Οperations Support key functionalities

Enelai platform comprises an integrated environment to monitor, manage, modify/improve and measure the virtual assistant experiences across all your digital channels. The platform encompasses the ability to monitor conversations in real time (or later offline for QA or Compliance), enables agents or subject matter experts to intervene when it’s necessary to maintain a positive User Experience and to help improve performance over time, and can also function as an Assistant/Consultant for the contact center agents during their conversations with customers.

Human-assisted Learning

Enhance the deployed language models via a simple graphical user interface with minimal effort and zero machine learning expertise. It’s as simple as typing new phrases, words, product names, regionalisms, slang, then identify Entities within these phrases and assign them to existing system Entities that comprise the User Tasks. 

This user-friendly tool also shows all the user phrases that were not recognized in production during the previous day(s), allowing a contact center employee to review and assign them to Entities and swiftly improve the training models of the Virtual Assistant for highest recognition accuracy.  

Human-Assisted self-service

When this feature is enabled, enelai virtual assistants will escalate difficult user request to live agents for accurate and timely interpretation. In transactions where the system has difficulty understanding what a caller requested, enelai asks a human agent to invisibly intervene by quickly interpreting the user phrase for the application.

The data from these interventions is tagged, stored and used to re-train and improve the virtual assistant language models to interpret future requests with accuracy, thus decreasing the number of calls that need to be escalated to agents. This learning loop mechanism makes it possible to deploy new applications to a production environment as quickly as possible, and then increase the accuracy with the help of human agents during the first few weeks. As a result, the time required for development, training and fine-tuning is dramatically reduced and end-users perceive the system’s performance as being on par with speaking with a human agent.

Supervisor tools / QA

Enelai allows your team to supervise any call and chat in real time by simply clicking an on-going call/chat. The same GUI with added data (and the system’s decisions in each dialogue step) is used for reviewing calls/chats after they are completed. The supervisor can open any completed call or chat, listen to user audios (if the input was speech), review & evaluate the transcribed text, the system’s decisions, re-classify the results (in case of erroneous system decision), or identify actions that will help to improve the self-service process/dialogues.

The data from the QA reviews are saved and used to improve the virtual assistant language models.

Divert Calls to Chat

In several cases, it’s simpler for a user to complete a task –such as a customer service or technical support question- using a graphical environment rather than using the IVR voice-only User Interface. For these Intents, it’s best to divert the IVR callers to the enelai chatbot in the Mobile App.

When this is enabled for certain Intents, the enelai Virtual Assistant in IVR will give customers the option to switch from a phone call to a GUI-based messaging conversation. If the user agrees, an SMS with a hyperlink is sent on their mobile phone to continue the chat from the point it was left in the IVR.  For Enterprises that wish to promote their Mobile App, the Assistant will also request user consent to download & install the App in their phones and then continue the chat. Customers will be able to easily resolve these issues and won’t have to wait on hold to get answers to their questions, schedule appointments or make transactions.

Contact us
for more info and case studies