Future-Proofing Your Developer Career in the Age of Agentic AI with OutSystems - Vivek Ganesh

This talk provides an overview of OutSystems, a low-code platform that has been operating globally for 25 years and in India for the last six years. The speaker emphasizes the significant developer community, with nearly one million developers worldwide utilizing the platform for application development, particularly in sectors like banking, financial services, and manufacturing. The session highlights the importance of integrating generative AI and agent-based solutions into the application development lifecycle, mentioning the need for governance and quality assurance due to the increased reliance on AI-generated code. The speaker also shares a success story from Griham Housing Finance, illustrating how OutSystems has streamlined their loan approval process to be paperless and significantly faster, benefiting their customers. The discussion concludes by inviting developers to engage with OutSystems through upcoming conferences and educational opportunities.

It's a good opportunity for two things. One, for me to let all of you know about OutSystems as an organization which has been there globally for the last 25 years and India for the last 5 to 6 years. Second, uh what's in it for you is the opportunity for all of you as engineers, developers, coders to be part of the OutSystems ecosystem that I will speak about.

Uh I think the previous session, the gentleman from Tesco spoke spoke about building applications from scratch, which I'm sure is an aspiration for all of you, most of you here. Some of you are probably already doing it. You probably want to take it to the next level. So, firstly, I would definitely want to thank customer base that we have across the globe, which is predominantly across banking,

financial services, insurance. So, these are enterprises which trust a platform like OutSystems. And I'll speak about what that platform does more in detail. Second is large manufacturing organizations globally, which is also extremely complex business processes that they have, apart from the regulations, complaints requirements that banks, insurance, all of those have. And then the developer community. I'll speak about how big the developer community is globally and how

we are rapidly growing that in India as well. So, quick highlights. Why is a technology like OutSystems important for our customers and the developer ecosystem with what is happening across the globe in our industry in the tech industry with the explosion of generative AI a couple of years back and agent AI as we speak right now. First, we are very close to 1 million developers globally who

are actively developing on OutSystems platform. It's pretty simple .NET and .NET Core but using OutSystems developing and building applications for customers like banks, insurance companies, manufacturing companies. That ecosystem is about 1 million globally. In India, we are very close to 100,000 developers who leverage OutSystems as a platform to build and take applications to production. The other few things I'll speak about is the certifications. We have about

close to 70,000 developers who are actively not just building but also certified for As we speak today, we already have close to 4,000 developers who have taken it to the next level in terms of application development in terms of agent builders, right? I'm sure you've heard of agent builders, but when I speak about how you do that on OutSystems, there are about 4,000 developers globally who are

already building agents, certified to build agents on OutSystems agent workbench and take it These are the few customers that I spoke about out of the 4,000 enterprise customers. So, first if you look at banking, why is this important for you? I'm sure all of you would want to be subject matter experts one day. I've been in the industry for 20 years. Most of you definitely looks look

much younger. You're probably in the industry for 5 years, 10 years, 15 years. Irrespective, You want to be working in a particular industry. The gentleman before me presenting here was from an industry like retail, right? Tesco is one of the largest retail giants in the world. And similarly, when you look at the retail giants that are you you utilizing OutSystems as a platform, banking giants, whether it

is, for example, in India, the largest bank in India is leveraging OutSystems for their entire home loan process. Similarly, you see KeyBank, you see Axis Bank, you see Western Union, and in manufacturing, you look at Schneider. Graham is another NBFC in India, and I'll speak about the Graham story more in detail. So, pretty much from an industry verticalization standpoint, banking, insurance, manufacturing. Now, from a credibility standpoint,

since none of you know about OutSystems before, right? We've been there in the industry for the last 25 years. From the time the term, I'm not sure how much of how many of you would know, and I'm sure you would have heard of it, the rapid application development term started, right? Let's say it probably started end of '90s, early 2000s. From there, taking it to the next

level, around 2004, 2005, and then actually working with an established player or a research provider like Forrester was between 2014, 2015 is when OutSystems and Forrester together coined the term low-code no-code, right? LCNC. And then, for the next almost 10 years, one decade, OutSystems has been the leader in that low-code no-code. Now, in the last couple of years, that's definitely not enough for you guys, all of

you as developers and enterprises for their consumption. They wanted it to What did we have? Wipe coding. Enterprise wipe coding. So, I'll speak about how you do enterprise wipe coding on OutSystems. But then in the last 6 months, enterprise [clears throat] wipe coding is not enough. Organizations, enterprise engineering community, developer community want to take it to the next level of leveraging agents. >> [snorts] >> They don't

just want to build applications. Now, they want to build agents. It's either the agent sitting on top of the application or the agent itself being the first interface to people like you and me as customers who are leveraging that application. These are few >> [clears throat] >> facts which are relevant for you when you work with your end customers. 90% of the engineers will be leveraging AI

for their code in the next 2 years. And I think none of you would be surprised with it because most of you are already leveraging it when we write code as of today. 60% of the developers will have in their desktop, in their mainframe computers, in their laptops at least three AI coding. We probably already have this, right? So, it's again not very surprising. And some some

people probably more use more than three in our in our laptops and in our devices today. 40% [clears throat] of all the code that's written across an enterprise will be AI generated. And I think this is probably the first thing that you've already seen in the last 6 months to 1 year. Now, next level of data points to create anxiety, but also find solution among the engineering

community. Close to 70% of the IT leaders, either those are the leaders that we work for, the CIO org, CTO org, or your customer org, right? It's the product that you're building and delivering it to the CIO's org, to the CTO's org, to the CDO's org, or let's say it could be an head of AI, head of data engineering. They are extremely not confident or feeling insecure

about managing AI's governance. So, one of the first thing that I spoke when I got on stage was industries which have high complaints and governance. Banks, financial services, insurance. On an average in the IT industry, the applications that all of us are building as developers, engineers, 40-50% of that is being utilized by BFSI as an industry. Banking, financial services, insurance, which includes the smallest of a fintech

which probably operates in India or a new-age NBFC which is leveraging technology for reaching out to their customers. >> now this is a ridiculously high number that you see, 2,500% increase in software defects. And this is the reason it's not surprising for any of us is we know if I only use AI and generate code and I directly take it to deployment, I'll probably be fired. Right?

My app will crash on the first click. And there's a reason behind it. Because the governance issues that we spoke about, the guardrails issues that we spoke about, the quality of that code that was generated by AI. And I think in the previous gentleman who spoke here spoke about a couple of instances, right? Saying your LLM does not have a great memory. How do you know that

the code that you generated for 10-15 minutes before for the same requirement you generated 10-15 minutes again later, will you get the same code? There's no guarantee, right? So, which means the code that you generate purely based on AI tools will not be enough. Which means there would be defects. Which is where people in this room, for example, the developer ecosystem, the 1 million which leverage a

platform like OutSystems building applications for as critical as bank a manufacturing company, an insurance need to be involved in this process. Now, I'll take it to the >> next stage in our topic today. There is a clear context gap. Which means you saw the previous gentleman show you about the knowledge graph, and probably I'll double click and go one level beyond that to talk to you about

the enterprise context graph, right? Why that is important. But as of today, we clearly know there is a gap on enterprise context. how does this AI tool LLM know my my requirements, my customer requirements, my business process workflows to build an application specific to that. Whatever that generates, the code that's generated would be extremely generic without the enterprise context, without the proprietary data that the enterprise owns.

of course, the changes one of our customers to make it very easy to understand. He was conducting the CIO was conducting a huge celebration dinner for the entire organization. The CIO gets on the stage and he says, "Today, the business sends a change request in the morning. What used to take, let's say about 2 to 3 weeks between the change request coming in, iterations of understanding the

change request, and actually taking it to testing by the time we take it to production and the business is happy with what the changes that we've done. Let's say it used to be 2 to 3 weeks, which also they would be very happy with. Today happens in 24 hours. Which means bottom line, any changes that needs to happen on an application level, application layer, is extremely faster

than what it used to be before. Thanks to AI generated LLMs, AI generated processes. Now, there is a clear gap which I spoke about, which is the enterprise gap, which is one, the business requires a particular speed at which applications need to go to market. And there is a challenge when you only purely depend on AI generated code and try to take that to production, either in

the form of security breach or in the form of the code quality in itself. So, there's a clear enterprise Imagine your world was like this. Irrespective, you can't change the multiple UIs that you would have for multiple applications, the multiple workflows and the business logic that is required for your business. Whether you are in the retail sector, bank sector, or you're running your own startup, or let's

say you're building an application for a manufacturing company. They would have multiple business process. There's multiple business logic. of course, there are multiple data sources. So, we have something called data fabric in OutSystems. The reason behind data fabric is to say, if a customer has data coming in from sources like SAP, coming in from Oracle, coming in from MongoDB, coming in from Snowflake, you cannot ignore the

multiple data sources that an organization needs to use. Then, you cannot ignore the governance and complaints that the organization requires. Last but not the least, a bank, large bank in India would want to be on prem. They don't They're not very happy going on cloud. Vis-a-vis the largest quick commerce company in India like Zepto, for example, will be very happy to go on Similarly, if you look

at the new age tech companies would be very happy to go Or some of the manufacturing com- companies, organizations would probably want to do a hybrid. you need to be in a scenario where the application that you build is got to be readily deployed on prem, on cloud, hybrid. And imagine you had a platform in between to do all of this. In simple terms, you work for

an organization, you're part of the developer ecosystem, engineering ecosystem. You have to build an application for that organization. Whether it is customer facing or internal. The first complication comes in saying there are multiple complex The second complication that comes in is there are multiple data sources coming in. I need to pull data from the database. It could be a traditional database like Oracle database or it could

be new age MongoDB for for example, vector data. And it could be as new as the data observability tools that the data's coming from. But that platform that I'm using to code needs to be able to integrate with all of these data sources. And does that platform have enterprise guardrails, GRCs, security compliance, which OutSystems in this case has for Right? Last but not the least, can I

host it on AWS? Can I host it on my customer's on prem? Or can I do a hybrid solution, right? So, that's the advantage that a developer, an engineering team would get when they build on OutSystems vis-a-vis trying to do that traditionally on high code. Right? If I can even use that word anymore. Right? Traditional coding does not give that opportunity for them because they would have

to sit and build every single interface on their own. Vis-a-vis when they leverage a platform like OutSystems, which can integrate all of that on one common platform. Now, a quick overview in terms of what are the things that have happened in the last 6 months since the explosion of agentic AI. >> Number one, the last couple of years, enterprise wipe coding {slash} conversational app generation {slash} natural

language query leveraging development. All of these things mean the same. Enterprise wipe coding {slash} wipe And that is something that our customers have been very actively using a tool called Mentor. So, now the new generation, next generation Mentor is GA, which all our customers, large customers that I spoke about, banks, financial services, insurance, manufacturing companies, retail companies actively use. So, which means their developers use OutSystems to

generate code, to validate code, to do the quality check that is required for the code along with their common sense and along with their with the security guardrails that OutSystems has for the last 25 years. Right. The next one is with the increasing leverage of Can you stop the developers in an organization trying to use wipe coding platforms? It's very difficult, right? Somebody's using lovable, somebody's using

cursor, somebody's using Claude. It's very difficult to tell them not to. What can we do in in place of that? Saying you probably have built the code on some new age wipe coding platform. At least bring it to a same safe for two things. One, to ensure we build in enterprise context. this code needs to work for this bank, for this manufacturing company. So, let's bring it.

Number one. Second, I want to ensure there is absolutely no security threat when this code goes for production. Last one is of course building that developer ecosystem. Integrated developer ecosystem to ensure that if I have 2,000 developers in my can I have one common platform? They could be using any wipe coding tool to generate 40-60% of the code. But can I bring them on to one common

platform which will ensure as an enterprise? I'm not a mom-and-pop shop. I'm an enterprise. As a bank, I'm responsible for end customer data like all of us. As an insurance company, I need to have those data for so many years because of the regulations. Similarly, a manufacturing company cannot afford to lose its business process to its competition. Because that's how their R&D IP revolves around, right? Since

I spoke about mentor, now I'll take a couple of minutes to talk about the next tool that we had launched last year in September called Agent Workbench. The industry was not happy with enterprise wipe coding. The industry was not happy with natural language for query or for programming. They said let's take it to the next level, which was saying building agents on top of applications. Imagine the

platform that I spoke about where you're able to build code or you bring in code from anywhere else. Now you can also build agents and take those agents to production. All of this happening on OutSystems through agent workbench. Where you go, build agents, test them, put the business logic in terms of where that agent needs to take that decision and take it to production all on one

single platform called OutSystems. I'll finish my presentation today. Do I have another 3 4 minutes? 5 minutes? With the story and the real life example of one of our customers in India. And since I kept talking about banking, financial services, insurance, complaint, security. How does this impact the lives of people, right? So Griham Housing Finance, erstwhile Poonawalla Fincorp, and then they were offshoot. Now they're called Griham

Housing Finance. Varun is the CIO of the organization. So Griham gives loans for people like farmers or it could be a small Kirana shop owner in tier two, tier three cities of India, right? It could be villages. So it's basically one level below affordable housing, right? So it's probably the basic first house that 90 to 92% of the Indians, none of us belong to that, right? We

probably belong to the first uh let's say top 6 7% Indians in terms of the economies of scale and the uh uh income bracket, right? 92 to 93% of the Indian population that falls into that particular bracket. It's probably one house that they buy in their lifetime. That's the dream that they live with. Typically for the last so many years, what happens with them? They go through

a home loan process with any bank. There's a lot of hard documentation that needs to be done. There's a lot of pilferages, there is a lot of leakage. And on an average, they probably take two to three months, sometimes six months to get a loan. Right? So, what GRUH Housing Finance has done with our systems in the last one to one and a half years is one,

make that process 100% paperless. This is a public case study, you should go and read about it. And [snorts] this making it 100% paperless is not to so say that we will go green. It's 100% paperless to ensure the process does not have any pilferage, any leakage, which means what that shopkeeper says is what is entered, is what goes, and they actually get the benefit of doubt

when the loan has to be sanctioned. Number two, what process used to take in terms of I'm sure all of you know the loan approval process that takes. There is a business rule engine, there is an underwriting process. All of this traditionally, even in the best banks with technology investments, with people investments, would probably take two to two to three weeks, right? Now, today, GRUH Housing Finance

has brought it down to three to four days. And their goal is to bring it down to to three to four hours, and eventually 40 minutes is their goal, that a farmer, a shopkeeper owner, will apply for a home loan, and probably get his dream for the house to get a home in 45 minutes approved. And which is that happening today, let's say about three to four

days is possible. And all of this is possible of the technology that they use, and how fast they're able to build the entire loan origination process, the entire customer onboarding journey process. How do they ensure that this is not This is 100% paperless. It's 100% secure process. And eventually, it makes a difference in our people's lives, which is very similar to the previous speaker talking about saying

whether it's retail where the food needs to get to people or in a banking space where people need to own their house. So, that's how as a developer community that we all think saying the application that I build cannot crash because it's going to impact people's lives. It's either in the hospitality industry, healthcare industry, or it's in the banking industry, insurance industry. With that, I will finish

my presentation and show you guys the QR code to scan. if you if you want to start your journey mastering the agent workbench which I spoke about. If you want to build agents leveraging OutSystems platform, that's the QR code. I'll leave this here for a few seconds before I move to the next slide. There's one more QR code which I will show you to you guys for

one of the upcoming developer conferences. The platform's called OutSystems Developer Cloud. ODC OutSystems. Go and research about it. Go and research about agent workbench. Go and look up at ODC mentor. So, a mentor is the web coding platform. Agent workbench is the agent builder. And this is the QR code for you to register for the upcoming developer conference on August 8th. This is the second time Krishna

will be doing it in Third time. This is the third time that we'll be doing it in India. Last year we had about 700-800 developers there and our goal is to get to about 2,000 developers this year. Uh and I'm I'm sure most of you have been part of AWS summit, MongoDB summit, and I'm sure conference summit, a lot of these summits. So, this is purely for

developers, the developer day for Thank you very much to GIDS team and thank you very much. All the best. Let's enjoy building applications and agents. Thank you. >> [music]