Why AI Alone Can’t Win: The Low-Code Advantage with Zoho Creator & Zoho Qengine - Vignesh S

In this talk, Vignesh, a solution expert from Zoho Creator, discusses the limitations of relying solely on AI for enterprise application development. He emphasizes that while AI can quickly generate prototypes, it lacks the robustness required for production-ready solutions. The speaker introduces the concept of the 'Hello World illusion', explaining how initial results from AI may appear satisfactory but do not meet the comprehensive demands of enterprise needs. He highlights the advantages of using low-code development alongside AI, illustrating how low-code platforms simplify the development process, reduce the lines of code needed, and ensure enterprise governance from the outset. Vignesh further explains how Zoho Creator integrates with Zoho QEngine for testing, ensuring that applications are not only functional but also reliable and compliant with industry standards.

Hey all. A very warm welcome to everyone who's present here. I'm Vignesh, a solution expert from Zoho Creator, and uh we are going to take a uh a discussion on why like AI alone isn't enough uh to win the whole enterprise race that's happening lately. And uh So, let's start with the um the base uh of this uh discussion, that is What I feel is uh we

live in amazing like an incredible era where a simple like a line of prompt can spark a whole application. Like it has I see AI as uh like a superpower that has empowered like every one of us to be able to like uh uh step up uh and build an old application of us. But but when we look beyond the prompt, so is that the only uh

uh thing that's necessary to run an enterprise ready solution? Because what I want to say is these applications that the AI uh assistance or the or the all these agents that provide us are good for the day good for the day one, but they're not enterprise ready. So, to be an enterprise ready solution, what it takes is is is going to be the core of this uh

discussion. So, let me talk about this experience, that is the Hello World illusion. So, let me explain why I call it an illusion. So, we all would have experienced uh something similar to this. So, when I like used to like explore new programming languages, like the the this is obviously the first thing that every one of us would have learned. So, the try like printing out Hello

World. So, at that moment, we get a sense of satisfaction and the the completion that we going to ace this programming language. But that's not entirely true. They are miles ahead. So, that is that that the exact scenario that we are facing with AI today. When we like get like create an application with AI, it gives us something that works on day one. But, for it to

sustain the enterprise uh the standards and uh to be able to work uh like solve the the the the daily usage everyone, we it it requires a lot more. So, this is what this is how I put it out. So, AI helps us definitely in giving us an excellent prototype. But, will it be a production-ready solution immediately? Not at all. So, so what could be the solution

for this? So, before that, so we also have uh the this is the the infrastructure iceberg, something that I would like to show. So, what we often get to see on the front end is is the UI and basic projects that the AI would be able to definitely provide us like within minutes. Just within minutes, you you will be able to see all this that is the

tip that you see over there. It is the UI and basic logic. But, the base is what matters. So, who's going to give us who's going to help us to to to scale everything and uh the the disaster recovery and encryption and everything. So, all these comes with the foundation of low-code. So, that's where we see the synergy of AI plus low-code. So, when we talk about

low-code, so then let's just have a a quick comparison on how many how many lines of code that we are talking about. So, when you are building an application entirely with help of like uh AI and when you're generating everything from scratch, it's going to be like millions of lines that's that's being generated over there. That because we are we are we are not you're not just

creating the application's functional uh functional part, you're also setting up the base for the application for it uh to be able to like add the each and every submission of the data that you're putting in into the into the database and starting from everything, you'll have to work on the authentication and a lot more stuff. But, when you compare it directly with the low-code part, the low-code

approach. So, it's just like 5 to 10,000 code and even that is not going to be like uh like written entirely entirely by us because the low-code approach, that's what it highlights. We have a lot of uh UI-based development itself in the for the primary uh part of the application development. So, what you what's the like the the bottom line of this is so, you will just

focus on the business layer of the application development, not you know, you will you will probably minding the hosting and the whole maintenance of the application when it comes to low-code approach, especially with Zoho Creator. Uh this is what I mean when I say that you don't have to like uh write each and every line of code by yourself even when I mentioned that it's going to

be 5 to 10k lines of code. So, primarily it's going to have like this kind of a visual code complaint development which uh we will be using to build the base of the application and which also like involves minimal coding later on for solving complex logics and complex calculations with or whatever you'd like to perform. And it's all these applications that we develop on Zoho Creator also

device agnostic. So, I'll uh quickly come back to this point and explain you a lot more on that. Now, since we were talking about uh the lines of uh codes that we'll have to go through uh manually to to review each and everything when we when we build something using AI. So, let's like just compare this quickly. it it we can easily generate millions of lines of

code, but to to to to to review it manually, it it takes a lot of effort from a whole uh testing and QA team for to be able to like review that and uh come up with all the hallucination task done because it is AI is known for for hallucinating within when it is generating codes that code code that is for like lakhs and lakhs of lines.

But the low-code advantage, you're not coding everything from scratch. You're only focusing on the business layer. For example, if I'm building an application like for let's say I'm going I'm building an application for a small clinic, let's say. So, I'm just going to create a patient management system. So, when I'm building an application for that on a low-code uh platform, all I focus is on how the

process is going to be automated, not where the data is going to be stored or how the these users are going to be logging in or even like how I'm going to have stored this data securely or all the or whether since this is a clinic medical requirement, so whether this data is going to be secured under this and whether is it complied with HIPAA, you don't

have to mind about anything. So, you just focus on automating the functional part. What kind of data I need to capture, so add those fields into your forms. And what kind of automation that I'm going to do, so when I when this when I register this person, this patient, I want to add them into my portal, so I send them out an automatic invitation. So, all these

are the functional parts that you're going process when you're developing things on a low-code platform like Zoho Creator. And finally, I would also like to add another service of Zoho that is Zoho QEngine because once you're done with the development, you also get to test it completely with the help of QEngine. QEngine is our automation like manual automation testing platform. So, where you can test everything entirely

on all the platforms. So, it's uh to to to emphasize on uh the power of QEngine, so as you might know, Zoho has like over 60 plus services, 60 plus products in we have in Zoho. So, each and every application, each and every product that we own own in Zoho are being tested with the help of QEngine. So, let's say like if we have product update or

some launch that's coming up, before each and every release, we run like around 15,000 like test cases with the help of Q engine for all our products. So, that's how powerful Q engine is. And there's also to emphasize on how it is easy for you to test everything on Q engine. So, it's not just for the web applications that you build. And it's not just about UI

testing. You can test everything including your APIs and custom workflows and and whatnot. So, all it's common for And so, this is what I was referring to as device agnostic. So, there's like a huge differentiator I would say when it comes to developing developing approaches between AI alone and with the synergy of AI plus low So, with AI approach, you can like definitely get an application that

works for you on that day, but that application whether is it like instantly accessible on a mobile as on a mobile or is it like also compatible to be to to run on an iPad? So, all these things come to your mind. So, it's not it's obviously it's not going to happen immediately. But when you take the low code approach, immediate native applications. So, it's not about

maintaining separate code bases for different platforms like iOS or Android or web. No such things. Once you develop the application on web, it's instantly accessible and not just as browser applications. It's not any progressive web apps or containers that just run on the browser. We support native applications that you can instantly download from Google Play Store and access these applications on your iPad or or even on

your any Android device. Like it's accessible on mobiles and tablets and web at the immediate moment that you finish the applications development. And all these applications that we're talking about are also like completely like rebrandable and white label. So, you can also like like Uh, post the deployment post the the development, you can deploy them as uh rebranded applications onto these Google Play Stores and App Stores.

And now let's come to the enterprise governance. So, in the whole like a corporate world, who did what and how secure the data is the this is the most important thing. After all, it is uh information technology that we are working in. So, this is the core part. This is something that completely lacks when we focus on an AI completely AI-driven solution, like an AI only solution.

So, this is the most important factor, I would say, when it comes to developing applications for enterprises. So, when you develop applications on Zoho Creator, so these compliances are not an afterthought. It's it's it's your compliance from day one. Any it it could either be a very very minimal application, a very personal very very personal application that you've built for yourself to to to mark your to-do

list with just a single form and two fields. It doesn't matter. It's it's it could either be a minimal application like that, or it could be a full-fledged core core business running application, like a manufacturing process uh where we manage the complete process of a manufacturing industry starting from the sales order and receiving them and processing them into production plans and converting them into complete um uh

like delivering the logistics part of it. So, it doesn't matter. Any application that you build on Zoho Creator is going to be compliance from day one. So, then along with all the complete uh centralized admin access with uh role-based access control where you define which exact user which which exact user should uh be able to access which exact uh information. And even within that, you'll be able

to define like uh in in a granular level uh restrictions, like to whether to let this person import data or whether to to let this person export data from this application. All these secure measures are available in Zoho Creator from day one. And other than that, the transparency part, automated audit trails. Um since you were I just talked about who did what is the most important question

in the IT industry. So, you get all that done automatically. So, you don't have to like build something by yourself in this application. You just focus on building this business layer application that is for example our uh clinic management application which I'll also demonstrate in a while. the rest of the part are going to be a standard. So, it's not an afterthought And the bottom line, so

these are like the numbers that speak for themselves. you can see that how fast the development has changed with uh the the combination of AI and low-code and how fast the like the and how well the the development timelines have reduced and like overall it the the the combination of AI and low-code is benefiting us. And so, this is what I call the winning formula. So, AI

I use the artificial intelligence as for the for its speed and Zoho Creator uh like the platform Zoho Creator for its foundation and finally Zoho QEngine for its reliability. So, with the with this as a trinity, you'll be able to like develop enterprise-grade applications which are also reliable since you are like generated manual test cases, automated automation test cases like uh n number of test cases that

would run and make sure that the application that you have built is fully reliable. So now, I'll quickly take you all through uh a demonstration of both Zoho Creator and QEngine. So, where I'll just uh create an application with a single line of prompt. Um, just a second. So, this is actually the Zoho Creator homepage. So, where you can find this as a platform, you can find

that there are multiple applications that exist over here. Each one like having built for its own purpose. So let me like quickly show you how I can create a new solution on a local platform like creator. this Zoho Creator Edition actually comes along with uh two more types of solution which using that you can also create like BI Analytics uh solutions and also integration flows which helps

you to integrate with uh 1,000 plus uh third-party solutions like cloud applications which is almost like in lines of Zapier and the BI Analytics almost like uh a Power BI alternate tool from Zoho. So let's focus on building the local applications. So while creating a new application we'll find these many different options. To start with we have the the the the the most commonly uh widely used

option in Zoho Creator that is creating applications from scratch. Because the very reason like people choose uh a local platform is they also have very unique requirements. So it's not something that you can get out of the box then that that that defeats the whole purpose of a local platform. So creating applications from scratch has been like has been like the the the thing of Zoho Creator

because AI is something that we are seeing in the the recent years but the platform Zoho Creator has existed from 2006 actually. In the coming September we are actually like going to celebrate the 20th year year of our platform. So even without AI or uh these uh these latest uh features that we seeing Zoho Creator on its own is a very reliable platform and it has like

been uh like battle tested for almost two decades now. And a new option and another option to create an application from an existing list of templates. So we have some uh common templates for uh most of the uh common requirements that we come across we also like segregated them here. And finally you can import it using any existing files that you have like the data files and

uh supported file formats are also listed down below. The common ones like XLS, CSV, JSON, numbers all that are supported. So when you import your data file, the the the platform analyzes the kind of data that you have in and converts that into a web application. So, that is also as I mentioned earlier, instantly accessible on your mobile and tablets as well. So, now let's talk about

the the the recent feature that is creating applications using Zia. So, Zia is those in-house AI assistant, actually. So, let me click here. I just I have prompts with me to create the application that I was talking about simple clinic management application. So, we have that prompt with us for here. So, the prompt is very simple to build a patient management portal to track the medical records,

track the doctor appointments, and insurance claims, and also an approval process automated approval process for surgical procedures. So, just with this simple prompt, so you and also additionally, so when you all can actually this can also be like you you all can also try this out for free. So, we provide a 2-week free trial for all the users. So, when you're trying out, you can also like

apart from this text prompt that you give, you can also upload some additional documents, supporting documents like a BRD or RFP or any process flow diagrams that you have. So, it enhances it it actually helps the application, I mean the the AI assistant in creating a better So, let's start building. So, let me like show you how does the the AI the AI assistant transform that prompt

into working solution. And so, these are all the use cases that the AI assistant has taken the Zia. So, just with the prompt it has like assigned few modules for medical recording, appointment scheduling, insurance claim processing, and surgical procedure approval. So, this is the app schema that the AI assistant has built for us. So, during this whole process, we are not planning about where it's going to

be hosted or where or how this is going to be maintained or anything. This is pure business layer automation. We're only focusing on the the the the problem that we're trying to solve here. So, this is this is the data model that the AI assistant has created for us. So, each one of these rectangular boxes that you see here would be available as individual forms for you

to access them when the application is ready. And each one of these are also different field types that we have in Zoho Creator. So, as I scroll down and you can see that each one of these tables are also related to each other through these these lines of arrows because they're also like actually related with the help of lookup fields from one table to another. So, now

let me try to create a new application. And before that, if the AI assistant has missed out any of your expected modules, you can actually like click here to add custom and specify that as well so that it also gets included. let me create the solution. So, what actually happens from this is creating the apps based components, that is the forms and reports. Basically, forms to capture

the details and reports to visualize them and a dashboard at the top of the application and designing the basic layout of the application. So, all these are also like customizable for you. Once this is done, you will still be able to customize this and take it further from where the AI has given to you. once the app components are done, it starts working on the RBAC part.

So, it creates permission profiles for possible users who could access this So, it just takes like minutes to create a good prototype for what you're looking for. Just a second. So, now we are actually going to be done with the application component components. So, now it's time for user permissions. This is the RBAC part where it's creating permission profiles for different types of users who could use

this application. Since we involved an insurance claim module, it has also created a permission profile for an insurance agent. So, now we're done with the user permissions. Now, this is time for blueprints. So, blueprints is where the the process like proceed like take proceeds further. So, where you will be defining the the the next steps from once for example when they once an insurance claim has been

submitted and which which department has to take care of it next and what kind of flow that it should proceed. So, all these automations are done with the help of blueprints. And now workflows are happening. So, workflows where it just creates some call like simple workflows uh automations and like validations. And then to reminder notifications for prescriptions and all those common ones. So, now the whole application

is ready. So, now let me click on access this Yeah. So, along with sample details, we have the application generated generated for us within a minute or so. So, here here you can access each and every module that does for us. And all these are some quick links that you can like click on and navigate in between the application easily. And down below we have few charts

and a whole medical records report embedded. All these would also be accessible as individual conference from these sections. For example, let me check the doctor profile. So, this is a doctor profile form where you can add details of new doctors and this is the existing report of all the doctors whom we have in this clinic. And to add more details or to import bulk information, we can

easily click here and import these details. And even while exporting, we support these many different types of file formats to be exported. we also like ask the AI assistant to create an automated approval flow in the back. So, let me just show you that. So, this view that you see here is the end users. This is the live view. This is what the end users of the

application would see. But, as an administrator of this account, I have this option to edit this application. So, this is what the back end of the application is. So, when you go here, you can like add like basically customize this application and develop it further. Whether you want to create new modules, you can do that from here or any workflows that you want to create can be

done from here. let me show you the blueprint that it would have created for us. So, this is the procedure approval blueprint, which we had mentioned in the uh prompt. So, just with the prompt, it would have created a process for the flow. So, initially, the the procedure is requested and from there the user gets to choose whether they want to submit it for review or whether

they want to cancel it. And from there, it goes further whether you want to keep it review and then whether you want to approve it or reject it or escalate it for higher approval. And each one of these actions can have an owner. So, you can choose which which person for example, if you have like a permission profile as senior doctors or department heads. So, you can

just choose that. People under this permission profile are going to be the owner of this approval process. I mean, specifically to to approve the procedure. Or you can basically, you can assign owners for each and every uh approval in this flow. So, this is the blueprint. So, from here you can also decide what happens before this action and what happens after this action. So, by by after

we mean, you can choose whether you want to like uh send out a notification or uh or make an integration call to an API integration call to another another system that you use and update some information over there. And whatever automation that you want to do can be done from here. And then, let me just quickly show you how we can use uh QEngine along with this

application that we just created. So, let me just go to this doctor profile form. let me also simultaneously open up the QEngine tab in then in the in the same So you have test cases here. So you can create a new test case. So let me create new one. it says going to be for clinic app uh doctor testing. I'll just choose the folder to which it

has to belong to. So now we are creating a new test case. So this this demo is going to show you uh how you can create an automatic uh let me create one. So I can actually start recording and all I have to do is get the link that is the basically the URL of the doctor profile form and paste it on to my So once this

is done, I can start recording. So I don't have to manually code what exactly it has to check for. So now it has opened the same uh form in a new tab. Then I just enter the details. So this is for testing this how well this form functions. So just enter the name of the doctor. And the specialization of him. And their email. So all these information

that I enter here are being recorded. So with this like with whatever information that I want to fill in and finally I can scroll down and submit using our Q engine, we are actually recording all these actions. And now that it is submitted successfully, I can stop this. Open. So, now it has shown all the actions that we had done in the other tab, that is in

the Zoho Creator form that we just And I can save it. So, now you can see that a whole script has been generated for what it has to test for. We can quickly save this. So, this is like an internal testing. So, we just like using our offer test applications and then it's credentials to to take it through for the testing, because it happens in different windows.

Each Each one of these testing would be like taken as a new concurrent login session. So, but while testing it with with in your organization in real time use cases, you can also like replace these credentials with variables where you can easily mask them. Okay. Done. So, now I'm saving it. And now let me run it. I'm just choosing the browser version, and I'm opting it to

run. So, you'll be able to see the whole test happening in live along with screenshot previews of what exactly is happening. Yeah, now so you can see that the the test is happening actually that it has successfully scrolled to the element that is the doctor profile form. Just a second, let me restart it. So, we we we actually see live screenshots in the right side of this

window. Once the test is done, you'll also get to with the screenshots that were captured, you'll also get to like play it along like a video. So, you you you have seen what the exact test uh it it has run through the our application. So, as this happens, let me also like uh go to the same form now. So, as you can see that one of this

test uh entry was manually made by me when I was uh recording the the the the the But, you can see that there are two more records of it which has like come into this application as new records which are actually added by our Q engine from here during the test phase. So, apart from this, you can also like make more customizations to your code. For example,

from here. So, now we are just adding static information to these forms, but you can like generate more test cases from here with the help of Zia. So it will be help you with a lot of more manual test manual test cases. And we are also coming out with a few more uh additions in the in the in the in the in the upcoming months. That So

where these manual test cases that you are generating with the help of our AI assistant Zia will also get converted into these lines of code very soon. Couple of the top two months. So we're pretty much done with uh the demo as well. So if you have any other questions uh the end, so please feel free to like shoot them out. We will discuss on them. Thank

you very much everyone for joining this talk. >> [music]