Qlik Embedded Analytics and AI Deep Dive

I'm Lucia Alice. I'm a senior product marketing manager for the ISVs and Embedded Analytics, and I'm here with my friend David Shannon, and he is the, uh, product manager for embedded analytics and ISVs, um, solution as well. Uh, we heard about, you know, Nick and Kyle and our friends from AWS talk about the value of ai, talk about predictive AI and, you know, generative AI and how this add value to, uh, all the business, right?

So we are here, David and I, to talk about how we can unlock this value by embedded the AI and analytics into applications, into solutions that our users, our end end users leverage every day. How to make this value, it is Intel intelligence available to them where they do their work. So we work with, uh, um, the ISVs, uh, team very often and their applications like healthcare or FinTech or even, um, you know, Salesforce or ServiceNow.

So we're, um, um, they, they're looking for information at their hands, real time insights, uh, at their hands, right? So it makes, uh, it, it was a time where it know we have analytics in a separate, um, um, window in a separate application, a separate context. It's more and more we're looking to for this into the, this, um, intelligence as part of the, uh, context and, uh, you know, ma helping to make real time decisions.

So, uh, this is why we are going to walk you through the demo that Dave, you're gonna show later on. But, uh, before we get into that, I wanted to kind of give, um, insights into what is really driving this, uh, need for embedded analytics, right? So first of all, it's, um, is this need for real time information.

So, um, more and more business are looking to be more effective and have the information at their hands, right? The AI is great, the po you know, predictive is great as, uh, Kyle mentioned. You know, we can get, um, great, uh, insights for that.

But if this is not embedded, this is not part of a, uh, workflow or a system that this information is not at the end user at the right time, there is no value, right? Uh, so that's the need for real time decisions. The second one, it's the, the end users, right?

So today we have, uh, it's no longer just the business analytics or the data scientists. They're leveraging AI and data analysis. Um, so even though though this more advanced users, they're looking for more self-service capabilities, uh, more altering by itself, but also as Kyle mentioned, we have our business users.

They're looking for, um, the information at their hands via text, but, you know, just, um, answer of their questions, not looking for, uh, dashboards or reports. So this need to have different solutions, the answer to the need for the answers, the need for different, uh, users as well. And the third one, it's the operation efficiency.

So it's to have, uh, this analytics and AI embed into the workflow and automated to make business more effective. It's interesting statement from IDC. Uh, and the question for you would be, um, I mean, reading this, you know, customers are willing to pay your premium applications with integrated AI capabilities.

Do you think that customers know what those AI capabilities are that they need or should be using? Good question. Um, probably not the word ai, but they're really expecting intelligence insights, you know, data-driven insights.

If you think about like today, um, as in users, we are all kind of looking for applications that in give your answer to us, right? So, and, um, you know, chat or any others, but give you, uh, intelligence at, uh, uh, you know, um, as part of their applications. So that is the concept.

They're looking forward to pay more for this intelligence. Right? So is it fair to say then that they give you a business description more than a, than a really an AI description of what you're Fair, fair enough.

Leading off of that, for the ISVs that you're working with, what value point are they driving towards? Like, you know, the typical value point for ai, right? Is productivity.

I don't really subscribe to that, but I get it. Uh, is what are they, when you say intelligence, the ISVs are building to do to, to enhance their, their customer's experiences, how and their customer's business outcomes? How, Yes.

So when in the con, actually, this is the context for ISVs. This comes from our reports to ISVs. So Lever providing them this intelligence or this data insights, right?

Inside their applications, some of our ISVs customers, they're actually able to charge some premium prices for that, uh, as well. I'm sorry, I meant their customer value points. So that's good.

We want the ISVs to make more money, but they're doing it by providing features or capabilities. Yes. Like predictive intelligence, for example, right?

We have one of our customers that, for example, they have, uh, uh, A CRM, you know, or a supply chain, and they are able to predictive for, uh, provide for their income in customers predictive insights that before they use not to. So for this type of insights, they can charge for premiums, and that for the end customers actually decap the, the, um, the value to a, be able to predict. Um, in this case, it was like the meat, right?

Um, um, the forecast of, uh, waste and, and stocking of meat. Um, save them so much, uh, you know, um, money reduce their costs. So the, and customers definitely see that value, the leveraging ai.

So, um, it's, I I would say for the ISVs, um, segment, it's a shoe starting, right? They are still, there's A category of additional features, like, like, uh, forecasting or prediction. There's a category, and then there's probably also specific features like chat-based interaction or that, Yeah, it's, it's the ISV building on behalf of the end customer.

So the end customer doesn't have to, that's kind of the, the night Yeah. Accelerator here. Yeah.

I don't, I don't wanna press you on the point, but it's, it's sort of getting at, to my sense, it's getting at, okay, so something has changed in this ISVs software, and that's something allows them to charge more. Yeah. It also delights the customer more, maybe like that sort of thing.

So what are those somethings? And those somethings are either like an additional capability that didn't exist in the platform before, or better interactivity with the current capabilities. Those are two examples.

Yes. Uh, well, better insights. Yes.

Mm-hmm. Thank you. Great Question.

I too, okay. Sorry. No, we're good.

Great questions. I, I had jumped that. Um, great question.

Um, so this is like, um, talk, when you talk about embedded, right? So talk about this modern embedded, um, analytics we're, there is a shift to, right? What being used to what was embedded before, just like a chart or a reports or just a graphic.

So because of the, um, this need to have ai, more interactive, more insights, self-service capabilities, we're, uh, just only dashboards. It's not, uh, enough anymore, right? Talk about like click answers, right?

Some, some users, as I mentioned, they are looking for just a answer for a question. So all this, it's available, uh, through embedded capabilities. It's, uh, you know, a service a, you know, ISVs for as one of, uh, our segments.

So, um, in terms of like, what, what click offers for this, like this robust capabilities, um, what's the difference, right? So it allows, uh, our customers ISVs or even enterprises, right? 'cause they also have enterprises.

They embed their solutions to, uh, allows a richer personalization, customization of their applications. Um, they also allow them to scale this. Like for, especially for IS Espec, especially for our ISVs, um, scalability and the multi-tenancy environment is very important.

So it's because we're a SaaS platform, we allow then this, uh, very elastic, you know, uh, flexibility and scalability. Also, this new model embedded application, they are not only, um, at the desktop anymore, it has to be available at mobile, um, devices. Like, for example, some of our svs, they use that into like, uh, a supply chain as a supply chain manager, like, you know, on, on or on a store, a retail store.

So this information has to be, uh, mobile where they do their work. Of course, as we so many times here today, it has to be secure and governed, right? Uh, because, you know, the, um, we talk about data and especially the IV is talking about the i, the ISVs data.

So the security and governance, it's very important to that. And, um, and in bad analytics, um, must be like, as a very, uh, seamless intelligence inside application, the end user. They, he does not see what is inside there.

They don't know that, for example, click, click is part of their application. So here, let me, uh, kind of walk you through, uh, uh, what are click's capabilities and how we kind of stand out at the market, um, in, in value capabilities. I kind, we kinda separate this into, uh, four, um, four, uh, buckets, right?

So the first of all, first it's embed into the modern web. So click has, it's, um, it's, it's on framework. It's called clicking Bad, right?

It's, uh, give us the possibility or give the customers the possibility to embed the full application or pieces of that you choose it. But we can to embed the full application and give to the end users all the self-service capabilities of clicking click sense, for example, right? So he can do analysis as, uh, Dave is going to show that in the demo, uh, after all.

Also, um, the clicking bed, uh, doesn't require, uh, collecting cookies and is a light framework and the full, uh, integrated with the modern, um, modern, um, um, other frameworks like React and it's is built. So if you were gonna build a mock architecture microservice API driven cloud native headless, you would use the embed into the modern web. I didn't catch all of that, by the way.

I was, that was pretty good. JS speak. That, that was, yeah, that was very good.

Yeah. Yeah. I, I'll show you some more whatever demo we do the demo.

Okay. The demos using, I didn't mean to go deep, deep, but No, no, no. It's, it's good fun building the demo.

Oh, yeah. He had a lot of fun. Yeah.

Okay. Yeah. Um, Let's Not say the other three aren't bad.

I'm just saying that, was this the one that caught my attention? Okay. Okay.

Um, so yeah, I, and if you, uh, you know, they have mo we have multiple UIs options, uh, there, but also if you wanted to go a little deeper, you can customize that, leveraging, clicking bad. Um, the second one is the analytics and predicted, um, predictive and generative ai. So all the, uh, our capabilities and the one that are coming, um, in, you know, in our, uh, roadmap that Nick and Kyle mentioned, they're all native embedded by clicking bag.

Um, and also we, uh, able to deploy that into the multi-tenant too as well, um, which is very, very important to ISVs. Two, we not only deploy, but um, we were able to, um, escape that. Um, the third pillar is the deployment and management.

So, um, this is, you know, um, the deployment and management of, uh, the clicking bed. It's, um, secure by design. So at each mo multi tenancy, it's, uh, has its own security.

And we also able to control entitlement and, uh, capability bank, and we also have a capability bank. And pretty soon we will have a, um, a cloud console that will, will, uh, be able to manage the whole entire, uh, uh, entitlement and all the multi-tenancy that, and last but not least, is, um, the end user altering. Uh, so that is what really differentiate click, as I mentioned, we can embed the whole entire application and, uh, this gives the end user the possibility to not only customize the applications, but also to, uh, bring your own data to that and, and do your own analysis.

So, um, and here's all kind, um, how everything comes together, right? From, um, the access the data, integrating, create the, uh, solutions, creating the experiences, deploy that into a multi-tenancy environment, security by tenant, embedded that into, uh, you know, um, ISVs applications for example. Or just like regular applications like sales, Salesforce and Teams, and ServiceNow, you name it.

And many others, um, available into mobile, uh, in a web. And also, um, create leveraging to the end users, um, self, you know, their self-service, uh, capabilities. And with that, oh, that's all because we have, um, you know, it's a C native solution and totally API driven, so we can automate the whole entire pipeline of that.

So if that, because I like to also show more than tell, so I'm gonna bring David to kind of show our demo. Great, Thank You. Yeah.

One of the nice things that I've picked up, so it's SaaS platform, so we kind of hold the state so you can run against it with things that aren't stateful. Mm. And we maintain a state so that you can be a little bit more iterative.

So I had a lot of fun building this stuff. I found out only a few weeks ago that I had the opportunities to do this session. And, um, I'm a product manager, not a programmer.

So I admit I did use a little bit of AI to help me build this out, but actually it's amazing what you can do with some published APIs. Um, and that's been my journey for the last few years. So what I will introduce you to, and I'll refresh that to set myself up, but is a fictional ISV, um, as I say, a lot of, I'm building this smart chain, no relation to anything that exists or could exist, right?

So just set that out there now, um, as it happens, that's a live URL, so you could try it if anyone wants to have a Oh yes, straight in. Okay, very good. Um, and what you, what it basically does is it shows you the end to end.

So I'm just impressed you got the domain. Well, ISV demo, it's a little bit like, little bit too generic, but I was quite pleased with it. Yeah.

Um, so what we basically have yes, is we have the end. It allows a end customer to come along and to actually subscribe to spool up all of the infrastructure and tenant and data that we need, and then land them into an interactive experience and end to end. This is all production now.

So there's, you know, there's not many futures here, but you'll gain all of the stuff that we've just heard about in the near term anyway. So it's all kind of built in and all evolve as we move through. So as a focus on ISVs, one of the things that we really focus on is being able to provide differentiated experiences, right?

So make the product more compelling, improve stickiness, improve that consumer experience that people expect from these applications. And so we talk a little bit about being able to do things like pricing and packaging, customizing the experience your users have, and to differentiate seeing who's gonna sign up here, I'm gonna do it too. So maybe I'm gonna get in a queue, but let's say I wanna sign up for this, so I'm gonna sign up.

Uh, and what it's gonna do is it's gonna start onboarding and behind the scenes. So let me just, uh, let's call it app me, Acme Corp. And I'm gonna choose some sample data.

Now, obviously if this was a real product, we'd be onboarding your data and you could use QTC or click down and cloud or any of your other data components to do so. And that's really the strength of having all of those published APIs. We have all the frameworks and Lucian talked about click embed, but we also have click API, which is a TypeScript framework that actually has all of the APIs wrapped within it and everything that you see in this demo.

So, and behind the scenes, we are setting up a new tenant. Um, we are adding the users to the tenant. We are creating spaces, we're creating the, uh, groups and roles that we need.

And we're creating some knowledge bases, importing some apps, you know, indexing those knowledge base and then also, uh, creating some assistance that you can use. So what's going on behind the scene? And now while I wait for it to, uh, speed of thought very nearly, this is, I was for You privately.

Well, thankfully we're done. So I can now come in and as a new customer, I now have a totally isolated tenant all set up within this infrastructure. This is one click Cloud tenant for this customer behind the scenes.

Um, it's reloaded an app, and we've got a whole bunch of stuff here straight away that's loaded off of that experience. If I give you a quick overview of this app bearing on not too long on time, but I've had a little bit of fun. So I've included a few things that you've already seen, so I don't need to go into too much detail, but we've added the answers experience up here with some examples.

So, you know, how do I override product categories? So questions about how to use this application. Uh, tell me about our coffee products.

So questions about data that's actually in the application. We've got the direct access to a clicksense app and the engine. So I've put some KPIs on here.

If I press the cog, it'll, it will load the master measures from the app, and I can obviously customize that in my web app. And this is showing kind of the more customized, deeper integration that's possible if you're using click API to really dig into the hyper cube and grab the data from underneath. And then for some slightly more legacy technology, uh, we've got Insight Advisor down the bottom and the questions API, which is generating insights.

And so you can ask different questions here, and of course this is going to evolve as we get more of this into answers. And you'll see, you know, there's plenty of things that we can do in this space. So hang on, you're, you're, you're kind of blowing my mind here because I'm thinking about, I've been thinking since last year and beyond about how, uh, what Qlik has been doing for a few years now, um, allows the, uh, you know, maybe more outside of generative ai, which is only a piece as we've discussed, but certainly within AI for quite some time, allowing data scientists to create, um, or tune or otherwise utilize AI for advanced analytics, um, and acquisitions and so forth have ensure data quality and so on and so forth, right?

And that's how I've just tended to think about it. It's like a direct model. This is like a tiered model for doing the same thing, because now you have a supply chain specialist, this fictional example, doing that possibly across, you know, in some very nicely compliant way across a broad set of customers, um, creating or tuning or otherwise creating a model that now then individual customers can go in mm-hmm.

Mm-hmm. And tune further or develop further for their specific models. So each step of the way accelerates, uh, time to value for the utilization of AI using not just the expertise at each level, but you know, the, the, the, the most relevant data sets at each level.

Yeah. The ISV brings the expertise and the tooling, and then the end user brings the expertise in the data and in the use case and the problem, it's What their own particular model as the, whatever their business is meet was. Yeah.

You know, well, yours was meat. Mine might be, I don't know, cheese and it's gonna be different. Yeah.

But across all of either food or goods or whatever. Wow. That's, um, that's kind of cool.

I kind of like that. I, I'm pretty pleased with how this turned out. So anyway, so if I, if I, yeah, another, if I, if I dig down slightly more.

So this is, um, you know, showing you an overlay of a bunch of different things that you can do. If we dig down, I've, I've set out a number of different ways that you can interact with this. And you know, I talked about this concept of being able to give different experiences to different users and give the customer control.

And in here, you know, some of this is, you know, the app, right? Obviously. Um, but underneath everything's powered, well, not everything, but most things are powered by click.

Anything, it's data, anything that's the ai, anything that's the insights is all click. And anything that I change in here is going to be pushed down to the tenant, you know, as they make the request instantly. So, you know, I've got this organization here, I've got some users, you know, I talked about having some subscriptions that allow 'em to do different things, to, to limit, you know, what tiers to build a product offering that's very compelling and drives people.

So yes, money for the ISV, but also stickiness improve the, uh, quality of experience. And in this example, um, I could do something such as invite, uh, you know, one of my colleagues, you know, I'm a customer of Smart Chain, I can invite one of my colleagues along and actually I'm gonna just do that now. Uh, now let's do something generic and then, and I'm gonna, I don't really trust them.

I don't really trust, trust John. So I'm gonna put him as a consumer. I don't really want him to be going in and making changes to any of my stuff.

Uh, that sent invite and you can now join. And obviously this invite process is all built in the app, but it's going to be doing, sending down to the tenant and creating the relevant users and setting up the access control of those users based on their role. Is this, is this the API endpoints then for adding users, like in the, in the cloud?

Yeah, exactly. Yeah. Management console effectively, right?

And the whole thing behind this uses a mixture of of machine to machine and all machine to machine impersonation to do that depending on what the user is trying, How, how much of the UI stuff is click embed, then how much of that is like changing UI objects just based on what's straight out of a click app and how much of it is deeper API Work, I'm pleased that you asked. Yeah, so the, the dashboard here has got, um, a little bit of deep deeper API work, but it's only listing, you know, those master measures that I talked about. So here I really have just done list measures and if I click on one of these, I felt failed, say the preference is it's a live demo, okay, I'm gonna blame the network.

But that is going down to the click out. That's only, you know, simple API call 'cause it's done using the click API framework that abstract some of that complexity. So Using the scaling properties that are on the click and embed calls to like leverage the, Well this one, this one's the kpi.

So it's raw data. Okay. Okay.

But then, you know, if we come back and then look at one of the other sheets, so analytics then, so if I say I have an analytics sheet, analytics is the analytics experience that we have in I got you. I see. Yeah.

And you, you know, obviously that is then customized slightly, you know, you can do things like add some metadata to determine which are core sheets or whatever else you want to do. And I've called this something slightly different. And of course you've got the usual, you know, selectivity et cetera, that you'd have in click, uh, click sense native.

And then let's say, well Just last thing that was the click click platform operation, API in the background there for spinning up all the, adding the users and all that. Yeah. We have the, yeah, the tenant service and that's using the multi-tenant service.

Yeah. So yeah, and, and as of, yeah, as you actually spun this up, this was all done directly against the APIs, but it also does some things in automations and I can show that as well. You've then got things like the analytics altering, so allowing the customer to build their own experiences for their users.

And this is using that full fat click sense client. So everything that you would do normally, right? If I wanted to duplicate this sheet, it's gonna be a bad demo because I'm just duplicating it.

But uh, and then if I wanna publish that, then all of a sudden that's gonna be available to anyone that's coming back into this native analytics experience, this lightweight analytics experience. So you've made your changes and you want to, you know, roll with them and now it's a custom sheet. So there's various different ways that you can do this.

Didn't take me very long to build, as I say, two weeks around my day job. Uh, managed to, to pull this together, use render for This. Yes.

Fantastic platform. Yeah. Amazing actually.

Good demo. Good demo. Good.

And then pages, pages was, so pages anyone's used, client managed click sense. Uh, this might look a little familiar, but this is showing a little bit more like going down to analytics charts. So down one level from having the whole sheet layout, you are then building the sheet layout.

So I could pick a bunch of sheets here and let me just drag that one from that sheet and I'll drag a few of these from this sheet, you know, but whatever, I can build my own sheet, uh, my uh, sheet. Very good. I'll make it visible to everyone and I'll just say the first sheet, whatever, right?

I'll save that page. Fantastic. That worked.

Uh, I'll drop into pages and then of course it's right here and you can do a bunch of things that you know, this is using the lightweight framework so it loads very quickly. You can turn interactions on and off. I've turned them off in this case so that they can't, you know, come in and mess on things.

They can still zoom but they can't make selections. And then I've added a component here to do things like download it to Excel just through calling the reporting APIs. And obviously you can do, you know, the usual stuff because it's all um, native objects.

One of the other nice things is that you can build everything to be, you know, like fully responsive. So you know, you can take it right down. It works really beautifully on a mobile device.

Um, there is a problem with logging it on Safari, but you know, yeah. Oh I'll fix that. Um, you've then got, uh, predictive, I'm not gonna dwell on this too much because we just had a fantastic presentation on that.

But you are the usual stuff going down to the click predict and uh, ML APIs to generate those bits of feedback reports. Same thing, you know, reporting services, you have it either in the Clicksense app or you can generate them. You really have a lot of choices as to how you do this sort of stuff.

The user that I just invited though, so if I was just to come and in uh, accept this and I'm also using superb base, which is also superb base. Yeah, yeah. Fantastic.

So there we go. John Smith. Fantastic.

That's now gonna set up my account, obviously done on the tenant, bounce you back in and then of course I'm gonna see a cut down experience because I'm not trusted and I've just got, you know, analytics pages. Nothing Yeah. That I can actually use to create.

So really quite straightforward in doing that. And then on the click side, so let's say you actually wanted to see what's going on in this tenant. It's got the monitoring for all of that.

When that was stood up, and I'll just open these two automations. This one is the, yeah, monitoring out setup. And this is the trigger, the trigger fired.

So we created a new record for a tenant. It's just a triggered UL web URL pretty much. And then it triggered a run of the reinstall of the reload apps.

You can see it's still going, it's deploying for 23 tenants now. Um, and you can see that it's just deploying all the apps and this one's already published. You don't have to make it yourself.

You can pick this up off of click, open source, uh, repository. And then at the end of all that, and I've put a little backdoor in just so I can log into the tenant to show this. But let's say I go back to here, That's come up here.

I can see my tenant. I've just sort of snuck it here. Now you as an end user won't be able to access this 'cause it's all old impersonation.

You're not like a real user that has a real or login on this tenant. I can log in because I had it myself. Um, and I can see the tenant's real, I can see all the content that's on that tenant.

There's knowledge bases, there's assistance, there's apps, there's data files, all sorts. As the end customer, I can then pop up here and I can say, well not really for me, you know, I've had enough come down here and delete my account. Little drastic, but I just wanted to show the offboarding.

Right? Bringing you around to that lifecycle of saying you can then have a customer offboard themselves. You are not, you know, all hands off.

Come back to this tenant, I'll refresh it. Tenant's been shut down and that be purged within whatever period you define 10 to 90 days. That could be pretty freaking dangerous.

That's why I had that warning about the admins at the bottom and all. Well Yeah, that's been all, He just took out the pop-up that says, are you sure? Are you sure you are sure He didn't have to type his own name in, which is the dead giveaway that you are definitely not the person you say you are when you delete.

So this is actually the, the whole thing I was talking about before, the SaaS, the SaaS apps, especially the email apps obviously had places where that needed to be backed up. This looks like another place where it said, do you guys have a uh, a third party app or a partner that works with you to protect the data? So if a mistake is made and something is deleted when it shouldn't have been that someone can retrieve it.

Yeah, there's partners on the show floor that provide exactly that. So awesome backup, you know, services they use the public APIs in, in an instance ISV there, it's actually tends to be less of a concern 'cause you are most things that deploy programmatically. So you own all the core content.

There might be some customer data, but that's coming from their system. And then the only thing that's really the delta is the sheets and the bookmarks and the content they've created in click sense and maybe some questions, history if you want to keep that and that you can all grab via the public APIs. So it's really not, and there Are partners that do.

Okay, that's Absolutely, yeah.