Analyse sentiment in emails - Sentiment Analysis
Overview
Our sentiment analysis pattern, available in Marketplace in Builder enables the analysis of the content of incoming emails and determine if, for example, they are positive or negative. This assessment can be passed back into Enate Work Manager so that Agents can tell at a glance what the tone of the mail is as headline information as they start to deal with it.
This pattern is available via the following providers:
EnateAI Setup
EnateAI is a zero configuration option supplied by Enate which uses our own Azure hosted OpenAI Instance for the GPT3 Engine.
This instance is shared with anyone that enables EnateAI so if you have high volume/high demand then you may be better off configuring your own Azure Open AI instance and using an Azure Open AI Adapter Version.
The setup of EnateAI for all patterns is super simple - simply click to activate.
OpenAI GPT3 Setup
OpenAI Adapters allow you to talk to the OpenAI API directly via your own OpenAi account and API keys, this will be using the OpenAI API which is open to everyone. You can also monitor usage of API requests sent via this API in the OpenAI interface.
To setup OpenAI for sentiment analysis, follow these steps:
Click on 'Activate' on the 'OpenAI GPT3 - Sentiment Analysis' option in Enate Marketplace.
In the resulting pop-up, enter the following information:
the URL of the instance - this is the base URL for OpenAI, i.e. https://api.openai.com/
the API key - this you can create from the OpenAI API keys page https://platform.openai.com/account/api-keys
Click to test the connection
Once the connection has been successfully tested, click to activate. Your Sentiment Analysis pattern provided by OpenAI is ready to be used.
Further notes of interest
The OpenAI API keys page is where you can create further API keys and manage your existing ones.
You can monitor the usage and breakdown of requests of the OpenAI account on the OpenAI Usage page:
Azure OpenAI GPT3 Setup
AzureAI allows you to use your own Azure instance of OpenAI which you can manage and control, giving you a great deal of control and detail of the information going through the API.
To setup Azure OpenAI for sentiment analysis, follow these steps:
Click on 'Activate' on the 'Azure OpenAI GPT3 - Sentiment Analysis' option in Enate Marketplace.
Enter the following information, all of which can be gathered from your Azure OpenAI instance, in the resulting pop-up:
Resource Name - this is the Azure Resource Name which can be found by navigating to your OpenAI Cognitive Services in Azure, accessing your Azure OpenAI Resource and opening the resource. The resource name can be found in the overview page, next to 'Resource group'
Deployment Name - to find this, go to the 'Model deployments' section in Azure OpenAI studio and then click on 'Manage Deployments'.
Then click on the 'Deployments' section on the left and the deployment name will appear on the right - this is the value you need.
API Key - this is configured back in the OpenAI Portal, from the 'Keys and Endpoints' section on the left. This is where you can generate new keys and manage existing ones. Enter one of these keys.
Click to test the connection
Once the connection has been successfully tested, click to activate. Your Sentiment Analysis pattern provided by Azure OpenAI is ready to be used.
UiPath Communications Mining Setup
UiPath Communications Mining provides communication analysis and to help automate your processes.
Note that the Sentiment Analysis pattern works slightly differently when using UiPath Communications Mining as your provider. Communications in incoming emails are still analysed, but the assessment is not passed back into Enate Work Manager. Instead, the assessment remains in you UiPath Communications Mining system, allowing you to auto-tag and categorise emails in a certain way according to how you've set UiPath Communications Mining up.
The following video takes you through how to setup UiPath Communications Mining for sentiment analysis.
Last updated
