Researcher Gets Skills — and Smarter File Uploads

Researcher, got to remember this week. So I've got a few features to show you. The first one is around skills.

And so I've got a project here and I'm going to use some skills. But first of all, I'm going to nip over to the admin area. This is where you can create a skill.

So it's an org admin thing. And we're going to use this one here. So we're going to have a look at it for a start.

So you give the skill a name, and then this description is used by AI to trigger the use of that skill. So here, it's going to be used when we're summarising variances. Okay, and then in the instructions on the next step, in this particular case, it just wants to control the format.

So we've got the headline formatted like that, and then we've got this sort of rag style bullets going off. So that's a skill. Now I'm going to Add that skill to our project.

And I'm going to upload something as well. So we're going to pretend there's been a finance review call. These are the notes from that call.

So I'm just going to ask a question. about variances and I'm hoping that that skill is going to be used to drive the format of the response. And we can see it coming through. So we can see it's given us that headline and then those rag style bullets.

So that's project level skills. In addition to that, there are skills that come from the central product connectors now. And one of those connectors that's got some skills is the Financials Connector.

So I'm just going to connect to my data set. Now I want to do a health cheque on a particular customer. So I'm going to choose this skill.

I could choose more than one skill if I want to, but that's the one that I want to go for. So, I'm going to do a health cheque on Nulla. So these skills are particularly useful for a very broad product like financials, because the idea is it's going to guide it to use The.

The right stuff. Oh, this has gone off paste. I'll just try one thing.

But the idea is it's supposed to guide it to what part of the system to actually query. I don't know why that's not working. I'll move on to the last. the last demo, which is round uploads, uploading spreadsheets and CSVs, which you've always been able to do.

But now, this is a little bit more optimal. So you'll have seen that this is quite a large spreadsheet, but that's uploaded quite quick. And now if I go into it, you can see that it's actually storing the data and it's understanding the data.

So for example, it's understanding that some of the numerical data is actually numbers. Let me close that and we'll ask it a question. So now it should be able to get the answer a lot more efficiently, so a lot quicker and more accurate as well.

So there we can see that that's come through. So yeah, don't know why, don't know why it's had trouble with that one. But hopefully you get the idea anyway.

That's it from me, Elliot.

Jon Coello 22:57

And Carl, just to add a bit of context, just to cheque my understanding is right. Clearly, anything around the skills that are attached to either connectors or more generally could be connected to a template, a project template. So as a soft code is delivering this to

Carl Rippon 23:12

Yeah.

Jon Coello 23:13

their customers, they can ship with these sets of templates that are already set up. with these connectors with the right skills or with the skills that are connected to the project, right?

Carl Rippon 23:22

Yeah, exactly. Yeah, so yeah, you can put the skills, you can wrap them up in a template as well, yeah.

Jon Coello 23:26

So the customers don't have to think about this stuff, they can just create a new project based on a template and not start asking the right question.

Carl Rippon 23:33

Yep, exactly. Cool. Any other questions?

Cool, I'll hand back to you, Elliot.