Legacy System Integration: Modernization with AI Innovations

WEBVTT
00:00:00.161 --> 00:00:06.663
Hi and welcome back to the Enterprise Alchemists, a podcast from SnapLogic for enterprise architects and people interested in architecture.
00:00:06.663 --> 00:00:09.230
I'm joined by my co-host, as ever, Guy Murphy.
00:00:09.230 --> 00:00:09.791
Hey, Guy!
00:00:09.791 --> 00:00:10.973
Good morning Dominic!
00:00:11.621 --> 00:00:19.685
And this week we also have with us Simona Todoroska from IWConnect, our valued partner, and I'm not saying that just to be polite.
00:00:19.685 --> 00:00:21.330
Hi, Simona, Thanks for joining us.
00:00:22.260 --> 00:00:24.545
Hi Dominic, Hi Guy, Thank you guys for having me.
00:00:28.676 --> 00:00:29.338
Absolute pleasure.
00:00:29.338 --> 00:00:33.393
Do you want to quickly introduce what you do at IWConnect for our listeners, who may not be already aware?
00:00:34.197 --> 00:00:34.780
Yes, indeed.
00:00:34.780 --> 00:00:37.265
So I'm Simona.
00:00:37.265 --> 00:00:46.929
As Dominic already introduced me, I'm a solution manager slash architect in IWConnect, but as well managing our European side of the business.
00:00:46.929 --> 00:01:03.969
If you will, that means indeed wearing too many hats, but I'm kind of bridging the gap between technology and business, which is basically trying to enable technology for resolving real world problems for our clients.
00:01:03.969 --> 00:01:05.531
So yeah, that's me.
00:01:06.353 --> 00:01:06.894
Fantastic.
00:01:06.894 --> 00:01:19.968
And yes, if you've been at the SnapLogic events at the end of last year, then you'll have seen the IWConnect booth, and if you missed out talking to them, then it's well worth checking out their website.
00:01:19.968 --> 00:01:21.405
We'll put that in the show notes.
00:01:21.405 --> 00:01:22.662
Guy
00:01:22.662 --> 00:01:23.727
, why don't you lead us off?
00:01:23.727 --> 00:01:27.347
Because you had some questions you particularly wanted to talk about with Simona.
00:01:28.251 --> 00:01:31.302
Yeah, no, absolutely so for our listeners.
00:01:31.302 --> 00:01:36.948
Sometimes I probably keep on referring to the past, and this podcast is not going to be different.
00:01:36.948 --> 00:01:50.254
Just quickly wanted to touch upon a couple of trends that we're seeing in the marketplace that aren't hot and sexy and aren't always going to directly link to AI, though actually that is changing.
00:01:50.956 --> 00:02:07.000
If you look at about 10-12 years ago, with the rise of the iPaaS platforms in the marketplace, we saw a very interesting dynamic and I was there with some of the early platforms in that space in which the vendors themselves actively avoided dealing with legacy.
00:02:07.000 --> 00:02:17.588
It was for lots of good reasons, sometimes commercial reasons and sometimes the scale and the complexity of dealing with this, and that was fine.
00:02:17.588 --> 00:02:33.207
That allowed the ipass marketplace to accelerate at pace into the marketplace, develop new hybrid architectures, but it kind of left an iceberg in many enterprise architecture strategies because of this disinterest.
00:02:33.207 --> 00:02:56.290
And what's been interesting is in the last two or three years we've definitely seen the rising concern of this from our customers in SnapLogic, because basically many of these platforms are coming to end a life or they're becoming fundamental bottlenecks to innovation, because often they are sitting there with thousands, tens of thousands.
00:02:56.290 --> 00:03:05.409
We even have one client in North America, with hundreds of thousands of integration patterns sitting inside of, effectively, black box legacy.
00:03:05.409 --> 00:03:10.890
Simone, is this something that you're starting to see happen around your clients?
00:03:11.879 --> 00:03:27.947
Yes, yes, indeed, you are pretty much on point, Guy, about the statement of previously potentially vendors avoiding migration, and there is a reason for it, right, it's basically one of the hardest projects are the migration projects.
00:03:27.947 --> 00:03:36.385
You usually get the questions such as why fix something that is already working or it's not broken if you will.
00:03:36.979 --> 00:03:38.544
If it ain't broken, don't mess with it.
00:03:38.544 --> 00:03:39.806
That's a safe rule.
00:03:40.328 --> 00:03:41.591
Yeah, that's right, it works.
00:03:41.591 --> 00:03:46.201
No one knows how, no one knows behind the scenes what's going on, but it it works.
00:03:46.201 --> 00:03:51.663
No one knows how, no one knows behind the scenes what's going on, but it just works, and usually it's.
00:03:51.663 --> 00:03:55.425
How do you win on that one and basically try to migrate out of it?
00:03:55.425 --> 00:04:07.491
And I've recently watched the BlackBerry movie, which is basically a migration, if you will, from, let's say, a smartphone that was working right.
00:04:07.491 --> 00:04:13.253
Nothing was wrong with BlackBerry, but when iPhone came into the picture, it completely changed the world.
00:04:13.253 --> 00:04:24.218
So basically it's kind of a similar comparison and usually what I ask our customers is when they say, no, but that works, that works well.
00:04:27.779 --> 00:04:30.033
I ask them a few questions that pretty much they should ask themselves, if you will.
00:04:30.033 --> 00:04:35.086
A few of those are basically is the solution really future ready?
00:04:35.086 --> 00:04:37.125
Is it really scalable?
00:04:37.125 --> 00:04:42.365
Is it according all of the security standards that you need to have?
00:04:42.365 --> 00:04:48.992
Is your current integration platform just working?
00:04:48.992 --> 00:04:52.870
Or basically it's driving your business forward?
00:04:52.870 --> 00:05:00.206
So they may be missing a lot of opportunities just because of you know this works.
00:05:00.206 --> 00:05:01.689
It doesn't really matter.
00:05:01.689 --> 00:05:09.261
But on the other side, I completely understand that for a manager it's very hard to take a migration project First.
00:05:09.261 --> 00:05:10.165
Who will fund it.
00:05:10.165 --> 00:05:13.370
It works, so let's just don't touch it.
00:05:13.370 --> 00:05:17.028
Also, a lot of people left.
00:05:17.028 --> 00:05:26.615
Documentation may not be there, so figuring out the as-is state is the hardest part of it.
00:05:26.615 --> 00:05:29.863
And then the migration itself happens afterwards.
00:05:29.863 --> 00:05:34.874
But figuring out the as-is state has been a lot of challenge so far.
00:05:35.939 --> 00:05:38.685
And that is precisely it, isn't it.
00:05:38.685 --> 00:05:45.949
You know that it's working, but if it ever breaks, then you don't have the people who originally built it available.
00:05:45.949 --> 00:05:51.872
You don't have good docs, almost certainly, so that would be bad, but that's hypothetical.
00:05:51.872 --> 00:05:54.504
That hasn't happened yet and until it happens.
00:05:54.504 --> 00:05:55.548
Cross your fingers and hope.
00:05:55.548 --> 00:05:57.072
Is that about the size of it?
00:05:58.642 --> 00:05:59.363
Yeah, pretty much.
00:05:59.363 --> 00:06:10.689
Yes, you're, you're completely right, and but when it happens, it's a complete catastrophe for the whole company because they're relying on it quite a lot.
00:06:11.961 --> 00:06:15.151
As our listeners can hear, we're all in violent agreement with each other.
00:06:15.151 --> 00:06:32.129
I think one of the accelerators, especially in EMEA I'm starting to see, is finally the shift in the SAP install base, because with the SAP decision to extend support, they kind of kicked their own migration strategy down the line.
00:06:32.129 --> 00:06:59.841
But now I'm seeing that manufacturing retail in store base across Europe is starting to move to their new cloud platforms and this is forcing a clear look inside of the legacy processes and suddenly the legacy integrations become a critical issue, because I can think of at least two accounts I've spoken to in the last six months where they it's not just integration patterns that have embedded in these systems, but actually business logic.
00:06:59.841 --> 00:07:06.173
And again, at that point your migration is even more complex than just point to point integration.
00:07:06.173 --> 00:07:13.923
Simone, so when you work with your clients you talked about the questioning how do you go about it?
00:07:13.923 --> 00:07:26.038
Because I'm sure you're not just sitting there with an army of millions hand analyzing everything, because that's been tried and that's been failing for the last 25 years in the marketplace.
00:07:27.079 --> 00:07:27.800
Yeah, that's right.
00:07:27.800 --> 00:07:45.713
Indeed, what we are trying to do as an approach and as a project, usually approaching immigration is basically use or reuse our knowledge that we've gained so far.
00:07:45.713 --> 00:07:54.168
What I mean by it is not really the people knowledge, but as well automation behind the scenes that we've built through the years.
00:07:54.168 --> 00:08:02.300
So let me give you an example of one successful migration project that we have done previously.
00:08:02.300 --> 00:08:06.911
So it was for a big telco company.
00:08:06.911 --> 00:08:12.613
Those guys had relying on Cisco Service Grid.
00:08:12.613 --> 00:08:19.730
It's an old legacy integration platform for integrating ITSM integrations.
00:08:19.730 --> 00:08:39.407
So basically those guys used the Cisco Service Grid to integrate their clients onto their ITSM, where they will open them incidents, cases, problems I mean the standard ITSM story.
00:08:39.407 --> 00:08:55.412
Now, that was working for more than 10 years for them, and they've implemented it with the multiple consultancies, companies, different vendors and whatnot, and documentation of course wasn't there.
00:08:55.919 --> 00:09:05.813
Nothing was really existing, and figuring out the as-is state was the basically hardest part to do, as I mentioned.
00:09:05.813 --> 00:09:18.700
So reverse engineering was one of the first thing that we kind of do, but not with people but with technology.
00:09:18.700 --> 00:09:31.894
So basically, what we have leveraged is acceleration of migration, which we are trying to kind of build as a strategy.
00:09:31.894 --> 00:09:48.004
What that means is basically, we build these accelerators, or speeding up tools that are speeding up the migration, which are actually analyzing the as-is state.
00:09:48.004 --> 00:10:02.494
Give you the as-is state as a report, meaning what is there, how complex are your integrations, how much of those can be reused, how much is future-ready architecture in your new platform?
00:10:02.860 --> 00:10:09.866
We were migrating this to SnapLogic, so basically the analysis phase was almost fully automated.
00:10:10.708 --> 00:10:25.971
It gave us the report of what can be automated, what really needs to be rebuilt by manual work and what needs to be changed from architecture perspective, and also gave us documentation of the as-is state.
00:10:25.971 --> 00:10:44.094
Once we've kind of deployed or used the accelerator, we already kind of did half of the job and then we kind of planned the migration, which is always, you know, a big bank, is not something that you would like to do.
00:10:44.094 --> 00:10:52.373
This was working for more than 10 years, so let's make sure that it will work once you migrate as well.
00:10:52.373 --> 00:11:36.774
So we planned the phases based on the report that this accelerator gave us and basically we achieved the numbers which, if we compare it to the traditional way of migrating, which we want to avoid as to your point, guy way of migrating, which we want to avoid To your point, guy we got numbers such as 65, I think it was the actual number of speed of delivery that we've overcome with this acceleration, and that's an important factor to point out that this analysis phase is actually the most important part of the whole operation.
00:11:36.940 --> 00:11:39.807
It's not the implementing the new thing.
00:11:39.807 --> 00:11:40.770
That part's easy.
00:11:40.770 --> 00:11:45.971
And to underline that point, I recently dealt with the aftermath of a failed migration.
00:11:45.971 --> 00:11:59.649
Someone else not IWConnect, I hasten to add had been into this customer and had migrated from Informatica Power Center to SnapLogic, but what they'd done was just a straight up lift and shift.
00:11:59.649 --> 00:12:21.089
They didn't really think about the semantics of how things are done in these two rather different platforms, and so they'd replicated exactly what had been done in Informatica, but that was far from being the best way of doing things in SnapLogic, so they were having terrible performance problems and we ended up having to redo almost all of that work.
00:12:21.089 --> 00:12:26.826
So that analysis phase is so crucial to getting a good result at the end.
00:12:28.008 --> 00:12:55.052
Yeah, you are completely right, Dominic, and indeed we have been dealing with Informatica Power Center as well, migrations to different integration platforms, and these accelerators that we've built, including the Informatica Power Center 1, actually are overcoming that gap, are overcoming that gap when I say this is basically analysis phase.
00:12:55.111 --> 00:13:19.735
Yes, figuring out the as-is state as the most important part and then deciding basically what is a lift and shift, really, what you can do, such as, I don't know, maybe mappings, converting into the respective, let's say, mappings in the new platform, but the actual architecture, or compressing the flows or making it future ready for the platform that you've chosen to use.
00:13:19.735 --> 00:14:04.986
Going forward is the most important part and for it, in this accelerator, we do use AI and Gen AI as part of these accelerators where, let's say, with AI, we do this kind of automated system analysis, if you will, and mapping the future integration architecture, and with Gen AI, we basically get the documentation of the as-is state and potentially getting the migration strategy as a suggested one or a project plan, if you will.
00:14:04.986 --> 00:14:07.010
So that's how we do it.
00:14:09.722 --> 00:14:12.510
So, Simona, you've been talking about the external tools and your processes.
00:14:12.510 --> 00:14:21.470
Have there been any aspects of the SnapLogic platform you found useful as obviously a candidate platform for dealing with these sorts of strategic projects?
00:14:23.234 --> 00:14:25.562
Yes, indeed.
00:14:25.562 --> 00:14:40.773
I mean I've been working with SnapLogic for, I think around a decade now, and SnapLogic when I started and SnapLogic now it's a completely different platform which I always want to say we evolved together.
00:14:40.773 --> 00:15:05.929
So I mean I have been working with Tipco, mulesoft, snaplogic and all of those platforms.
00:15:05.929 --> 00:15:17.398
You name it citizen integrator approach as something that really is in the sense of the actual word.
00:15:17.398 --> 00:15:36.926
When I say this is basically I've did the comparisons of multiple platforms, I've did, let's say, sometimes a customer doesn't know what platform they would like to use and usually we do the matching.
00:15:37.460 --> 00:15:40.605
Of course, not for every use case.
00:15:40.605 --> 00:15:45.394
Every platform is really applicable.
00:15:45.394 --> 00:16:30.073
But once you do the let's say comparisons and what are the strengths and what are the really incorporating another tool or code or anything to kick off a project, even a migration one, because in these accelerators, a lot of behind the scenes is SnapLogic itself.
00:16:30.073 --> 00:16:32.804
Behind the scenes is SnapLogic itself.
00:16:32.804 --> 00:16:38.168
When I say this is basically a lot of these accelerators contain SnapLogic pipelines that are doing the conversion.
00:16:38.168 --> 00:16:43.764
Not every tool can do that and not it's to spin up.
00:16:43.764 --> 00:17:00.785
Basically, these accelerators with SnapLogic so far in my experience has been the most powerful, let's say, thing to do going forward and with it the client doesn't really need to install anything.
00:17:00.785 --> 00:17:13.368
They don't need to do anything on their legacy platform or servers, but actually just, let's say, potentially use SnapLogic as their new platform.
00:17:13.368 --> 00:17:23.174
So that's I will take as a major win usually in all of my migration projects using SnapLogic.
00:17:24.060 --> 00:17:36.336
I love that you jumped straight to the ease of use there, because sometimes when we talk about the no-code nature of the snap logic platform, people assume that that means it's only suitable for simple use cases.
00:17:36.336 --> 00:17:39.163
But you're actually coming in as platform experts.
00:17:39.163 --> 00:17:41.730
You know this, this platform, deeply.
00:17:41.730 --> 00:17:47.752
You're trying to accomplish pretty complex tasks and still it's the first thing that you jump to is there the ease of use.
00:17:47.752 --> 00:17:49.743
It just makes it really easy to get stuff done.
00:17:49.844 --> 00:18:22.332
You think about it, you do the analysis, as we were talking about before, but when you're trying to to do the work and get people to adopt it, that's the important factor yes, indeed, and uh, especially for for migration projects, uh, as such, uh, the let's say the, the client where you are trying to do the migration will always ask you but I have a team that already knows this platform how would I do the transition and training?
00:18:22.332 --> 00:18:38.751
And that's really one of the, let's say, important stuff with Simplogic, because the training and the intuitiveness of it really really comes at ease for almost every customer that we had so far.
00:18:40.759 --> 00:18:40.921
Cool.
00:18:40.921 --> 00:18:42.606
Thank you very much for sharing that.
00:18:44.131 --> 00:19:02.006
So you mentioned before that you used AI and Gen AI in the migration process and the analysis phase, but people who've seen the IW Connect booth at events like SnapLogic Zone Integrate will also have noticed that you talk about Pete the virtual assistant.
00:19:02.006 --> 00:19:03.977
He's also all over your LinkedIn.
00:19:03.977 --> 00:19:10.771
Do you want to explain a little bit more about Pete and Alex and the agents that you built and where they come in?
00:19:11.875 --> 00:19:15.481
Yes, yeah, and thank you for the question, Dominic.
00:19:15.481 --> 00:19:26.301
So, indeed, Pete and Alex are agents that we have built using AI, gen, ai, and I mean kind of our experience so far.
00:19:26.301 --> 00:19:31.538
How we come up with these agents is basically resolving an issue for us.
00:19:31.538 --> 00:20:01.099
We had documentation was always like, as we all know, developers doesn't really like to write documentation, doesn't really like to do any how to say maintenance of it, if you will and we kind of tried to bridge that gap and save that time, but as well have an updated documentation all the time.
00:20:01.099 --> 00:20:05.778
Going back to the migration, this is really important.
00:20:05.778 --> 00:20:11.241
So we've built initially PEAT as an agent.
00:20:11.241 --> 00:20:14.874
It wasn't really taught as such, it was built.
00:20:14.874 --> 00:20:20.516
It was even called, I believe, IWDG or something which was IW Document Generator.
00:20:21.659 --> 00:20:22.741
Very catchy!
00:20:23.471 --> 00:20:50.078
Very catchy indeed, and basically it was something like a pet project for us to help internally, for us to not write documentation, to help our customers not, let's say, billing hours, if you will, to them for writing documentation, and try to automate that as much as possible.
00:20:50.078 --> 00:21:00.316
And once we've shared that with multiple clients, this is how we do documentation everyone was like can we have that?
00:21:00.316 --> 00:21:01.057
How do you do it?
00:21:01.057 --> 00:21:01.680
Can we have it?
00:21:01.680 --> 00:21:10.640
So then we kind of grouped and thought about it how we can productize this, let's say, project.
00:21:10.640 --> 00:21:22.280
So that was a few years ago when things were a little bit different and AI wasn't really that much of a hype.
00:21:22.280 --> 00:21:24.857
But we've even done now.
00:21:24.857 --> 00:21:51.630
We've been using it the traditional one and trying to basically have something that will keep your documentation up to date, to make your support teams aware of what is going on within your integration layer and having the documentation always in front of you.
00:21:51.630 --> 00:21:57.526
It works, for it documents everything on confluence.
00:21:57.526 --> 00:21:59.832
It works with SharePoint as well.
00:21:59.832 --> 00:22:04.601
You can also download PDFs if that's your way of documentation.
00:22:04.601 --> 00:22:08.137
But that's not just.
00:22:08.137 --> 00:22:10.286
That's not just it.
00:22:10.346 --> 00:22:13.317
With Pete Now we have added multiple features.
00:22:13.317 --> 00:22:23.558
One of them is a code review, and what we're trying to make with Pete is basically making your compliance expert compliance expert, if you will.
00:22:23.558 --> 00:22:42.615
So the code review as well was kind of a gap that we've tried to bridge with Pete, because it was always a challenge of how you do a code review, what is the process, how that works, and with every client that's different.
00:22:42.615 --> 00:22:49.925
It's conversations and meetings and whatnot, and with this, basically you have it automated.
00:22:49.925 --> 00:22:55.163
It analyzes the integrations built in SnapLogic.
00:22:55.163 --> 00:23:13.615
It gives you a report of your environment, health, how that looks like at the moment, what you need potentially to fix, what is a warning and what is a complete error, if you will, something that you want to allow in production.
00:23:13.615 --> 00:23:29.282
And our whole idea is basically you just put the API of it in your CI/CD and it kind of blocks a deployment if your code is not on the level where you want it to be.
00:23:30.852 --> 00:23:32.739
I love how we've come full circle.
00:23:32.739 --> 00:23:46.390
We said that the problem statement for many customers of yours for migrations is that they don't have any documentation of what's been built, and so the first thing they're going to build with the Gen AI features is something to generate docs, so you don't have that problem in the future.
00:23:46.390 --> 00:23:57.380
So you're not just solving the immediate problem but setting the customers up to have to rinse and repeat a few years into the future, but you're actually giving them something that's going to help them structurally.
00:23:58.589 --> 00:24:04.862
Yes, yes indeed, and it's basically nothing that you will need to do with it.
00:24:04.862 --> 00:24:09.079
It's basically a click of a button, or even you can.
00:24:09.079 --> 00:24:17.652
You can just put it as an automation within your layer, so it's not like you have a bot and you ask it, but it's more like a.
00:24:17.652 --> 00:24:27.516
It's a, it's an agent behind the scenes that does that work for you and that's something else I'm seeing as the gen ai market matures.
00:24:27.576 --> 00:24:35.820
Initially, everyone wanted to build a chatbot and that was all everyone was thinking about, and now we're seeing more and more of these features as kind of just in the background.
00:24:35.820 --> 00:24:37.676
They're working automatically.
00:24:37.676 --> 00:24:41.121
They can be called by an API so you can integrate them within workflows.
00:24:41.121 --> 00:25:02.415
It's just becoming seamless and that is, I think, a much more mature approach, because lots of things it's exciting once to talk to a chat bot, but you wouldn't want to spend your whole day in a chat interface asking very verbose questions, so you just, you know, figure it out, Machine gets on with it.
00:25:03.691 --> 00:25:12.544
Yeah, you're completely right, and Alex, the other agent, is basically doing a similar job from that standpoint.
00:25:12.544 --> 00:25:37.392
However, that's a little bit more on the almost let's say almost every company uses Scrum Masters, follows an Agile approach of development and what Alex does is basically playing Scrum Master in a way where it automatically does the storytelling for you.
00:25:37.392 --> 00:26:13.099
It gives you a complete spin-up of all of the epic stories for anything that you are trying to build and basically what it does is learns from your environment and your platform so it can be better and better every day and have, let's say, create for you all of the stories boards in JIRA or Azure DevOps so you can speed up the whole development process.
00:26:14.570 --> 00:26:24.759
But speaking of things that do perhaps work better as a chatbot, there's also the other feature that you built, the natural language to SQL query generation.
00:26:24.759 --> 00:26:26.797
Do you want to talk a little bit about that?
00:26:26.797 --> 00:26:35.550
Because SQL queries, once you get beyond the trivial, that's also a big migration pain, and so it's maybe worth talking about those for a minute as well.
00:26:36.372 --> 00:26:37.614
Yeah, yeah, sure, I mean.
00:26:37.614 --> 00:26:44.832
Worth mentioning is all of these things we are implementing and continue to implement.
00:26:44.832 --> 00:27:02.855
This is just a few examples that we have done right, and NL2SQL also came as, let's say, a resolution of a problem that we have, or our customers have, and we're trying to make their lives easier.
00:27:02.855 --> 00:27:25.893
What that does, indeed, is a chatbot in a way, where you ask it questions and, directly behind us, it generates you a complex queries, store procedures or, let's say, best way for querying your data sets.
00:27:25.893 --> 00:27:36.895
What that does is basically you have a simple interface of a chatbot and you ask it I don't know.
00:27:38.151 --> 00:27:55.803
You are, let's say, a retailer where you have different products and you want to get the info of which product generates you the most revenue, and basically that's what you ask which product does give me the most revenue in the last month?
00:27:55.803 --> 00:28:16.800
So, behind the scenes, it does querying of your data warehouse, of all your info, where you have all the data, and gives you an output of basically a text where it says I don't know, these sneakers are the ones that generate you the most revenue.
00:28:16.800 --> 00:28:50.690
So, behind the scenes, imagine how big of a query is that merging multiple tables, depending how you have your data warehouse built, of course, but this is basically used for our customers that want the business users to get their data really fast without really needing to go to the data team and ask them to make them a report or whatnot.
00:28:50.690 --> 00:28:58.243
With these bots, they can easily figure out what's going on within their enterprise.
00:28:59.150 --> 00:29:03.020
And going back to empowering the users and citizen integration once again.
00:29:03.020 --> 00:29:27.941
And also I love that you've mentioned two AI and Gen AI use cases, and both times, almost your first statement was we had a problem and we wanted to solve it this way, and I hear far too many people saying we wanted to do something with AI and we went looking for an excuse to use the cool tech, and those tend not to end particularly well.
00:29:29.650 --> 00:29:32.982
I was shoehorning the tech into the use case when it should be very much the other way around.
00:29:32.982 --> 00:29:35.875
You start with the use case, start with the end in mind.
00:29:35.875 --> 00:29:40.531
We want to make it easier to generate documentation or sql queries, or what have you.
00:29:40.531 --> 00:29:52.258
This ai tech looks like it would solve the problem, but you have to have the understanding of the problem first yeah, correct, correct, you're completely right and problems are usually repeating.
00:29:53.098 --> 00:30:09.063
For most of the customers, I mean, no matter in which industry or what technology they use, problems are very, very similar and once we see, let's say, a recurring problem, we are trying to resolve it and trying to automate it.
00:30:09.063 --> 00:30:15.750
Before the AI world, it would be an automation some of the things and you would call it automation.
00:30:15.750 --> 00:30:19.339
Now, everything is AI, which not necessarily.
00:30:19.339 --> 00:30:19.961
I agree.
00:30:19.961 --> 00:30:33.173
You need to kind of make the differentiation what is really an automation and where you would use AI, what is really an automation and where you would use use ai?
00:30:33.173 --> 00:30:42.250
Because even before ai, automation was there and we've been leveraging it a lot to to overcome a lot of problems love it, the analysis phase.
00:30:42.570 --> 00:30:56.046
Just as we said in the beginning, it's always that it always comes down to analyzing the problem, understanding really how to solve it yeah, yeah, that's right I think that's a great place to leave it.
00:30:56.145 --> 00:30:58.295
So thank you so much for your time, simona.
00:30:58.295 --> 00:31:01.040
This has been an interesting conversation for me.
00:31:01.040 --> 00:31:03.953
I hope the audience also found it interesting.
00:31:03.953 --> 00:31:24.352
As a reminder, if someone sent you this as a standalone, it's a series that you can subscribe to look for the Enterprise Alchemists, wherever good podcasts are downloaded, and there is also a companion website that's under the Integration Nation forum where you can find the full transcripts and also the comment thread.
00:31:24.352 --> 00:31:30.917
If you have any thoughts, we are always eager to hear from our listeners, but for now, thank you once again, simona.
00:31:31.778 --> 00:31:34.201
Thank you, thank you guys, thank you for having me.
00:31:35.060 --> 00:31:36.041
And thank you to our listeners.
00:31:36.041 --> 00:31:40.726
This concludes this season of the Enterprise Alchemists, but we will be back soon.
00:31:40.726 --> 00:31:41.507
Watch this space.
00:31:41.507 --> 00:31:46.285
Until then, thank you once again.