TopQuadrant How to Build an AI-Ready Data Foundation
فصل
-
0:00
thank you everyone for joining us this afternoon um we are very excited to be here and joined with all of you with all of you guys um to chat about this topic it's something that we've been spending a lot of time thinking about um working w…
-
5:00
And one of the core ones that we focus on is the fact that the historical data stack was not necessarily focused on, you know, it was not built for this moment, right? We had a storage layer and there were a lot of ETL tools that helped kin…
-
10:00فصل 3: a model around them. 299s · Speaker 2
a model around them. It's thinking about kind of core entities and components of your organization and making sure that they're in a single source, then they're referenceable. Taking the metadata around all of your structured and unstructur…
-
15:00
different chemical names over time um compound names they they're they have different active ingredients themes and different you know um in different geographies. All of this information is sitting often in unstructured formats in differen…
-
16:42فصل 5: in and how you might change a project plan depending on variables that might impact it. 303s · Speaker 2
in and how you might change a project plan depending on variables that might impact it. You know, a shortage of labor, a shortage of a particular good, a weather events, things like that. And how can you actually change the project plan in …
-
21:45فصل 6: models that already exist into the TQ data foundation to help customers that don't, not have to start from scratch. 83s · Speaker 1
models that already exist into the TQ data foundation to help customers that don't, not have to start from scratch. The second pillar here is activate. We're trying to, you know, we're trying to, once you build the context layer, we want to…
-
23:08فصل 7: you, Amr. 302s · Speaker 2
you, Amr. And I've shared my screen. Can everyone hear me and see the screen? Looks good. Yep, so thanks for that. I think Amr went through really well, you know, why AI needs context. I hope that all... make sense. Hopefully I don't have t…
-
28:11
level data to those higher level classes that we've defined in the model. For our customer example, that could mean, okay, you've defined risk status as a thing that exists. What are all the different kinds of risk status that we would want…
-
33:13فصل 9: I want to do something really cool with it. 300s · Speaker 2
I want to do something really cool with it. We have you covered. We have integrations with all the big AI players. And I will just show a couple examples. Here's one of Claude Desktop. So you could put Claude Desktop, connect your data foun…
-
38:13فصل 10: just some things like, you know, as we think about, you know, for you all, what's next, right? 105s · Speaker 1
just some things like, you know, as we think about, you know, for you all, what's next, right? And the way we think about, you know, when does it make sense to start having a context layer or start building one is, you know, first start thi…
-
40:01فصل 11: thought and we'll open up for some QA if there's questions from the folks on the call. 302s · Speaker 2
thought and we'll open up for some QA if there's questions from the folks on the call. Great. Thank you so much, Amar. This is Ingrid, Head of Marketing here at Top Quadrant. We do have some questions coming in. If you have a question, go a…
-
45:04
single source of truth just what we've seen so like you can have you can manage um customers in your crm for example that's fine but jira and other systems need to know about that and so um you know just saying put everything in one place a…
-
48:50فصل 13: Yeah, I like that, Stephen. 69s · Speaker 3
Yeah, I like that, Stephen. Even when that logic is in specific applications like SAP or things like that, there are still often breaks to other systems, right? And there's still kind of jumps that when we're doing work as humans, you're ta…
-
50:00
here and here, and then think about how you reconcile it. Great. Thanks. A couple more questions, and then we'll close it out. Is it necessary to use a knowledge graph as a context layer? Our answer is yes. But that's mostly just because we…
-
51:32فصل 15: What was the original question, Ingrid? 89s · Speaker 4
What was the original question, Ingrid? Do you need a knowledge graph to make this context layer? Yeah, I think that was it. Is it necessary? Yeah, I think I like... I mean, yeah, I don't know. It's like a, I can imagine there being context…
-
53:02فصل 16: And I think that's one of the core things in the platform that we help customers do is we help you connect. 105s · Speaker 2
And I think that's one of the core things in the platform that we help customers do is we help you connect. core business concepts to your data. And then, you know, you can model your policies within the product and you can bring in, whethe…
-
54:47
yeah oh god yeah i'm not steven no no i i i yeah you get me excited all those cool things we could show um and i think that that's the access control is like kind of an easier one right because that's like black and white and you can say th…
-
56:39
Yes, and of course, you know, with some updates to the shackle constraint language that Holger and the team, the open source team on that committee is working on, that should be really excellent as well. Looking forward to that. Yeah, so ok…