ChatGPT is an AI-powered chatbot extension for Zoho Cliq, designed to assist users in generating tailored responses to various queries using Natural Language Processing (NLP) technology. The bot offers context-aware replies, summarizes conversations and files, outlines drafts, and much more, all designed to save time, boost productivity, and enhance communication efficiency.
The extension is available in the following regions:
United States (US), India (IN), Europe (EU), Australia (AU), Canada (CN),Japan (JP) and Saudi Arabia (SA)
Installation
- Go to the Zoho Cliq Marketplace and search for "ChatGPT for Zoho Cliq", or install it from here →.
- Specify whether you want to install the extension for your entire organization or specific teams and click Install.
- Accept the extension's permission requests, and you're good to go.
Onboarding
- Upon successful installation, ChatGPT's bot will post a welcome message with a disclaimer on Cliq.
- Click the Get Started button to consent and grant the necessary permissions for the extension to work.
Note: After granting permission, the bot will send a message to request data privacy consent. You must accept the OpenAI terms and conditions to continue using this extension.
Configuration of Open AI Tokens in Extension
- Follow the steps in the ChatGPT bot's message and get the secret key from the OpenAI website.
- Admins can select a configuration type within the extension to set up the OpenAI Secret Key or Token via bot or Widget.
Note: Sharing API keys may lead to sensitive data leakage and pose a security threat. Make sure to handle them securely.
Via Bot
Click the configure button or navigate to the Configure bot menu action and choose the token.
In the Configuration tab, click Configure Authentication Mode to set the token configuration for the organization.
Note: In extension version 1.6 and above, admins can set tokens with a size of 100 and above (project-based keys).
There are three options to select a token configuration type.
- Configure for entire organization : This option will use one API key (that the admin provides while configuring) for the entire organization.
- Allow employees to configure : This option will allow each user to use their own API keys from their OpenAI accounts for the extension.
Reconfiguration
The authentication mode can be modified through a bot or widget, and the token can be reconfigured if a new key is used.
Via Bot
Navigate to the Configuration bot menu action and select Token.
- To modify the authentication mode, click the Change Auth Mode button.
- To reconfigure the token, click the Reconfigure button.
Via Widget
Open the ChatGPT Widget.
- To change the authentication mode, click the Change Authentication Mode button at the top-right corner of the widget.
- To reconfigure the token, click the reconfigure secret key button.
Assign Keys for Specific Roles
- This feature enables admins to efficiently assign keys in bulk by creating and managing groups based on user roles or email patterns.
- Groups can include individual users or roles such as Admin, Member, or Developer. Groups can also be created based on criteria like the initial letter of the email address
Managing Groups:
- Use the Create a Group button in the widget or bot to set up a new group. After creating a group, you can assign users or roles based on your needs.
- Once a group is created, you can update its information by clicking the Update button. This allows you to edit the group name, modify roles, and reconfigure tokens for the selected group.
- If you no longer need a group, you can use the Delete button.
Admins can configure the extension's PII (Personally Identifiable Information) settings to enhance security. By default, the PII configuration is set to Track. To modify these settings, navigate to Configuration > PII via the bot menu action.
Why Configuring PII is Essential?
- Data Leakage Prevention: This ensures that AI does not unintentionally share or process sensitive information, such as personal identifiers, financial details, or other private data.
- Regulatory Compliance: This service helps organizations adhere to data protection laws, such as GDPR, CCPA, or HIPAA, which require strict handling of personally identifiable information (PII).
- Risk Mitigation: This reduces the likelihood of mishandling sensitive data by applying predefined controls to messages.
- User Trust Enhancement: Creates confidence among users that their sensitive information is protected during interactions.
- Flexible Security Measures: This allows admins to customize actions based on their organization's security policies.
After selecting a configuration type, admins can define PII keywords in the textbox, separated by commas. The following configuration types are available:
- Only Track (Default): This option tracks and logs the number of Personally Identifiable Information (PII) detected in messages. No further action will be taken.
- ⚠️ Alert: If PII is detected, the sender will be alerted and asked to confirm before further action is taken. Open AI execution will only proceed after confirmation.
- 🚫 Block: Messages containing PII will not be allowed to be sent by the user.
- 🗑️ Remove: Any detected PII will be removed from the message before it is sent to Open AI for execution.
Features
Image Generation
Generate an image using DALL-E based on a prompt or a description.
- Bot menu action: Navigate to the ChatGPT Bot to find the Imagine bot menu action for generating an image based on your specified prompt.
- Slash command: In any chat, use the /imagine command to initiate the image generation.
Utilize search and browsing capabilities to find answers to your inquiries through text generation models developed by OpenAI in the Cliq platform.
- Slash command: Use the /chatgpt command to ask any question or request the AI to generate text-based content, such as a code snippet or a blog post.
- For example, you might ask the AI to create a meeting agenda to discuss the progress of the week's goals.
- You can mention @chatgpt in any conversation and ask questions, the bot will respond within the same conversation
Playground
The ChatGPT Playground is an interactive platform offered by OpenAI. It allows users to experiment with and refine the responses generated by ChatGPT. In Zoho Cliq, you can set up a playground configuration in the ChatGPT Bot to create customized responses based on your preferred data model and examples.
Accessing the Configuration
- Open the ChatGPT bot in Cliq.
- Click the ⚙️ icon at the top-right corner to open the Playground configuration menu.
Choosing Models
Under the Models section, select the text generation model for your responses. Currently, the following models are supported:
- GPT-4: A highly advanced language model ideal for complex reasoning, creative tasks, and detailed comprehension. It provides high accuracy and versatility.
- GPT-3.5 Turbo: A faster, more cost-efficient version of GPT-3.5, optimized for conversational applications and high-quality text generation.
Note: Each model has different usage costs, so choose one that fits your budget.
Setting Example Prompts
Under the Examples section, you can select predefined prompts to tailor the bot's responses. Supported prompts include:
- Create Study Notes: Refine content into concise study notes.
- Mood to Color: Convert text descriptions into colour representations.
- Classification: Categorize text into predefined classes.
- TL;DR Summarization: Summarize long text into a brief overview.
- Summarize for a 2nd Grader: Simplify text for a second-grade reading level.
- Grammar Correction: Convert ungrammatical sentences into standard English.
- Interview Questions: Generate thoughtful interview questions.
- Chat: Engage in conversational interactions by default.
Fine-Tuning Responses
After setting an example prompt, refine the bot's responses using the following parameters:
- Temperature : Set this to control the randomness of responses. Higher values produce creative and diverse outputs, and lower values generate more focused and predictable results.
- Maximum Length : Set the maximum number of tokens (words or characters) to manage output length for generated responses.
- Top P : Top P influences response randomness by considering only the most probable words.
- Frequency Penalty: Setting the frequency penalty discourages the overuse of common phrases to encourage varied outputs. For example, if a word like "awesome" is overused, a higher frequency penalty encourages synonyms like "remarkable."
- Presence Penalty : The presence penalty reduces the likelihood of introducing rare or new words. It helps simplify text or maintain readability.
- Inject Start Text : Adds specific text after the user's input to guide the bot's response.
- Inject Restart Text : Adds text after the bot's response to maintain a structured flow.
Saving Your Configuration
- After adjusting all settings, click Save. Your configuration will be stored and available the next time you use the bot.
- Interact, experiment, and engage with various models, prompts, and settings to find the solutions that best meet your requirements
Outlining
Turn text into a clear outline to organize ideas and decisions.
- Message Action: Use the Create Outline message action on any text to draft a quick, rough outline of shared ideas or decisions.
For example, let's say your manager requests you to write a blog on a particular subject. You can use this message action to get an outline of the blog that you have to write, and you can build upon it.
Summarization
Effortlessly generate concise summaries of long chat conversations or shared files to spotlight essential points, making information easier to access and understand!.
- Slash Command: Use the /summarize command to provides a concise conversation summary in any channel, thread, or direct message as bullet points or highlights.
- (Note: Summaries are based on the last 25 messages only.)
- Message Actions:
- Summarize This: Use the summarize this message action to condense a long message to provide a clear gist without the need to read through the entire content.
- Summarize this File: Use the summarize this file message action to summarize an attachment from any chat.
- Ensure your attachment is in a supported file format and that your pricing plan includes OpenAI Assistant features to enable this functionality.
- Depending on the file size and content, summarizing the file may take some time. During this period, a processing message will appear in the chat. Once the summary is ready, click Check summary results to view it
Important Note:
Cliq leverages OpenAI's Assistants API, an external service based in the USA, to process the uploaded file and generate a summary. The entire attachment is securely processed to provide a clear and actionable overview of its content.
Exercise caution when uploading files containing business-sensitive or personal information (PII). Only share attachments that are compliant with your organization's data-sharing policies.
The ChatGPT extension for Zoho Cliq is an incredible asset that can elevate your team's productivity and communication! Its cutting-edge natural language processing capabilities deliver quick and precise answers to your questions and easily support various tasks.
Whether you're eager to enhance your workflow, foster better communication, or boost efficiency, the ChatGPT extension is invaluable. You can try this extension from the marketplace or add it to Cliq by clicking the button below.