk calls
  • Calls can be flagged for further attention and remediation
  • Comments made and an audit trail details actions and outcomes

Daisee can invoke different Scripts depending on the business situation, and many customers have complex requirements that they need to check across different products or geographic situations.  

Script Adherence is an integral component of our Programmable Scorecard and the rating weight can be set to ensure that problematic calls will be highlighted for rapid remediation.