Introducing Apla, the AI Account Planner Agent.
Introducing Apla, the AI Account Planner Agent.
Join the Webinar
This agent retrieves 50+ messages per channel (or individual DM history) over a specific time period, then writes a summary based on the things you’ve said you care about.
This agent can be adapted to handle a wide-range of use-cases. For example, you could summarise resources shared, questions asked by the team or customers, bugs squashed, and so much more. The possibilities are endless.
We recommend that you set high-risk tools, especially customer facing actions, to "require approval" until you're happy with how your agent is performing. Then you can change them to "auto-run" so the agent can complete work without your supervision.
💬
Build a tool to retrieve 50+ messages over a given time period from your chosen Slack channel (individual or team). You can change the number of messages retrieved, the time period and the channel. The data retrieved will include the sender, the message and attachment details etc.
Build or use this tool ->
💬
Build a tool to post a message to your chosen Slack channel (individual or team). It uses a built-in Slack Message tool-step provided by Relevance, which will guide you through connecting your Slack account.
Build or use this tool ->
These are the settings we used to configure this agent. Every setting is completely customisable. We recommend that you get this agent working using our default settings, then start experimenting with making small changes.
Create & Configure an agent ->
Slack Summariser
This agent summarises Slack channel histories. You can choose which channels it summarises and what kind of things it shoud summarise (e.g. questions asked, resources shared, team highlights etc).
No triggers for this agent, but the "Post to Slack" and "Slack:Retrieve messages" tools both require connecting to your org's Slack account. The tool steps above will show you how to build and use these tools.
GPT-4o
You are an expert at writing summaries of team activity over the last week. These are the tasks you must do:
1. Retrieve the last 50 messages from every single channel you have been given.
2. Summarise what has been happening across those channels. Write the summary following these rules:
- Start with a TL;DR paragraph at the start of the whole summary. The TL;DR should start with: "> 📣 *TL;DR Channel Summary for this week (since Monday):* " before the individual channel breakdown. Do not use markdown besides that. Double check that this is a paragraph.
Now for each of the channels, whether that's one or more:
- The pod name is the title, capitalised without hyphens.
- Wrap the pod name in single asterisks, like this: *Pod Name*. Do not use two asterisks, this isn't markdown. Double-check you have used single asterisks to wrap the name only.
- Bullet point the summary, using this symbol "•" as the bullet. Do not use dashes for the bullets.
- Do not use hashes anywhere.
- Reference people by name.
- Exclude anything that happened before this Monday.
Finally:
- Double check that you have followed the title, bullet point and all other formatting rules.
3. Post the summary to slack. Show your thinking out loud too, but double-check that you have posted your summary to slack.
To use this AI Agent for summarising your own Slack channels, you need to do the following:
1. Clone this agent via the "Use template" button at the top of this page, and click "Create Agent".
2. Create a Slack bot with permission to post to your workspace channels, and copy it's API key (tutorial for this).
3. Add your API key to your Relevance account in the Integrations page. Give it the name "slack" so your agent will know where to find it.
4. Connect Slack to your Relevance account (this will involve choosing channels you can post to): The Tool steps above have tutorial links for this.
5. Invite your slack bot to the channels you want to post to (/invite @your bot name).
6. Save your agent, then ask it to summarise your chosen slack channels. You can also ask it to retrieve a specific number of messages. You can adapt the core instructions above to your own use-case.
