Rethinking Workflow SaaS: building the aHawkAI inspections app

1 "Jason Masciarelli" (3295821568)
00:00:00.928 --> 00:00:23.599
Thanks for joining us again today and I'm Jason Masciarelli. I'll be the host and you're at the launchpad webinar for today. Before we jump in, I just want to highlight a couple upcoming events. So we've got our provider spotlight, another one coming up on 24 March around payment.

2 "Jason Masciarelli" (3295821568)
00:00:23.599 --> 00:00:42.990
Investigations with and then at the end of March, we've got a launchpad live community call with our existing providers and developers so you know follow us on LinkedIn and make sure that you register for those as we get closer to those events. But with that we'll get started. So.

3 "Jason Masciarelli" (3295821568)
00:00:42.990 --> 00:01:02.990
Today we're gonna focus on really rethinking workflows in SaaS, leveraging AI blueprint, agentic automation and really working with one of our longtime launchpad providers, which is Aaseya. We've had a great collaboration with a Saya over the last couple of years.

4 "Jason Masciarelli" (3295821568)
00:01:02.990 --> 00:01:22.990
And one of the things that they've done has been really innovative in terms of how they've looked at the area of inspections and they've built out a product on Launchpad called Ahawk AI. And so we're gonna spend time really featuring that, digging into their story, understanding really where they saw the opportunity in the market, how they're leveraging Launchpad.

5 "Jason Masciarelli" (3295821568)
00:01:22.990 --> 00:01:42.990
And really how they're helping out clients and envision, you know, where they're gonna go in the future with that solution. So pretty exciting. I'm happy to jump in quickly here. So I'm I'm happy to be joined by Stan Dilulo with Isaah. He's the head of North America sales. And I also have Tim Randa.

6 "Jason Masciarelli" (3295821568)
00:01:42.990 --> 00:02:16.730
Typically Tim and I are doing these things as a co team, and he's a fellow solutions architect with launchpad. I've been with Pega for a long time and really deep expert on Pega technology and and Pega capabilities and as part of the Launchpad provider success team. And again, I'm Jason Mashrelly. I run go to market and I also oversee our corporate venture fund Pega ventures. So with that, we will get into into getting started. So, the agenda looks long, but we're gonna move quickly through.

7 "Jason Masciarelli" (3295821568)
00:02:16.730 --> 00:02:48.620
And keep it interesting and keep it engaging. You know, really as I said, today is all about, you know, featuring what Aaseya has done innovatively on the launchpad, platform. And so really we're gonna set kind of the context in terms of what the market is around inspection and compliance workflows. We're gonna basically talk a bit about ASEA and how they're building a hawk AI, and really why they've chosen Launchpad as their platform of choice to build that capability.

8 "Jason Masciarelli" (3295821568)
00:02:48.620 --> 00:03:03.470
Look at the product, show you how it was built, show you how it's delivering innovation in the market, and then wrap up and we'll have time for questions at the end. And so with that, I will jump in and hand it over to you Stan.

9 "Stan DiLullo" (3173026048)
00:03:03.470 --> 00:03:20.190
Sure, thank you Jason. So hi everyone, the, A Hawk AI was, you know, founded out of inspections. Inspections apply to a very broad range of industries.

10 "Stan DiLullo" (3173026048)
00:03:20.190 --> 00:03:40.190
Addressing critical imperatives such as safety equality, regulatory, operational compliance, and, and even more fire and safety, health and food protection, municipal county state zoning, environmental construction, civil ent.

11 "Stan DiLullo" (3173026048)
00:03:40.190 --> 00:04:03.050
Engineering and really product manufacturing as well. So next slide. So what are the challenges that we're addressing with the inspection processes? Well, one, they're very manually intensive and filled with inconsistent guidelines. They struggle to support.

12 "Stan DiLullo" (3173026048)
00:04:03.050 --> 00:04:24.128
Support centralized management and performance reporting, corrective actions are inconsistent, prone to errors, lack in timely resolution and regulatory complexities and non compliance and violations become significant costs due to flaws in performance and quality.

13 "Jason Masciarelli" (3295821568)
00:04:24.128 --> 00:04:44.150
Hey stance. So I just want to jump in for a second. So I I was when you guys 1st told us back, you know, when you started this journey with us that you want to do something in inspections, I was like inspections, like that's not something I've ever thought about much. But as we've gone on this journey, I've realized like just how broadly applicable it is and how important it is to health and.

14 "Jason Masciarelli" (3295821568)
00:04:44.150 --> 00:05:05.390
Safety and so on. And, you know, as you go through some of these challenges I think, you know, folks just need to understand this is a highly manual process that is critical for it to be accurate and high quality. And so I think what's exciting is what you guys have done to this and love to hear you talk to more about kind of what intelligent inspections are and and what kind of that transformation is.

15 "Stan DiLullo" (3173026048)
00:05:05.390 --> 00:05:27.899
Exactly, and I think one of the the other main aspects of it is you know it's a large field force, and therefore with that and the manual processes you you get those inconsistencies and and not only flaws in what you're inspecting, I mean that's, that's the intent is to identify them, but flaws in how.

16 "Stan DiLullo" (3173026048)
00:05:27.899 --> 00:05:45.809
You're you're inspecting and the guidelines and the and the standards and the consistency. Then on top of it, which we'll talk about as well is then what are the right corrective actions? Where is the follow up? What are the resolutions? How are you monitoring it?

17 "Stan DiLullo" (3173026048)
00:05:45.809 --> 00:06:08.607
All of that, you know, is around us every day. You know, I I had a fence put in, so I had to come and and double check that I wasn't on my neighbor's property I mean that's a very simple example, but there are much more critical, inspections and and compliance that that needs to take place. Otherwise, there's penalties and fines.

18 "Jason Masciarelli" (3295821568)
00:06:08.607 --> 00:06:25.867
Yeah, that's cool. And so like you're taking basically like an AI 1st approach, really thinking about, you know, how automation and AI can help, you know, augment and support those folks out in the field such that they're able to focus and do a better job, you know, at what they're trying to do and be more accurate as well.

19 "Stan DiLullo" (3173026048)
00:06:25.867 --> 00:06:51.619
That's right. Augment and and insert and embed intelligence into the workflow. Yeah. So going to the the next the next slide. Yeah. So with all that said, introducing, introducing AI our platform transforming inspections with AI driven workflow automation. We, we kinda, we alluded to it in the in the previous slide.

20 "Stan DiLullo" (3173026048)
00:06:51.619 --> 00:07:12.119
Right? It's end to end solution, improving compliance, reducing violations and delivering insights into the hands of inspectors, right? From planning to post inspection monitoring, allowing productivity and quality guidelines.

21 "Stan DiLullo" (3173026048)
00:07:12.119 --> 00:07:32.119
Through case management and process orchestration helping to avoid the cost of non compliance and faults, which I mentioned before, right? With AI assistance providing contextual decision support throughout the process, inspections shift from being airproned manual processes.

22 "Stan DiLullo" (3173026048)
00:07:32.119 --> 00:07:37.765
Two intelligent compliance management.

23 "Jason Masciarelli" (3295821568)
00:07:37.765 --> 00:07:51.425
Awesome and so why don't if you could just maybe people on who don't know much about Isaah so maybe just tell us a little bit about Aseyah and then also why organizations are working with you folks and why you're the right organization to bring this to market. Sure.

24 "Stan DiLullo" (3173026048)
00:07:51.425 --> 00:08:12.049
Sure. Why I say it? Well, the the as I, as I mentioned before, right? Custom, well I don't know if I mentioned it before but customers recognize our industry expertise, the power of AI insights that we bring and and the sp.

25 "Stan DiLullo" (3173026048)
00:08:12.049 --> 00:08:44.487
Need at which we can deploy. All of that is in addition to the business KPI benefits of of our solutions. We focus on workflow driven enterprise solutions with embeddeded AI. That's how we deliver business applications, not generic, you know, platforms, if you will. Sure, this allows us to achieve business outcomes and operational value for, for our customers in in a significant way and at scale. Yeah.

26 "Jason Masciarelli" (3295821568)
00:08:44.487 --> 00:09:05.825
And you folks have been doing more of, custom implementations of inspection solutions for a long time. You know, I think in India and Europe and US and so you've got a lot of expertise in this and I think a lot of that was about, hey, how can you take that expertise and codify that into a product that's more scalable and deliver more impact in the marketplace?

27 "Stan DiLullo" (3173026048)
00:09:05.825 --> 00:09:28.129
And at the same time configurable. I think if we go back to, if we could just think back to the original slide, the applicability of inspections crosses many verticals, right? It's not just municipality, it's not just manufacturing. It it is a very broad range and the fact that.

28 "Stan DiLullo" (3173026048)
00:09:28.129 --> 00:09:55.874
Our platform can be configured, right? To, to not only dynamically determine what kind of inspection which we'll, we'll see, but dynamically determine what are the criteria, what are, what is the process, what automations can take place, but also be able to modify and hone that based on the specific industry in which we're working.

29 "Jason Masciarelli" (3295821568)
00:09:55.874 --> 00:10:04.610
Great. Awesome. That's great. So let's move into, you know, kind of why, why you folks chose Launchpad to build this product that you want to take to market.

30 "Stan DiLullo" (3173026048)
00:10:04.610 --> 00:10:11.549
Launch launchpad played a critical role in enabling A Hawk AI.

31 "Stan DiLullo" (3173026048)
00:10:11.549 --> 00:10:26.969
Between the blue blueprint driven rapid development, prebuilt components, connectors, accelerators, an AI native architecture, can't tress that, right? Enough?

32 "Stan DiLullo" (3173026048)
00:10:26.969 --> 00:10:47.906
And the fact that you we've got a secure multi tenant SaaS infrastructure, a cloud native that that scales. So it was the right fit to help us quickly, right? Bring an enterprise application to the market.

33 "Jason Masciarelli" (3295821568)
00:10:47.906 --> 00:11:03.619
Yeah that's awesome. And, you know, I think you you guys really embody what we talked to, you know, providers and builders about, which is, you know, don't focus on the low level stuff that is everything that you have to build to bring a product to market for.

34 "Jason Masciarelli" (3295821568)
00:11:03.619 --> 00:11:22.786
Focus on really your unique IP and value proposition and the outcomes you want to deliver, and it seems like you've really leveraged Launchpad to really focus on your expertise and the value you provide your clients versus being caught up in lots of engineering and full stack development and cloud ops. Right. That's a distraction quite frankly. Right. Yeah.

35 "Stan DiLullo" (3173026048)
00:11:22.786 --> 00:11:30.048
And and and and and we partnered well with our parent company Yash, as well as, as, as well as with Pega.

36 "Jason Masciarelli" (3295821568)
00:11:30.048 --> 00:11:45.125
Yep yep and and we've even worked with you folks on looking at how we leverage your solutions, you know, that you're building on Launchpad within Pega as well, which has been, which have been exciting, so great, let's jump into the platform and talk a bit of more specific.

37 "Stan DiLullo" (3173026048)
00:11:45.125 --> 00:12:07.259
Yeah, so, so an overview of the solution architecture includes our four, our four main pillars, field inspectors are equipped with mobile 1st technology to get their job done. Workflow automation infuses efficiency and consistency in their process which we talked about where some of the challenges previously, right?

38 "Stan DiLullo" (3173026048)
00:12:07.259 --> 00:12:26.789
Agentic AI helps automate both, not both automate document validation, helps generate standardized inspection reports and detects compliance risks. Huge from a penalty and cost perspective in order to you know get ahead of that if you will.

39 "Stan DiLullo" (3173026048)
00:12:26.789 --> 00:12:51.012
We already mentioned the secure scalable SaaS environment, right? Enables flexible integrations with legacy systems and supports enterprise security controls, role based access control encryption, audit trails, transparency, and, and very important nowadays, right? Data protections.

40 "Jason Masciarelli" (3295821568)
00:12:51.012 --> 00:13:14.831
Yeah. Wow. Sounds pretty robust. I mean in terms of the timeline it's taken for you to build this and I know you folks have had a small development team, you know, engaged in this and so you've got a pretty broad deep, you know, set of capability that you're bringing to market here, which is which is awesome. It's great. So let's let's get into the the next piece here, which is how it works in the real world. How it works.

41 "Stan DiLullo" (3173026048)
00:13:14.831 --> 00:13:45.359
Real world, right? So again, you know, it addresses the end to end inspection I mean this is obviously, you know, a summarized view, but we addressed the end to end inspection life cycle from assignments through corrective actions and ongoing monitoring. So, we've got three scenarios. Actually, if you go to the next one, right? This kind of pretty simple on this one, but the three scenarios we'd love to cover today.

42 "Stan DiLullo" (3173026048)
00:13:45.359 --> 00:14:14.271
Is here, right? How agentic AI does the document validation, where we insert conversational guidance for inspectors out in the field and then how the results of the process itself can be generated through automation for report summary saving time and again delivering consistency into, into the process end to end. Awesome. Awesome.

43 "Jason Masciarelli" (3295821568)
00:14:14.271 --> 00:14:33.934
Great, so we're gonna jump into the the demo on each of these three, you know, kind of capabilities. Again, this is just probably the tip tip of the iceberg in terms of the capability of the platform. So you know if folks want to see more, they can certainly reach out to you and your team, but let's let's get into the demo.

44 "Stan DiLullo" (3173026048)
00:14:33.934 --> 00:15:04.949
So this is the, this is the start, right? The assignment as we were looking at before, right? An inspector, is addressing a performance slip delivered to them. The AI document parsing is actually validating the validity of this inspection assignment, which is critical with if you don't have the right components and you execute an inspection and it's an invalid inspection.

45 "Stan DiLullo" (3173026048)
00:15:04.949 --> 00:15:24.949
That's not great. And if you look at some of the wording in there, right, the name wasn't right, the signature which is I think is great. The signature was just printed. It wasn't an authentic signature. These are things that would make this inspection invalid. AI catches that ahead of.

46 "Stan DiLullo" (3173026048)
00:15:24.949 --> 00:15:42.779
Time before investing the time into the inspection itself, right? Things that an inspector might even miss tell you the truth, right? Yeah. Now an updated authorization is delivered. It reviews that authorization.

47 "Stan DiLullo" (3173026048)
00:15:42.779 --> 00:16:02.779
And when we get there, it up now says the inspection is valid, let's go ahead. So what we've seen here, right, is that the the inspector's permit was validated to proceed the.

48 "Stan DiLullo" (3173026048)
00:16:02.779 --> 00:16:17.472
Documents were verified in order to proceed. It reduces manual checks, improves compliance, and is obviously faster at document validation than, you know, reviewing it, reviewing it back up.

49 "Jason Masciarelli" (3295821568)
00:16:17.472 --> 00:16:18.414
Yeah.

50 "Tim Miranda" (2184978432)
00:16:18.414 --> 00:16:50.731
Well, I like to hear also that, you know, it's something that just sort of happened we take for granted, but you know you're dealing with attached files, you're dealing with file processing here. Lodged manages all that, so it's securely storages your stores your files and makes them part of your business data, right? So in this case, you were able to validate the file that was uploaded, but it's just part of the data that's part of the case, so you can treat it like anything else, do validations, AI decisions on it. And that just kind of comes with the platform. So that's cool to see. Yes.

51 "Stan DiLullo" (3173026048)
00:16:50.731 --> 00:16:51.950
Yeah.

52 "Jason Masciarelli" (3295821568)
00:16:51.950 --> 00:16:59.012
All right, should we move to the next one? Yep, please. Great.

53 "Stan DiLullo" (3173026048)
00:16:59.012 --> 00:17:31.879
So on this one, right, inspection assistant, the inspector is out in the field performing the inspection and is now engaging with our AI agent for guidance, right? Chat guidance, asking, so based on the things that I've seen, the checklist I've gone through, what are the risks? What are the corrective actions that should be incorporated? What if any additional checks should I.

54 "Stan DiLullo" (3173026048)
00:17:31.879 --> 00:17:53.629
I do all delivered through our AI agent. Things that, again, from a manual process perspective, some inspectors may, be versed in the corrective actions, some very well, some maybe not so well. Some maybe better versed in additional follow UPS, but.

55 "Stan DiLullo" (3173026048)
00:17:53.629 --> 00:18:02.576
But the AI agent really helps insert those insights for intelligent inspections.

56 "Jason Masciarelli" (3295821568)
00:18:02.576 --> 00:18:03.529
Very cool.

57 "Tim Miranda" (2184978432)
00:18:03.529 --> 00:18:08.351
Awesome.

58 "Stan DiLullo" (3173026048)
00:18:08.351 --> 00:18:19.079
And and even in that one we didn't really show up but things like referring to policy documents referring to guidelines, even referring to.

59 "Stan DiLullo" (3173026048)
00:18:19.079 --> 00:18:37.412
The historical performance of that entity's inspections can be brought up and and delivered to the inspector real time on, on the scene, if you will, all through the conversational agent, right?

60 "Stan DiLullo" (3173026048)
00:18:38.248 --> 00:18:53.077
We have this exact violation before. What was the correction corrective action last time asking and having like, like I said before, a interaction with the agent on site real time? Yeah. Great.

61 "Jason Masciarelli" (3295821568)
00:18:53.077 --> 00:18:54.673
Awesome. Let me get this.

62 "Stan DiLullo" (3173026048)
00:18:54.673 --> 00:19:16.230
And we might have missed it because it is there's a little nuance in there, but I'll I'll I'll leave it, I'll just speak to it. There's a footnote in there that says these recommendations are AI generated. Please verify for accuracy, right? Responsible AI, keeping human verification in the loop.

63 "Jason Masciarelli" (3295821568)
00:19:16.230 --> 00:19:25.611
Yep yep awesome. That's great. All right, let's go through the the 3rd use case.

64 "Stan DiLullo" (3173026048)
00:19:25.611 --> 00:19:35.425
Generating the report summary, yeah.

65 "Tim Miranda" (2184978432)
00:19:35.425 --> 00:19:51.972
Yeah, I like the contrast that you're showing here as well because you have this inspector out in the field and they'd be on a mobile device and they're able to work through a set of guided steps and forms, but they're also able to chat with the case as well and even here, get a summary.

66 "Stan DiLullo" (3173026048)
00:19:51.972 --> 00:20:07.529
And and and and so this, you know, provides a summary of the inspection based on again guided workflow, process orchestration, standards and compliance, and generates.

67 "Stan DiLullo" (3173026048)
00:20:07.529 --> 00:20:22.799
What you could see would be very time consuming, right? For an inspector to, to write up, if you will, right? Sure. The AI agent summarizes the findings.

68 "Stan DiLullo" (3173026048)
00:20:22.799 --> 00:20:41.459
Inserts some of the insights so things like that conversational agent that the inspector as he was going through the inspection may have asked you can decide whether or not to have that be a part of the, of the report summary. So things like.

69 "Stan DiLullo" (3173026048)
00:20:41.459 --> 00:21:05.553
The corrective action. As he verifies what the AI agent said, he can have that incorporated into the report. The history, he could have that incorporated into the report if he or she feels, you know, it's it's should be inserted into the report summary again all auto generated. Yeah.

70 "Jason Masciarelli" (3295821568)
00:21:05.553 --> 00:21:32.654
Wow yes, I imagine from an outcomes perspective, the inspectors can now probably get more inspections done. They're more accurate, they're more compliant, and probably their job satisfaction is better as well because there's less administrative work and they can really focus on more of the, the physical world stuff and the stuff that they probably quite frankly enjoy more as opposed to having to type up these reports at the end of the day.

71 "Stan DiLullo" (3173026048)
00:21:32.654 --> 00:21:41.189
Right and and it's the intelligence into the inspection, right? And, and consistency into the inspection.

72 "Stan DiLullo" (3173026048)
00:21:41.189 --> 00:21:59.639
And the consequences of, of poor inspections because of admin are pretty significant when you talk about penalties violations, when we talk about manufacturing flaws and defects.

73 "Stan DiLullo" (3173026048)
00:21:59.639 --> 00:22:19.639
That are way down the road of, of, of assembly, if you will, right? And now you need to, it's like we we used to always say in in development, right? Finding flaws and testing is much more expensive than finding flaws right during design.

74 "Stan DiLullo" (3173026048)
00:22:19.639 --> 00:22:32.813
And that's why, you know, these, these inspection quality gates are very critical to one not only be performed but be performed accurately and with intelligence. Yeah. Awesome.

75 "Jason Masciarelli" (3295821568)
00:22:32.813 --> 00:22:39.069
Fantastic. And do we have one more or did I make a mistake? I think you made a mistake.

76 "Tim Miranda" (2184978432)
00:22:39.069 --> 00:22:42.948
Yeah, there we go. Okay, there we go. All right, and maybe I should have an inspection of my work here.

77 "Jason Masciarelli" (3295821568)
00:22:45.792 --> 00:22:51.375
Alright, so why don't we, why don't we do the wrap up here, Stan, in terms of what we've seen?

78 "Stan DiLullo" (3173026048)
00:22:51.375 --> 00:23:11.669
Sure, so it's just a summary of what we saw, right? AI is embedded directly into the workflow life cycle, performing operational tasks like document validation, report generation, and contextual guidance, real time conversational guidance for for the inspectors.

79 "Stan DiLullo" (3173026048)
00:23:11.669 --> 00:23:29.009
It determines the parameters of what is required for the inspection and required for the skills of the inspector. We didn't talk much on that, but that assignment of that 1st, that 1st one, the permission slip and assignment is also based on.

80 "Stan DiLullo" (3173026048)
00:23:29.009 --> 00:23:59.493
The inspection, the category of inspection, the availability of of the inspector and the skills of the inspector, right? Very, very crucial parts of an underqualified inspector for a very important inspection is not a good thing and and and the insights that are delivered again real time end to end through process orchestration and.

81 "Jason Masciarelli" (3295821568)
00:23:59.493 --> 00:24:13.674
Very cool. And so I guess as you look at, you know, where, where the product is and where you guys are going, maybe just talk a bit about like where you see this going for Aaseya and and kind of the impact that you can deliver on the market.

82 "Stan DiLullo" (3173026048)
00:24:13.674 --> 00:24:33.499
Yeah, I, I think as I mentioned, as I mentioned before, right, A AI is designed designed industry agnostic platform. We support a wide range of compliance, regulatory quality, health and safety.

83 "Stan DiLullo" (3173026048)
00:24:33.499 --> 00:24:55.679
Manufacturing inspection requirements, the configurable platform is is is key. And, and even, even within those different verticals, the different kinds of inspections that can be configured into the platform is also very key and and gives us a lot of flexibility, even.

84 "Stan DiLullo" (3173026048)
00:24:55.679 --> 00:25:12.359
Even customized checklists for, for a given entity or or possible. The different scoring models, approval workflows, reporting structures. Again, we didn't talk much about reporting. You can imagine with a very manually intensive.

85 "Stan DiLullo" (3173026048)
00:25:12.359 --> 00:25:30.509
Inspection program, getting centralized reporting, aggregated reporting, management reporting is a nightmare in and of itself, right? Not only for trying the effort required to get it, but the accuracy of the information.

86 "Stan DiLullo" (3173026048)
00:25:30.509 --> 00:25:49.076
Frankly is is subject to flaws. Yeah. And so with that, the business value expands beyond just inspections. There are many other use cases within the scope of possible, and we've just listed a few here just representatives.

87 "Jason Masciarelli" (3295821568)
00:25:49.076 --> 00:26:09.799
Sure, sure. Well, that's awesome. So it seems like you've you've built a lot of depth and specificity around inspections but you also have a flexible capability to work with your clients to say, address it and tailor it to their specific needs. They've got some customization capability as well. And so I think that's really a combination of, you know, your.

88 "Jason Masciarelli" (3295821568)
00:26:09.799 --> 00:26:19.812
Your expertise and, you know, what you've done with the platform and then, you know, kind of what's been locked with launchpad. And so Tim I don't know if you wanna kind of highlight some of those things here. That'd be great.

89 "Tim Miranda" (2184978432)
00:26:19.812 --> 00:26:39.379
Yeah, I mean, I love what you've done with this application, but also your vision of where you're taking your industry knowledge and experience into those other areas like you mentioned, but that's exactly what we're looking to enable with launchpad, right? So starting with your AI blueprint, you're able to take your ideas, your domain knowledge, your like your company's.

90 "Tim Miranda" (2184978432)
00:26:39.379 --> 00:26:59.379
Extensive experience with this and turn that into an actual design and the starting of an application that you can demonstrate before you even started building and hit the ground running with your development. And the cool part is like you've mentioned before, right? As more and more customers subscribe to your app, their users start using it. The platform itself.

91 "Tim Miranda" (2184978432)
00:26:59.379 --> 00:27:11.210
If we'll scale up and out and meet your demand so you're really able to focus on like what are the outcomes you're looking for, what work does it need to get done on their behalf? And you're able to do all that with ALOC AI.

92 "Stan DiLullo" (3173026048)
00:27:11.210 --> 00:27:28.039
Yeah, and and and often, you know, people, people ask, how quickly, right? How quickly can we get started? The AI driven blueprinting and low code configuration really helps us deploy this core solution in just a few weeks or or.

93 "Stan DiLullo" (3173026048)
00:27:28.039 --> 00:27:56.112
At least the blueprint within a few weeks, we can go live in in approximately ten weeks. Blueprinting one to two weeks configuration, three to six weeks and another, you know, I don't know, three, well I do know actually another three to four weeks for testing and deployment. So overall average, we can be up and running full application in ten weeks.

94 "Tim Miranda" (2184978432)
00:27:56.112 --> 00:27:58.672
It's great to hear. Awesome. Awesome.

95 "Jason Masciarelli" (3295821568)
00:27:58.672 --> 00:28:10.199
Good stuff. Good stuff. So, so for folks that wanna get started and and want to work with you folks around the solution and and need help with this capability.

96 "Jason Masciarelli" (3295821568)
00:28:10.199 --> 00:28:30.199
You know, certainly go to your website, go to sales at.com. If you're looking to know more about, you know, how Launchpad can help build solutions that you want to build and take to market, you know, I'd encourage you to go to launchpad.io sign up for a Launchpad Explorer, which is a free account, and then reach out to us and we're happy to have a.

97 "Jason Masciarelli" (3295821568)
00:28:30.199 --> 00:28:50.199
Conversation around particular use cases and so on. And you can just do that through contacting sales. So, you know, we're, you know, Sayya and Launchpad have been working together closely for the last several years and, you know, I'd Saya's been a long time partner of Pega, so we're excited to have you and glad to have you on stand to represent Essaya and help take us through.

98 "Jason Masciarelli" (3295821568)
00:28:50.199 --> 00:29:16.172
To the story today. So before we get into questions, I just wanted to highlight, again, the next couple of events we've got, so we've got the modernizing payment investigations coming up in 24 March and then a launchpad live for our community on 26 March, so make sure you register for those. And then with that we will wrap up. I appreciate everybody coming and joining us today and Lena, do you want to guide us through some of the questions that have come in as we've been talking.

99 "Lena Lisitskaya" (154817024)
00:29:16.172 --> 00:29:33.835
Absolutely. So there I have a couple questions about launchpad, which I'll start with, Tim, maybe you can speak to this one. When you build on launchpad, does it support try and different AI models?

100 "Tim Miranda" (2184978432)
00:29:33.835 --> 00:29:51.799
Yeah, absolutely. So whether you're using generative AI as part of a step in your workflow where you're like, like they showed with summarization or analysis or validation, we we let you choose from many different models, whether you're looking at Amazon's offerings, OpenAI.

101 "Tim Miranda" (2184978432)
00:29:51.799 --> 00:30:06.793
Anthropic, Google, and so you can really pick and choose, but you don't have to bring your own endpoints, you don't have to bring your own keys or anything. We handle all that for you. And the same with the agentic processing as well.

102 "Lena Lisitskaya" (154817024)
00:30:06.793 --> 00:30:23.496
Got it. Thank you. And Stan I'll go to you. Many vendors claim AI capabilities. What makes a hawk AI different? And Jason Tim, you can jump in as well.

103 "Stan DiLullo" (3173026048)
00:30:23.496 --> 00:30:40.409
Excuse me. One of the, you know, the conversational aspect of it and the fact that it's embedded, the agents are embedded in the workflow its itself. And so it's, it's more than excuse me, it's more than.

104 "Stan DiLullo" (3173026048)
00:30:40.409 --> 00:30:56.579
Just the LLM, it's actually really helping to automate and and productize if you will, the, the, the orchestration of the process itself. And, and really.

105 "Stan DiLullo" (3173026048)
00:30:56.579 --> 00:31:14.849
That's, that's the big benefit is bringing insights into the hands of the, of the inspectors. Again, as we mentioned, it's a very manual process and and leveraging the agents in the AI for for the automations as well as.

106 "Stan DiLullo" (3173026048)
00:31:14.849 --> 00:31:34.344
The historical nature of of of entities inspections is, is critical and and with that embedded in the workflow itself, it really brings a lot of value to, to, to the process.

107 "Lena Lisitskaya" (154817024)
00:31:34.344 --> 00:31:41.152
Makes sense. Stan, what happens if AI makes a mistake?

108 "Stan DiLullo" (3173026048)
00:31:41.152 --> 00:32:05.489
I did mention before, I was glad I footnoted that out of the out of I think it was 2nd scenario, right? The transparency and auditability of the AI insights is built into, is built into the platform. It there is human in the loop.

109 "Stan DiLullo" (3173026048)
00:32:05.489 --> 00:32:21.779
Processing. So it's not it's not just a given, right? Like I said, there's the transparency human in the loop framework for quality and insurance, we ensure.

110 "Stan DiLullo" (3173026048)
00:32:21.779 --> 00:32:40.871
That there is human aspects that intervene especially in high risk or statistically significant scenarios where the AI needs to be, you know, double, double checked and, and, you know, trained for future learning as well.

111 "Lena Lisitskaya" (154817024)
00:32:40.871 --> 00:32:57.079
Makes sense, yeah. Complementing the human. And the next question is related as well. Is it feasible to replace traditional manual inspections with an AI agent inspector for primary oversight?

112 "Lena Lisitskaya" (154817024)
00:32:57.079 --> 00:33:20.495
While maintaining a human in the loop framework for quality assurance. Specifically, I'm looking to leverage intelligent sampling to ensure humans only intervene in high risk age case or statistically significant scenarios, a step moving in the direction of autonomous enterprises.

113 "Stan DiLullo" (3173026048)
00:33:20.495 --> 00:33:27.314
I think the answer is the answer is is yes. That's.

114 "Jason Masciarelli" (3295821568)
00:33:27.314 --> 00:33:57.329
It almost sounds like the question's asking about like deflection of inspection of physical inspections and maybe doing some statistical analysis and predictive stuff up front and then having people focus on higher risk things. And so I'm not sure if you folks have built that already out, but I I know that, you know, through your expertise and what Launchpad could do, we could certainly build that in and that would be certainly like that could be done. And a lot of that's gonna be based on the context of.

115 "Jason Masciarelli" (3295821568)
00:33:57.329 --> 00:34:17.329
The particular industry or the the specific client and the space that they're in because there maybe legal requirements or compliance requirements that actually require a physical inspection and there maybe situations where you could do some statistical deflection that says, hey, look, this is a very low risk threshold, and so we can actually skip these and then focus our.

116 "Jason Masciarelli" (3295821568)
00:34:17.329 --> 00:34:21.614
Are people on other higher risk things that are, you know, meeting certain criteria.

117 "Stan DiLullo" (3173026048)
00:34:21.614 --> 00:34:38.369
Yeah, and we and and great great point. And, and we did in some scenarios use AI models to examine product defects using, you know, smart AI driven sampling to identify quality issues.

118 "Stan DiLullo" (3173026048)
00:34:38.369 --> 00:35:02.193
Products flaws and things of things of that nature. But as, as you mentioned, the th the scenario, the inspection itself dictates the amount of AI versus human intervention or human oversight required. Yeah.

119 "Lena Lisitskaya" (154817024)
00:35:02.193 --> 00:35:12.711
Excellent. And let's finish with this one. So, how does Launchpad manages data privacy and protection?

120 "Jason Masciarelli" (3295821568)
00:35:12.711 --> 00:35:14.633
Timmy you want to take that?

121 "Tim Miranda" (2184978432)
00:35:14.633 --> 00:35:34.039
Yeah, so that's that's really a cruise kind of baked into the plat the launchpad platform so there's a lot of layers to it, right? Whether it's role based access control, attribute based access control, which you can build, configure right into your application, but also the data isolation.

122 "Tim Miranda" (2184978432)
00:35:34.039 --> 00:36:02.990
And tenant isolation, you know, the platform is built from the ground up to be multi tenant, so we're very robust ways to keep data from different customers in its own places as well as compliance with different industry regulations around that. But we, we looked to kind of take that off the table, but also give you the flexibility to configure the access and other things that you need based on your business requirements, so.

123 "Jason Masciarelli" (3295821568)
00:36:02.990 --> 00:36:03.794
Great.

124 "Stan DiLullo" (3173026048)
00:36:03.794 --> 00:36:06.791
And and the scalability.

125 "Jason Masciarelli" (3295821568)
00:36:06.791 --> 00:36:24.599
Yeah, yeah, auto scaling certainly is a is a huge one. And so, you know, imagine bring bringing a product to market, you know, Stan, you don't need to worry about, you know, and your team, you know, scaling to thousands or millions of users, you just focus on selling the product as much as you can.

126 "Jason Masciarelli" (3295821568)
00:36:24.599 --> 00:36:44.599
And then also in terms of, you know, agility and changing and addressing new market needs, you've got that flexibility to kind of evolve to the market needs and add in more and more AI capabilities as well. So, awesome, this was a great discussion. San, thank you so much again for, for joining us on the call and and walking us through your story and I.

127 "Jason Masciarelli" (3295821568)
00:36:44.599 --> 00:37:04.599
About what you've built with Launchpad and the exciting things you're doing with inspections. Tim, thanks for joining as well and and weighing in on some of the more product specific things. And with that we'll wrap up. Thank you Lena for making sure this was a success today and and getting through all the questions. So, have a great day and.

128 "Jason Masciarelli" (3295821568)
00:37:04.599 --> 00:37:07.053
We'll see you next time. Thank you.

129 "Tim Miranda" (2184978432)
00:37:07.053 --> 00:37:07.753
Thanks.

130 "Stan DiLullo" (3173026048)
00:37:07.753 --> 00:37:17.664
Appreciate it.