Mastering Agent Insights and Interaction Rules 🤖

Hey, everyone.

I'm Malin from Fynch.

And in this video, we're going to talk about agent, insights and interaction rules.

So Fynch operates really as your virtual marketing department.

And in order to do that, it needs deep persistent memory.

So there's lots of ways we achieve that, which I'll talk about in other videos.

But in this specific video, one component of that larger memory is the Agent Insights.

So to orient you once again, inside of the, Fynch workspace here.

And I'm inside of my account for, AI Powells, which is an independent bookstore here in Portland, Oregon.

And on the home screen here, you'll notice this banner is what I want to call out.

So there's this call out from my agent that says, I found three strategic insights based on your recent activity.

So this is something that will continually show up the more you interact with your agent in the workspace and projects and tasks and content and everything.

So this is continually learning in the background, your agent's thinking about all, of your recent activity and will surface these proactively for you to review and approve.

So I'm going to go ahead and click this, the review button here.

And what that's going to do is it's going to take me to my Agent Insights tab inside of my profile.

So you can also visit this directly inside of the My Profile tab via the main navigation as well.

But here you'll see these are the three insights that have surfaced that I can either accept or reject.

So this is a relatively new workspace.

So when you start off, some of the insights are pretty basic.

But as you continue to use the platform and your agents to do real power work, your insights get much more powerful and a lot more interesting and tailored to you.

So the more you use this and kind of coach and, interact, this just gets better and better.

So first one, you'll see this is actually really basic.

This is actually tagged as profile info.

So this is saying you explicitly referred to www.pals.com as my website and an AEO, report.

So it's tagging that that's our company URL.

Pretty basic.

I'm going to accept that and say, yes, that's true.

Next one up here is also profile specific.

It's saying, I recognize that the company's name is formerly Powell's Books.

that's also true.

Pretty basic can accept that one as well.

And you'll see this third one is actually based on interaction preferences.

So this is where these start getting much more granular and tailored to me specifically.

So all of my interactions that I have within the workspace and with the agent, this is going to be tailored to me.

So my other team members, who are also in the workspace and doing their own projects and we're collaborating, they're going to have their own recommendations that the agent makes to them, which would be different than mine because we have different preferences, we have different needs and ways in which we're working with, our agents.

So this one is saying, please provide clear, direct and transparent explanations of internal platform mechanics, tool capabilities and background workflows when answering any support questions.

And it gave some insight to say this came out of, a note that, I've been asking for this recently, in the last week.

as the agent's been working on some of my requests.

So this is also true and helpful.

So I'm going to go ahead and save this one as well.

So I've saved all three.

Any of those you can also reject if you don't want them saved to your profile.

But then what happens here is you can go into your profile tab, I can click Edit and then scroll to the bottom.

This is also a way where you can manually add agent interaction rules as well.

So let's say you're first getting started in the workspace and you know, you have specific jargon or acronyms or, preferences, no emojis, whatever it might be, those sort of really specific things.

You can come in here and just manually add and save to your profile if you want as well.

in addition to the regular insights that get surfaced from your agent, in that recommendation platform, in the recommendation process too.

So that's it.

Get started.

The more you start using it, the better your recommendations will be.

And excited to see what you built.

Thanks.