How to Get the Most Out of the AI Assistant 🤖

Hi, everyone.

I just wanted to talk quickly about how the assistant works, because I've been on quite a few calls with people and one of the things I keep getting requested is how do I make the most of it?

Like, how do I get the most out of the assistant?

And I want to talk about the way that it works behind the scenes because it will give you a better understanding of what you need to do and how you get the best results.

The way most people are used to working with AI is to use it in the same way you use Create mode in a way.

Although Create mode is accurate because it's based on your facts.

You make a request, it goes across all this knowledge and it comes back to you and gives you, something back.

Now, this assistant's different because this assistant can do that.

Obviously, it can reference all of your facts and all your information, but it's got what's called tooling, which is the ability to execute things for you.

And so the way that you communicate to it, when you need it to do those things is quite important.

And just understanding how it works, I think will be a bit of an unlock for some people in getting the most of it.

And to be transparent, it's kind of exciting this, because this is such early stage with this type of technology.

We're learning as you're learning.

and I think sharing this might help everybody.

So when you're working with this assistant, the thing to understand is the way that it is designed is that it will look at what you request.

So we might ask it to make a task, right?

So we'll go like, add a task to onboard a new customer, Dave Jones.

And when I make this request via natural language, it's then going to work out what it needs to do.

And it's got a host of tools.

Think of these, like, capabilities behind the scenes that we've developed to enable this to go and execute particular tasks.

And they're all based on the language you use.

So that's why we talk about it in natural language control.

Now when we ask this, the assistant is going to look at the request and then it's going to work out which of those tools to use to execute the task for us.

So.

And in this instance it is a task, so it's gone and added a task force.

And you can see that over here.

Now the key thing to understand here is that when you're communicating with the assistant, the clearer you are in kind of what you're asking it to do, the better result you're going to get back and the easier it's going to execute.

Now one of the ways this gets really complex is we have a host of different things that the assistant can do.

For example with NOAN network.

And I'm just going to jump into one of these, one of these profiles here, and just talk through kind of what can actually happen in this instance.

So the assistant can create any new contact for you in your NOAN network.

It can add any of these details and pass these details out.

So that's the name, the email, phone number, website.

If you add those, it will just pass it out.

It can update any contact with any of those in it.

it can add the company, company to it, it can add the name, the role, it can add any memos that you want.

Now the thing to understand is if you want to update any of those things, so you say I want to add a company to Drew Adams and it's, it's X.

The assistant will pick the right tool, tool to add the company to Drew.

If you want to search and you want to find out across your memos, can you look at somebody, you met at one event, who was it and you give an overview of who they were and then it comes back, it's going to go to the memos first and then it's going to look through that and then it's going to go look at who the people are.

So a really good way to think of this is you're giving directions to the assistant on where to go and what to do.

So, so if you were to search, you're not just using kind of more traditional referencing what we call rag, where it's going and looking at all this knowledge and extracting it.

You're actually like setting it off in a direction.

So the way I do that is I might say which people in my network work at anomaly and it will go and look at the people and then the roles and then come back with all the people I know who work at Anomaly, I might say in my memos I met somebody at X event.

Who was it?

Can you tell me who their profile is?

And it will go to the memos, look across the memos and then come up and find the person and come back to me.

And in many ways you're kind of almost just giving it directions of where to go and what to execute and you're kind of treating it like you would an employee.

If an employee was kind of in a library, for example, to go really old school and you were trying to give it directions on what to go get, you would tell it where it needs to go and which direction or what to look at and what to find.

In many ways you're talking to the assistant in the same way when you're recalling information in this environment.

so I just think it's really important to understand that it's all about directing the assistant for what you want it to do.

And then similarly the product names in here, everything you see and everything is labeled the way everything's labeled that is reflected in what the assistant knows and understands.

So if you said a private note or a private contact, it will go and add a private note or private contact.

If you ask it to add a role, it understands what role is, it will add a role.

If you ask it to add a company, it will add a company.

it knows all of these different things.

and so as a result you can be really clear and direct in what you're asking it to do.

And the clearer you are, the better results you're going to get back.

I've seen a few people who are can be a bit wishy washy in the way they're talking to it because they're not expecting kind of giving it direction, then it can do as much as it can, and as a result you might not get the best results back.

So it's great.

The advice I would give is really to actually play with it, give it directions, learn and understand how the product works and what everything's called and then just talk to it like an employee as if it's using the same product as you and I'll execute those things for you.