Knowledge Indexing - Slack

Project

Artificial Intelligence

Intermediate

Use Template

This is a 'Project' template which means that it contains a group of workflows that work together to achieve a particular aim

Overview

Copy

Extract data from their Slack conversations and turn them into Knowledge articles for a Support use-case by using AI.

If you need help reach out in the community in the #ai-help channel!

Set-up

Copy

1. Import the Project into your Tray instance.

2. Create a Data Tables in this Project. It should have the following columns:

   1. Link to Thread

   2. Original Message (part)

   3. Decision

   4. Reason

3. Create a Vector Table in this Project. It should have the following configurations:

   1. `Name`: Could be anything

   2. `Dimensions`: 1024

   3. `Metric`: Cosine

   4. 💡The Embedding Model this Project uses is the Amazon Titan Embedding Text v2.

4. Open the `Slack extract` workflow and open the Config Settings. Configure the Config variables:

   1. `latest_timestamp`: Slack threads **after** this timestamp will not be extracted.

   2. `oldest_timestamp`: Slack threads **before** this timestamp will not be extracted.

   3. `slack_channel`: The Slack channel where the workflow should grab the threads from.

   4. `user_to_skip`: A list of Slack user IDs whose Slack messages should be ignored. E.g. at Tray we ignore bot messages sent e.g. by a Jira bot.

5. Open the `Slack Audit + KB Creation + Vector Table Push` workflow and configure some steps:

   1. `data-tables-1`: Pick the Data Table (the one that was created in Step 2) in this Project for this step.

   2. The columns should be in this order:

      1. `Link to Thread`

      2. `Original Message (part)`

      3. `Decision`

      4. `Reason`.

   3. `vector-tables-1`: Pick the Vector Table (the one that was created in Step 3) in this Project for this step.

Next Steps

Copy

Pair this with our Slackbot and other AI Agent templates to compose your first enterprise Agent.

If you need help reach out in the community in the #ai-help channel!