So this is the agenda for today. I'll go directly to the topic. I'm sure everybody's excited, to learn about, how AI is transforming QA. So let's, dig deeper.As I show, in the agenda, I want to make sure we see the transition. Right? How we used to be and how, currently most of the companies are and where we are heading toward. Right? So, yeah, traditional QA, as you know, it's a very specific boxed timeline and boxed effort, meaning it's a sequential, task. Right? So you have a requirement stage, QA involved in early stage. They start planning for testing. They start writing development after that. So everything is happening in sequence. Right? When PM is doing, requirement gathering, developer started, the designing the product and everything, and QA, in parallel start, doing the test planning.

Then we do the test, case development, and we do the environment setup. We do the test execution, then we do the testing, find a bug, and then, we close the cycle. Right? That was the traditional QA, method. And then, we came to more faster agile methodologies. Right? So it's more modern. And we are doing exactly similar thing, but it's more in fast pace environment, I would say. So we are doing exactly similar thing, doing the testing, developing, you know, planning, designing, doing the deployment, doing the verification, finding bugs, but it's in more faster ways. So in traditional QA, waterfall model, we had, like, let's say, one month QA is going to test. Right? It was a, schedule timeline that we can go through.

In agile, in the other world, depending on the company, how long is your sprint? Based on that, how fast or slow, the sprint will be. Right? In our case, we have one week sprint. So all this iteration is happening weekly basis. Right? So it's much faster, but it's similar process that, we used to do in traditional. Then now we are in AI area. Right? So we are doing those similar kind of activities, but there are a lot of help now. Right? Before, QA have to spend time that, okay, we are doing the test planning. We are asking, we are checking PRD. We are checking, FS document to figure out, you know, where, we start writing the test cases, making sure all the edge case scenario is covered, which environment, how we are doing and all this thing.

We have to spend time on writing the test plan and everything. With AI, we have, I would say, we have an assistant who can help you with all this thing. Before test planning, maybe, it will take a day or two just to come up you know, go through all the FS, every, requirement, every design you want to figure out boundary cases, edge cases, and stuff like that. Now what we are doing using AI let's, for example, if, you are using cloud code or cursor. Right? You can feed your FS document or your PRD document or both, and then you have UI mock up. You upload those, with whatever, tool you are using. Right? And Cloud Code will come up with the test cases, come up with, edge case scenario in minutes.

Before we are spending time reading every line and trying to come up with those thing, now it's become ready within a, you know, five minute or so. And then our job is to basically check, did AI interpret whatever we are interpreting. Right? If the FS document is written properly, if PRD is written properly, it should be very accurate. Right? But still, sometime AI hallucinate. Right? So that's why we came to the, picture. Right? So we have to make sure we see what AI come up with, the test cases are relevant, the edge case scenario, the assumption, what they are doing. Everything is, to the point. Right? So is we are just having a helping hand, not necessarily, replacement of ourself. Right? So we are getting all these test cases, immediately.

And then we can actually plan for the automation as well. So what happened before even the code is developed, we can plan for our automation. As I was mentioning, the wireframe or the UI mock up that UX designers are giving us, we can take the screenshot and feed it into the AI. They can actually, Cloud Code can actually plan for the automation before the code is existed. Right? So the process becomes so fast. Now before we are waiting days to write the test plan, test cases, finding bugs, you know, executing. Now we can be even, prepare before the code is marked. So we have the test cases. Cloud code is helping us with writing the test cases, preparing for our automation, write the pseudo code, ready for it, and indicating where the bug can be.

And then, it will create a test suite. What like, after the March, we can run, run the execute the test suite and make sure find the results. So that's why I show this evolution, evolution of, QA testing. So now we have a basic information, then systematic understanding, then autonomous evaluation, intelligent test driving, and then full autonomous QA. So when we are done previously, we finish the manual execution. We provide our feedback that our testing is done. Now we'll be working on automation. Here, in the other hand, we are we'll be fully ready. Once we qualify, it will be ready with the automation as well. So that is the change we are seeing in the industry. So, of course, we had, some challenges. That's why we're seeing all this, change upcoming. Right?

So before we did, in waterfall model, obviously, we did most manual testing, and it was very time consuming. Right? Sometime, depends on which QA is doing what. Obviously, there were test coverage gap. Sometime we missed finding a critical bug. Then our test script was, you know, taking long time to maintain, and it was very difficult to, scale. That was the challenge in traditional QA. Right? Now transforming with AI, many of those concern are rectified. So that's why I come up with this diagram where you you can see whatever we discussed early on, this is where exactly we are seeing the transformation. Right? So I'll go with each of the point, in this, diagram, right, so that you can press on it. AI test generation as I was explaining. Right?

AI can read the PRD. AI can, read the FS diagram. Even, wireframe that I was talking about, if you have, UX design, you can fit into that. If you have user story, task, any Jira ticket with the details, you can fit into that. If you can fit into this Jira ticket directly, you can just, provide the information is in Jira, and you can upload those, information. It basically reduce all the manual tasks that we are doing day to day. Right? It speed up, any, manual intervention. It provide consistency, and you can centrally change your test suite very efficiently. Right? So that's the faster test creation part. And, of course, the coverage become better because of, you know, it's reading directly from the PRD, directly from the FS.

So it's become very, very, efficient to identify the edge case scenario. So, it reduce the risk. Then the, come to the predictive analytics. Right? So, previously, we have to go through the failure analysis to see that, oh, okay. You know, in this feature, this part is the, fragile area, and flaky tests are in this area and all that. We have to, every day when we are doing failure analysis, we have to focus on those areas and stuff like that. With, AI, you can actually analyze your report before a human eye look into it. Right? It's, AI already have all the historical data that previously this test, how many time this, test case was running, how many time it passed, how many time it fail, if it fail, why it fail.

Is this failure similar to previous failure, or is it a brand new failure? Is it a product problem, or is it a, script problem? All this thing, it can predict and analyze based on how you set it up. Right? And you can, detect your, defect early on. So, that is one of the very good benefit from AI. Right? And then the self healing automation. So, SQL, I think every all of us see this, like, sometime DOM change. Right? Look at our change, developer change it, and they didn't communicate with QA. Your automation is failing. So in order to help with that, AI can, you know, self heal. So you can use AI or Cloud Code or whatever, AI helping tool you are using. It can actually read the DOM directly from the UI.

And if there is a DOM change, the test case can be, in flight, changing that, new DOM instead of making the test case fail. So it's in flight change is possible. Before, it was just the hype that, oh, we can do this, this, this. But these are actually reality in my team. Everybody is using this. Our failure rate reduces drastically after we are using AI. Right? So and then it can adapt with your, UI changes. Maintenance time become very low, and it improve automation stability. Right? So, yeah, you can read through it. I don't have to read, through it, but I'm just giving you the example where, AI is helping us. Right? Then intelligent test execution. Like, let's say, if you are in weekly sprint. Right? So, some features are changing and some features are not changing.

So prior to AI, you have to identify, okay, which test case you want to run, which test case, less risky so we don't have to run. But using AI help, it can determine that, this feature is changing, so we'll run this suite. If it's not changing, source code is not changing, then they don't have to run those codes. So the release become faster using all this. It can optimize the, regression cycle, And, you don't have to run the redundant test cases. It can identify which test cases are the high value test cases and where it it needs to be execute. Enhanced test coverage, as I was mentioning. Right? So, AI can help us identify the gap, in existing test suite, going through the failure analysis, going through the edge case scenario.

You know, sometime QA deal with the information. Like, let's say, we don't have the data generation. We have the test case, but when we run, we don't have enough data to validate. So AI can help us simulate that data as well. So that will give, you know, enhanced test coverage, you know, higher confidence that, you know, our product quality is better. Reduce testing time. Yeah. We mentioned that, many time writing test plan to test case to automate everything become much, much faster. And also DevOps side. Right? If, anything that, pipeline need to be enhanced and stuff like that before, let's say, we go through the YAML file and try to fine tune which test suite to run when scheduled is that. Right? We can enhance that by, using maybe, you know, we we can generate the UI deployment page, and it can go through all this information, in UI.

You don't have to go through the YAML file. It's time consuming to figure out, you know, which build, which, test suite to run and all that. It can be, you know, click and play. So, yeah, you can just do the checkbox, figure out from the UI, and, run it. And all this thing can be possible using your AI assistant. Cost efficiency, obviously, since we are, doing everything faster, we are saving lot of, manual man hour. So it optimize, resource utilization and, of course, long term QA cost. Right? So I'm not saying that we are reducing cost by reducing resource. I'm saying we are reducing cost by reducing manual effort. So let's say if, you have five member team, we used to do five feature per week. Now since we have this AI assistant, now we can do 10 feature per week. Right? So that's give us lot more ROI.

So, yeah, that's very, very cost effective and, obviously, improve accuracy. Right? We a human can do lot of human error, very nitty gritty data, driven validation, defect validation. Everything is possible using AI. So accuracy is high. So, yeah, I just wanted to, give you, a snap of what type of metrics we can collect so that we know AI AI is actually benefiting us. Right? So, yeah, these are just the pointers. Again, every team is different. Everybody is tracking different metrics. So, your metrics might be different than mine, but, this is what you can see immediately how much, benefit we are getting by AI. Right? So faster testing cycle, we mentioned that many time. Regression, time reduced quite a bit. Release cycle become faster. Test coverage, increase, less number of edge case scenario, is dropped.

So now you have better edge case scenario, risk based scenario validation, and then reduce the defect, you know, so we can identify defect, in the QA, duration, not, you know, in the field. Some more, you know, faster failure, triaging. Right? So as where I was mentioning, right, failure analysis take hours and hours. Now using AI, you can actually provide the test report and say, figure out the historical data, you know, this test case, how many time it failed, how many time it passed. Is it a, UI flakiness? Is it, DOM change, or is it actually feature change, or script change? So all this thing you can, figure out very, very fast. Low maintenance because we don't have to spend that much time in the triaging. We need to spend less time modifying our script because, you know, it has context, actually. And then cost efficiency because, we are doing everything faster, so it's very cost efficient now. Of course, there are challenges, of course.

When we are using AI, we have to make sure that, the data quality, availability, everything is as per our, requirement. Right? The model accuracy. So you ask anything to AI, it will return back some data. But is that data accurate? Human need to look into those. Right? If we have existing framework, we have, code, best, coding guideline. Are they, are AI following that? If it create different type of, code base, then, you know, our existing framework become corrupted. Right? We have to make sure that we follow what we have right now, existing guideline, existing, best practices. So, basically, QA skill gap also another thing that I see because we used to work in different way. Now, AI is injected.

So now every QA engineer have to learn how to use AI. So that tool learning that particular, you know, using how to use plot code, how to follow the, you know, compliance, for the you know, which code you want to share with Cloud Code, which are not following the company compliance policy and everything. Those are challenge, obviously. Then, yeah, future of QA with AI. As I was mentioning, right now, everybody is using integrating AI in a, developer life cycle in QA life cycle. So, future will hold something similar. Right? So I see that a lot of autonomous testing, is going on. Continuous QA strategy is changing based on that. How we are doing AI assisted development plus testing coverage. Right? How fast we are coming up with that. Previously, what happened that, you know, QA members are doing it manually, then we are doing automation after the fact.

And then if we have too many backlog, then we might have a contractor. They are working on the side for automation. We are only working on the manual testing and stuff like that. Right? But in future, it's everything should be going in parallel. Right? When we are, feeding into the test case generation, test case automation planning, and everything using AI. We are just, checking if, accuracy for those, model and everything. Right? So we have more time to think about, what other feature we can enhance, what existing test intro we can enhance, how to optimize that. So we have much more creativity, to work with, basically. That bring me to the key takeaways. So yeah. I'm seeing that QA is transforming QA from reactive to proactive. What the what does that mean is before, we find a bug and then we react to it, we try to fix our test script and, making sure if it is a feature bug, then we work with developer and, try to fix it, update our test cases.

Now it will be more proactive. Right? Before even we keyways code marched to the master, it's already finding those bug because AI is helping us on that. Right? And then, our automation become more intelligent, so it can find more hidden bug. If there is a issue with the data, it can detect and help us finding those fine tuning. And then, yeah, obviously, QA engineers, have to evolve with, current time with AI scale and, you know, it's it's not just efficiency part of it. It's more like, it impact directly time to market, customer experience, operational cost, and, risk reduction. Right? So I would say, one thing, from this session, if you take, AI plus QA will be the faster, smarter, and better quality, deliverable for your product.

I'm sure everybody have this question in their mind that, can, QA testing be replaced by AI? I'm sure everybody else's mind also have the similar question that, oh, development will be replaced by AI or not. Right? From my point of view, what I see so far in the industry, I say it will be more of a multiplier than a replacement. Meaning, it will help us do our work faster, not necessarily replacing us. Because if AI is hallucinating, who have to look into it? Human eyes have to be there. So we are not replaceable yet. So that's what I wanted to mention. Yeah. With that, I wanted to ask you guys if you guys have any questions.