Worknet supports multiple options for leveraging GenAI natively in Slack. As a prerequisite, Worknet needs to enable its GPT capability, and the Worknet app must be invited to any channel where you would like it to take actions.
All operations described are based on the same WorknetGPT engine and the same data sources. It is possible to control access to data sources when needed.
The following four (4) are available.
- Stealth mode - channel message
- Stealth mode - thread actions
- Public mode - automatic response
- Public mode - @worknet
The following describes each one in details.
1. Stealth mode - channel message
Use WorknetGPT for you and your Worknet team without other Slack users seeing it. Worknet uses Slack's ephemeral messages, which are visible only to selected people and will disappear later.
In a new message box, type /worknet followed by your question. Upon posting, the original message will disappear and it will be shown as an ephemeral message visible only to you.
A few seconds later, you'll receive the answer.
The answer will include links to sources such as knowledge base documents, Slack messages, customer tickets, Confluence, and more. When using Slack and tickets, it will even identify the names of the people involved in the content.
You'll be able to provide feedback using emojis.
The message will eventually disappear from the channel.
2. Stealth mode - thread actions
Worknet allows you to consult with its GenAI when you are in a thread. Using Stealth mode is optimal when you do not want other people in the channel, such as customers, to view the response.
All actions are available when you are inside a thread.
Hover over the message and click the three dots at the top right corner. This will open a few options. Select WorknetGPT.
It will display the following ephemeral message. Click the "Select an item" option, which will list two options:
1. Get thread summary
2. Suggest a response
When clicking the "suggest a response" you'll get an ephemeral message with the suggested response as you can see below.
In this case, links to the data sources are also available, as well as the names of the people involved.
You can use Worknet's conversational ability to keep asking questions. These might include further knowledge questions, such as specific data about pricing in this instance. You can also ask it to change the tone, make it shorter, or translate it into a different language.
Feedback - you can provide feedback using emojis as well as textual feedback.
Automation - It is important to note that this option for ephemeral thread responses can also be automated. An ephemeral response will be posted to any message in the channel and the thread, making the response easily accessible.
3. Public mode - automatic response
This feature is designed to enhance customer service channels. Initially, WorknetGPT will provide the first response to inquiries, offering quick answers and deflecting routine requests. If the automated response is not satisfactory, a support team member can then join the conversation. Additionally, this feature allows customers (internal or external) to initiate the creation of a customer service ticket if needed.
It is enabled at the channel level. So any message in the selected channels will trigger and automatic response.
Users can continue conversing with Worknet and ask follow-up questions.
As in the previous examples, the response here will also include links to the data sources and the names of people involved in those responses.
4. Public mode - mention @worknet
This option allows Worknet users to control when are they interested to get Worknet to participate in the conversation.
Only when @worknet is mentioned in the beginning of the post Worknet will attempt to response. @worknet can be used in multiple places:
- The first post
- Any post in the thread
The actions that you can request it to perform include:
- Find info in the provided data sources - knowledge base, previous Slack conversations, previous tickets, Confluence and practically any dat sources.
- Rephrase
- Translate
- Make it shorter
- Anything that GenAI supports
Comments
0 comments
Please sign in to leave a comment.