Moderator
Hello everyone uh welcome to a joint webinar hosted by Pcloudy.
Moderator
And test house we are thrilled to have you with us today for an exciting session on Next Generation testing with AI Trends.
Moderator
Challenges and opportunities so before we dive in let's cover some housekeeping.
Moderator
Rules your lines are currently muted but feel free to submit questions during the webinar using.
Moderator
The chat option that you find at the bottom of your screen.
Moderator
And we'll address them in the Q a section uh at the end of the webinar.
Moderator
And also please note that this webinar is being recorded and we'll send out the link of the recording later.
Moderator
So that you can share it with your friends or colleagues or have a watch again.
Moderator
So now let's get to know the two awesome organizations behind this webinar.
Moderator
So first up is Pcloudy is a cloud-based testing platform offering 5000.
Moderator
Plus real devices and browser combinations for mobile and web app testing key.
Moderator
Features of Pcloudy include continuous testing devops Advanced automation AI based visual testing robotic process automation real-time test analytics and more with.
Moderator
Integration with popular testing Frameworks and tools and round-the-clock premium support Pcloudy is trusted by 300K plus registered users from.
Moderator
Small businesses to large Enterprises so next meet Guest House a.
Moderator
23 year old Global software testing and quality engineering company based.
Moderator
Out of London UK with a mission to contribute to a world of high quality software test houses help businesses deliver exceptional.
Moderator
User experiences through Innovative quality engineering and quality assurance solutions they.
Moderator
Focus on enabling total quality management for technology driven businesses is all about harnessing change and Innovation to benefit customers.
Moderator
So be cloudy and test house have joined forces in a strategic partnership to revolutionize digital testing and quality assurance next slide please foreign […]
Moderator
's Advanced cloud-based testing platform with test house extensive experience in software testing and quality.
Moderator
Engineering together we are better positioned to provide a seamless testing experience for businesses the key aspect of this partnership is mutual expansion.
Moderator
Into new markets with Tesla's strong presence in the Middle East UK us and Australia helping Pcloudy solidify its Global reach.
Moderator
So this is the beginning of an amazing partnership and we aim to help more businesses deliver unparalleled.
Moderator
Value to their customers around the world so now that you know a bit about both Pcloudy and test house it's time.
Moderator
That we move on to the main event so we are honored to have George chief.
Moderator
Solutions officer at test house as our speaker for the day George.
Moderator
Has more than two decades of quality engineering experience and.
Moderator
Has worked with numerous portion 500 clients in implementing agile testing practices it is also carried out more than three dozen maturity assessments.
Moderator
Using digitized Frameworks and value stream maps to Define roadmap strategies uh for agile devops and Cloud transformation.
Moderator
So some of the use cases is implemented include automated user story reviews self-healing automation engines and what enabled data provisioning.
Moderator
So with his wealth of knowledge and expertise we are confident that this session will be both informative and inspiring.
Moderator
So without any further Ado let me hand it over to the speaker for today George.
George
Good afternoon everyone this is George So today we're actually going to take a look at uh how is that we can actually use.
George
Artificial intelligence to improve or to enhance the way in which we actually do testing.
George
So you look at some of the interesting Trends you also look at what is happening in the quality engineering world what.
George
Is what are some of the day-to-day challenges that Engineers are dealing with and also see some innovative solutions which.
George
Can be applied to your day-to-day work so we're just going to initially.
George
Start off by looking at the factors that influence quality engineering we are going to take a deep dive into challenges.
George
You're also going to take a look at some of the tools.
George
Uh as well as Solutions and then we'll wind up with a set of best practices that.
George
Lack that La that will enable you to implement artificial intelligence-based testing Solutions in an Enterprise.
George
And then we'll also have enough time for Q a we've also planned a couple of demos.
George
Uh based on the time frame that you're having you'll try to include them as part of the solution […]
George
So if you look at software testing as a stream there is there are a lot of things that I've actually.
George
Uh that is that that has changed in the last I would say four to five years around software testing.
George
You might have seen that almost every organization now is under what you call the influence of agile.
George
So agile is no longer a trend will become a norm.
George
So almost every organization actually has agile projects running I would say like 95 percent of the products are still what.
George
You call are currently run and maybe in an agile mode and there is an equal amount of implementations happening around devops practices.
George
So the amount of devops adoption or the number of companies have adopted are devops.
George
Practices as well as platforms have actually increased significantly in the last I would say two to three years.
George
And there are organizations who have gone with traditional devops tools like Zen cleans puppet all of those combinations.
George
And then there are two there are cloud-based devops tools being used by organizations.
George
Who have who have already moved to the cloud or have plans to.
George
Move towards the cloud platform a good example of that is maybe something like Azure devops you.
George
Also see just like a devops adoption you also see a large number of customers moving towards digital based Solutions.
George
So I think it's digital first is what is being applied whether it's a small business medium business or a.
George
Large business everyone actually has a digital printing so there's a lot of Investments.
George
Going on into digital at the same time there is also an equal amount of focus around moving to the cloud.
George
So either the existing applications that are running on print are actually being migrated to the cloud or.
George
Customers are actually going ahead with solutions that actually are hosted in the cloud like a SAS based applications like CRM Dynamics or.
George
Now because of the adoption of devops and because you're actually having Cloud as well as digital transformation happening in many organizations the.
George
Pure functional testing kind of skill is not the only thing that you might need in order to survive in today's software.
George
Testing world because the type of testing that you might do on these applications let's take an example of maybe a cloud-based application it's.
George
Very different from what you might have done in the past especially.
George
With maybe like if you're a very senior person who have been in industry.
George
For the last two decades you want to start off with Mainframe.
George
Or monolithic kind of applications or you might have tested to a.
George
Three tire and a client server applications and then you might have also seen API based applications and now you're also seeing microservices adapter.
George
So and serverless architecture so the type of testing that you might that.
George
You might that you will have to do on these cloud or what you call digital based applications very different.
George
So you may need to actually have those skills and there is not a lot many people in the testing industry who.
George
Are having enough knowledge around these skills for example let it be the.
George
The fundamentals of Cloud solution like Azure or let it be advanced concepts around what.
George
Do you say cloud like maybe knowledge around the cloud infrastructure like dockerization kubernetes.
George
Kind of architecture or it may be specific type of testing that you may need to actually do on a cloud-based.
George
Application like chaos engineering for example using a tool like Gremlin or Russia or chaos studio.
George
So the type of skills that you might that you've allowed to process in order to test.
George
An application that is actually hosted in the cloud is very different.
George
And you might actually find that there are not many people in the industry who have these skills who can be readily employed.
George
So there's definitely a talent shortage there are plenty of jobs for skilled people.
George
But you may not actually have enough people who has the required skills that you may be looking out for.
George
And also one more interesting Trend that is seen is around low code low code development.
George
So instead of traditional development there are a lot of challenges that customers are facing around custom development Oneness.
George
Again there's a skill shortage to get people with that with.
George
That required skill it's not easy to get full stack developers the time that it takes to develop a custom application is actually high.
George
So customers have gone either with the SAS days package which is again a cloud-based solution like CRM Dynamics or Salesforce.
George
Or they've decided to go ahead with Solutions like outsystems.
George
Or Microsoft powerapps which are basically low code no code kind of solutions which can be quickly deployed and do a testing.
George
So how does it apply so one case is basically like the type of testing.
George
That you're trying to do on low code no code is very different you may not have you cannot actually follow the same strategy.
George
That you were following for your custom development to test low code no code applications.
George
So the the type of testing is different the speed is very important just because of low code no code since the development.
George
Is faster testing also needs to keep keep up with the pace of development happening.
George
And also I think many organizations even who have gone with custom development uh where there is devops the speed is a very important.
George
So you need to be accurate testing applications with speed but you may need a totally different skill.
George
Set to test these kind of applications but I haven't said that the.
George
Amount of budget that's available for testing has not increased in fact it has actually reduced.
George
So there are a lot of constraints that you might see around cost around making Investments to actually try.
George
Out new things or maybe Investments that are actually made.
George
On tools you may actually have constraints all over the place.
George
But overall what I wanted to cover is basically the type of testing that you might do is very different because.
George
Cloud devops as well as the adoption of no code platform you.
George
Don't have enough people in the market who actually have the required technical Talent or even the domain talent.
George
To actually test these applications and then it's also important to.
George
Uh understand the fact that there is constraints around the cost.
George
But there is definitely a demand uh to actually put in applications into Productions.
George
At a much faster Pace with high quality that intent is there but the only constraint that you might have is on the budgets.
George
That exactly given to these projects to implement testing now.
George
If you look at some of the challenges if you look at maybe from a what you say if you look from a test engineer standpoint.
George
You may actually have a number of challenges that you may have to deal with on a day-to-day basis when it comes to software.
George
Testing so you may you may be an automation engineer.
George
Who might have written thousands of tests but you may be actually facing issues.
George
Uh when you're trying to run them in a formal regression cycle or as part of a devops pipeline where.
George
You might be facing challenges with respect to a lot of tests failing because of script related issues.
George
So you you your test may not be very stable or.
George
Your application UI might have changed and you may actually have a lot of tests that are actually daily when.
George
You're trying to run it for uh what do you say a formal regression.
George
Cycle and you may have to spend a lot of time in order to maintain them the.
George
Other aspect can be like your developers are actually making all types of changes you might have patches going into production you.
George
Might have minor enhancements but you really don't know what.
George
Is that I should be testing for these changes because it is not known to you like.
George
What if what portion of the code is actually getting impacted.
George
So basically either to one thing you basically take a lot of assumption make assumptions and then do a regression.
George
Test as well as testing of these Pathways or you basically do a hundred percent coverage.
George
Testing the sense like execute everything that is available in your regression pack to.
George
Make sure that there's no impact but you're again spending a lot of time to actually run these tests.
George
So your first Factor you don't know what a state but at the same time to mitigate.
George
That risk what you're trying to do you're actually trying to run large volumes of tests whether it's automated or manual without really.
George
Looking at what is actually being changed by the development team.
George
So you may have longer test Cycles even though you're automated.
George
Uh your test Cycles you may be actually running it for long because you would actually run all your tests across different configurations.
George
As well as combinations you may also be collecting you may be actually creating a lot of data as part of your testing in.
George
The sense like you may be actually looking at maybe getting a lot of things.
George
What I would say within your test management tool or defect management tool.
George
But the data that you're likely creating is not really used for any kind of inside generation or to make any kind of prediction.
George
So that means that data is there but the data may not be used for the right purpose and then you may.
George
Also be facing challenges around the experience that you're providing to the users you.
George
May be actually writing doing technical testing but in most of the cases you we.
George
Basically fail to look at what the what is the pain point that your end customer is actually facing.
George
And we pay and we don't don't actually tweak our testing to suit the customers need.
George
So we just follow the technical requirements or the functional requirements but we don't we actually forget to look at or.
George
Identify what the pay what are the key pain points that are faced by the End customer.
George
And then what do you call carry out testing to align with those pain.
George
Points or solutions that a developing team would have created to address those pain points and then we may also have other hindrances.
George
Or other type of challenges coming in when it comes to testing.
George
You may actually face issues with respect to test database for example where.
George
You may not have enough what you call you may not get the right set of data proficient your environment on time on.
George
An average what you've seen is like it take approximately.
George
Around three to three and a half days to actually get the complete test data provisioned to an environment before the test cycle starts.
George
So you may actually face delays even to start testing.
George
So either uh so one is basically like you really don't know what your developers are changing you are not aligning your testing based.
George
On the customer's pain points you are you're executing too many tests without knowing what needs to be rigorous you're.
George
Creating data but we're not using it and then there are also what you call operational challenges.
George
With things like test data these are very I would say like maybe top maybe like six.
George
Challenges you may have another maybe 25 or 50 different challenges.
George
That you may face but it is just to give you a view of the day-to-day challenges that are actually faced.
George
By It software testing team now if we move ahead one.
George
Of the other interesting things that you might see is there are a lot of tools.
George
That have come in in the in the recent past around testing.
George
And some of these tools may actually have ai capability.
George
There are two things that you need to keep in mind when it comes to AI.
George
So it's a relatively I would say like uh maybe like eight years back when.
George
A customer had actually given me a use case where.
George
He or she was basically looking out for a solution to reduce the number of tests that they were executing for a regression cycle.
George
So they had thousands of tests that they were executing and they're looking for an AI based solution to.
George
Actually solve the problem around test execution but that point of time eight years ago there.
George
Are not many products that actually had this kind of capability AI.
George
Was being I would say AI was mostly a research topic at that point of time and there are not many companies.
George
Or custom solutions that are available which could actually do anything like test analytics to decide upon what needs to be actually regressed.
George
For a software change that is actually happening but then on I think the last I would say four years ago I think one.
George
Of the interesting things that I observed is basically like customers.
George
Who are trying to do proof of Concepts either by developing custom Solutions with the help of an innovation team or.
George
They're trying out with some of the products that are available in the market around testing.
George
Or basically testing tools which are clear AI capabilities many.
George
Of them did not go forward but I would say like maybe like a one percent of the companies are successful in implementing.
George
The solutions but in the last two years uh customers uh I think I could actually see a lot of maturity.
George
Coming into these uh at least in the commercial space even.
George
The open source phase there's not much of maturity that we could see there's.
George
Definitely a lot of maturity that is actually coming in that has come into the AI based.
George
Solutions around testing and we could actually see customers making having a great amount of success maybe.
George
To solve specific pain points that they were having around uh what do you say artificial intelligence or around testing.
George
Now so it is basically a maturing area uh it started off with what.
George
You call a research kind of topic then it became a proof of concept now customers.
George
Have been able to use AI to some specific problems you know I I don't think anyone.
George
Used to need to use AI to solve every problem.
George
So there might be 25 different challenges that you may have within software testing maybe.
George
Only there may be only like 10 percent of those challenges.
George
May only require AI to uh to actually what you call for a solution standpoint.
George
You may not need to apply AI for every solution.
George
That you have and that's one thing that you need to keep in mind you need to really know where.
George
To apply AI for what kind of solution because every.
George
Problem in testing doesn't require AI to be used in order to find a solution.
George
But there are multiple tools that exactly come in for various different purposes there are these are just some sample tools that.
George
I could actually uh potential one of the tool called Deep Blue we're actually going to take a look at that which is basically used for a.
George
Test generation there is a solution called certifier as well as there.
George
Is a solution called opky on the same company uh with whom we are partnering for this webinar rapid cloudy.
George
So we are actually having a solution which can actually do autonomous testing as well as a solution not test automation platform that.
George
Actually has capability to do self-feeding you have functionalized which is again a test automation platform with self-eating capabilities apply.
George
Tools most of you would have already heard about this product is actually a tool that is used for visual validation.
George
Or visual regression you have CE lights a test analytics based product which.
George
You're actually going to take a look around which is basically going to help.
George
You to streamline your test execution Cycles you also have a product called launchable relatively.
George
New entrance I think they came into the market around three years ago but this is a very interesting solution this.
George
Solution also helps you to cut down on the number of tests that you may be actually running for regression cycle they.
George
Basically use the AI models to actually look at your previous test runs and also recommend.
George
A test that you should be running for your regression cycle it also has a very interesting prioritization.
George
Mechanism that helps you to actually look at what tests you should be actually executing.
George
First in order to find the maximum difference if you have 100 tests.
George
That you want to run it will tell you what which are the top 25 tests.
George
That you should be executing in your cycle first so that you actually find the most important defects.
George
So there are various Solutions around test automation test data generation across.
George
All the streams of testing which have actually come into the market whatever you are seeing is maybe I.
George
Would say like 10 percent of that I would say that at least the four dozen tools that.
George
In the software testing world that actually yeah the AI capable you're only taking a subset of it.
George
The other interesting factor that I would like to cover before we move on to the AI based solution is.
George
Basically there has been a fear in the mind of testers earlier itself before the charge GPT came.
George
In the testing is going to be a stream that is actually going to be taken over by AI.
George
But I think we started hearing about AI ATS back until now I've not heard any questions you are losing a job because.
George
There is an AI solution that is actually implemented so I.
George
Think there's nothing to fear about whether it is start GPT which is a generative is AI solution or.
George
If you're actually dealing with a traditional NLP based or AI solution I.
George
Don't think there should be any appear in actually adopting any kind of AI based solution because they.
George
Have been designed to actually enable testers to perform better or they actually help testers to be more productive in the sense like resolutions.
George
Can actually help a tester to maybe create this faster in.
George
The sense it can be a productivity Improvement that can actually happen maybe there are there are capabilities within.
George
AI that can be used to generate tests maybe from something like an API definition file.
George
For example we want to generate a set of API tests and the solution actually ask.
George
You for the place where the API definition style is stored and then these kind of solutions will likely help you to generate.
George
Tests from those API definitions files in an automated manner.
George
So you're trying to automate the test design which is a good example of improving the productivity of test stress or you.
George
May be actually using Solutions like customer sentiment analysis or maybe you're actually analyzing the feedback of the customer using AI.
George
So that you understand what are the paid points that the user is actually facing.
George
Or it can be a solution where you're actually using video analytics to understand the Expressions the facial expressions.
George
Of the user to actually understand the pain points that the customer is actually undergoing when they are using your product.
George
So it can be used to improve the user experience of your the overall user experience of your product or.
George
In some cases we discuss about the challenge around the test analytics where.
George
Our test what you call date metrics where there's a lot of data that is available but you're not actually using.
George
This data to derive any kind of insights and that is when test analytics.
George
Would actually come into pixel you can actually look at the previous runs.
George
That you made you can actually look at the data as well as the defects that were locked in the past.
George
And you can actually have Solutions which actually has just send the text capability maybe a good example of sea lights it.
George
Can actually guide you on how is that you should be actually going about selecting tests for your regression cycle.
George
So it's a very simple example so there are there are.
George
So AI can also be seen as an advisor where the advices can be given on uh to a tester around.
George
The features that he or she should be testing or it.
George
Can be around the number of tests or the type of tests that he or she should be picking for.
George
Com for validating a particular build that your development team has given.
George
So automating of decision making as well as generating insights.
George
From the data that has been generated through testing is.
George
One of the good use cases where AI can be applied.
George
Now ai can also be applied to improve the speed of testing a good example uh again we.
George
Talked about launchable as well as the test what do you say DC light solution.
George
So Solutions will actually help you to upon the actual the real test that you should be running to validate a product based.
George
On the type of change that their development team has actually made which.
George
Is again a good example of how you can actually cut down on your automation cycle time or.
George
Even your manual cycle Time by actually reducing the number of tests that you're actually executing.
George
You're only executing the most valuable test this again helps you to cut down on your automation cycle time.
George
So it so AI also can help you to do testing faster.
George
And the same time there are solutions like self-healing which which will actually cover today which will likely help you to reduce.
George
The effort that you're spending on what you say uh maintaining your test automation assets.
George
So you're actually spending maybe like 10 hours to actually fix uh scripts.
George
Every test cycle uh because of the changes happening to the UI you can actually cut down all those effort because the internet helps.
George
You to reduce the cost of testing so look at AI from a standpoint where you can if you're actually where.
George
AI is trying to help you to be more productive it will actually help you with recommendations.
George
It will help you to give better experience to your customers it.
George
Likely help you to do testing faster and also at the same time it likely allow you to cut.
George
Down a overall cut cost of testing or you can actually reduce your testing budget by implementing some of these Solutions.
George
So today even though there are I would say like that around I have a commercial it is 40 plus use cases around AI.
George
You're actually going to take a look at some of the key use cases today.
George
So we'll start off with maybe test automation which is a stream where you.
George
Have seen a number of advanced is happening in the last I would say last one decade and the self-peeling.
George
Is one of the common capabilities that I've actually come in into a lot of commercial tools.
George
Even though you're not seeing this capability in uh what.
George
Do you say open source a lot but there are a lot of commercial Solutions like off key functionized.
George
And even price interests who have actually bought in the capabilities around self-paid.
George
So why is self-healing required in the first place so.
George
In today's world you you basically have application releases happening every day or every week.
George
And there is a lot of focus around user experience as well as the overall look and feel of your application.
George
So your application screens may actually undergo changes a lot and when you're trying to run these test scripts assume.
George
That you've created 500 test automation scripts for a for.
George
A product and if you're trying to run it as part of a devops pipeline.
George
Or you're trying to run it as part of your manual what you call yet manually triggering those test cases what.
George
Will actually happen is a 10 to 15 percent of the scripts may fail because there are changes that.
George
Are actually happening in the UI or it can be due to other reasons your test scripts are actually failed.
George
So assume that you ran 100 tests in the first on the first day out of which 15 of them failed.
George
Now someone has to actually do two things you can do two things one is basically if you have an NGA within.
George
Your team the engineer can actually go and look at those 15 scripts that have failed.
George
And then you you can actually try to go and fix those issues and rerun them or.
George
What you what directive the companies do is basically they try to run those 15 tests.
George
Manually because they may not have a person within the team who can straight away go and fix those issues.
George
So they approach that will take approach number two where you actually go out and run those 15 test scripts manually.
George
So you may not actually complete your tests on time your test cycle may get extended.
George
Because a lot of tests are failed because the first day itself you have 15 tests.
George
Failing and someone has to manually go and execute these tests and since you've implemented automation you may not have enough bandwidth.
George
In your team to run these manual tests also so you may really struggle to complete the execution.
George
On time now often there are Slip pages and after an execution cycle is complete.
George
You basically go ahead and run these scripts or fix these scripts and then what.
George
You call run them to make sure that it actually works the next step and this gets repeated for every test cycle.
George
So you actually spend a good amount of time either doing manual execution because your automation script did not work or.
George
You basically go ahead and fix the automation scripts manually.
George
So that you can actually use it for your formal regression site.
George
And test scripts tells you to multiple reasons it can be basically your UI is actually changed there are new objects that have been introduced.
George
So you may need to actually relearn those objects or.
George
The new objects may not be there in the object map that you have created you.
George
May actually have objects getting renamed the developers actually accidentally rename some of the objects or the developer directly.
George
What if they change the properties or some of the objects.
George
So your script is not able to recognize the objects when it is actually running or.
George
It may be an environment issue in the sense like you're trying to reach out to a website first time did not come in.
George
So you may actually have to look at what you say retrying.
George
It or your test script is very fast your application.
George
Is too slow or vice versa your application is fast your test script is slow you're facing synchronization issue.
George
So these are really some of the common problems that.
George
You may come across when you're trying to run a formal when you're trying to run your automation script.
George
So first thing what how this is going to help us.
George
If you are actually having a self-healing capability within your automation platform there are two modes in which you can actually run this dimple.
George
Event cell period so self peering the capabilities primarily used to fix these issues that you might actually face when.
George
You're trying to run your script if there are types of changes like the UI changing or new object getting added synchronization.
George
Issues or character facing environment related issues there is definitely a solution that will build into your platform that is capable of handling these.
George
Kind of Errors so that your test execution can be smoother and the maintenance of your script is actually taken in an automated map.
George
So if such a failure happens in your automation uh during.
George
Your test automation script execution time there's an engine that actually goes and understands why.
George
A failure has happened and it also tries to categorize.
George
What type of failure it is whether it is a UI it says whether it is a what you call screen karate ratio.
George
And then it tries to apply a solution on the Fly using widths we can actually go ahead and fix the issue.
George
So for example you may actually go ahead and try to increase the wait times of the scripts for example your script is your application behaving.
George
Too slow now instead of actually stopping the script and updating your variables.
George
Which are controlling your weight and you can actually straight away do that as part of your healing solution where.
George
You can actually improve the wait times of your script on what you call in a dynamic.
George
Mode or you might have new properties that have come into your object or existing.
George
Properties have changed in that case also your healing solution will can have the capability to go and fix that during run time.
George
So one is basically fixing these issues during runtime that is one mode of doing cell theory the other one of self-feeling.
George
Which is maybe a lower version of it is basically looking at whether.
George
I can actually do an impact analysis before I run myscripts to understand what type of changes have happened.
George
And then go ahead and tell the tool to actually go.
George
Ahead and fix all the issues that it is actually discovered.
George
When the auto impact analysis or the test impact analysis is actually done it can be done in two ways.
George
So ideally what happens is whether you're actually doing it in an offline mode or during runtime whatever.
George
Fixes are applied there is a script maintenance log that is actually created or updated and that is also given to the user.
George
So that user also understand what type of changes have actually gone into your what.
George
Do you say solution and then you basically run the script to make it work.
George
So the primary benefit that you actually get by implementing a self-cleaning solution like the functionized healing capabilities.
George
Or using cop key is you would have to spend a lot of time maintaining.
George
Your scripts everything is actually run in an automated manner even though there are changes happening to your application UI.
George
If the controls are getting changed the test platform will have the automation platform will be capability to go and fix these issues.
George
So you don't have to spend a lot of time maintaining.
George
These scripts which actually improves the ROI that you're actually going to get from test automation now let me show you a small example.
George
Itself on how this actually works so what I'm going to show you is basically.
George
Uh from an oracle Erp standpoint and uh yeah we have we have we are taking an example of the Oracle Erp application.
George
And we are trying to run the healing in an offline mode in the sense like this is not done during run time we.
George
Are actually using the release 22b version of Oracle and then comparing it to the previous Baseline.
George
Of 22A and you're actually using an instance which is uh used which is actually hosted in North America.
George
So we are basically running this impact analysis report on the 22p version to understand what has actually changed in the Oracle 22b.
George
Version to understand like I have written automation scripts and.
George
Also look at what exactly changed in the application and what are the changes that I may have to perform in.
George
My application in this case there are three modules that have changed and you can also see that there are five business processes.
George
And there are a number of test cases that are actually impacted.
George
Because of a small change that exactly happened in this version of Oracle now you're actually digging deep.
George
So we're actually finding out that it is the changes are actually primarily.
George
Happening has happened within the finance module and within the finance module we are actually going to a sub module called what.
George
You call procure to pay or a business process called procure to pay.
George
We're actually going deeper and then we're looking at the actual workflow.
George
So what you see here basically the test workflow that has been designed.
George
And you can actually find out that there are certain portions that are actually highlighted.
George
In red which actually show that there are a number of things that are impacted.
George
But we can actually narrow down on the particular test that has actually changed.
George
And then go ahead and find out what are the type of change that has actually gone in.
George
Uh as part of the new development that has actually happened.
George
So what you see here is you're actually going getting into the depth of the workflow.
George
You're actually trying to get into the screen where this thing has actually happened.
George
So the changes actually happened in a screen which are related to receipts.
George
So we'll have to go to that particular screen now and then figure out what uh what type of controls have been added.
George
So there is an edit receipt kind of a functionality that is available.
George
Uh within the form called account receivable and you can actually find out that there is a new control that.
George
Has been introduced inside the form which is actually causing its issue.
George
So what we can find out is there is an ad open receivables.
George
Field that has been added recently uh which is actually causing an issue with the script.
George
So now you know that there is there is something that has changed.
George
So what you can do now is you can actually tell the tool that you can you can go and self-feel.
George
This in the sense like instead of me going and updating all the scripts that are impacted what.
George
Can actually happen is you can actually tell the tool to actually go and perform a self healing where all.
George
The what you call objects that are related to this particular screen or the workflow they are updated.
George
Now and then what we can do is we can actually go ahead and run this what we call test workflow again.
George
So we are basically trying to run this workflow again.
George
So what you can see now is basically the test workflow will work fine because the issues have been fixed and you.
George
Can actually see that the test will actually pass them.
George
But at the same time it's also important for us to keep a track of what is what has changed.
George
So we can also create an impact analysis report or you can actually download and use the version of those impact analysis report.
George
So that you know what exactly changed and you can also use it for reference that's how self-healing.
George
Works is a very very simple demonstration from a package application standpoint but this can be applied the.
George
Concept is same but it's going to apply for any kind of application whether it is mobile whether.
George
It is responsive web or any of those any of those kinds of applications.
George
The factory taken a next example which is around the autonomous bot.
George
So what we're going to talk about basically the the testing of a mobile app assume.
George
That you have a mobile app that is being developed in your organization.
George
And you have to actually what you call there are the releases.
George
Happening every day and the development team is very eager to actually understand the.
George
Quality of these apps as soon as it is released for the testing team.
George
So traditionally what happens is if you ought to test the validate.
George
This mobile app you want to do the basic sanity test from maybe like five or six different devices it.
George
Will take at least half four days of time to actually get that kind of feedback but there are solutions like certifier.
George
Which Pcloudy has actually introduced which actually helps you to understand or which helps you to do something like autonomous testing.
George
So you don't have to write test cases to validate the build or.
George
The package that has been given to you you can actually use the certifier bot to actually validate your mobile banking.
George
Product or your what you call e-commerce product which is a basically a mobile app it's not just a functional aspect.
George
So basically run running what you call a set of smoke test on your application you know don't don't just limiting.
George
To the functional aspect as a part of this testing can also find out the other aspects of your mobile applications for example what.
George
Is usage frequency or what are the battery frequency you usage frequency or what are the type of CPU utilization.
George
That's happening how much of memory is being consumed so we can validate the basic functional Behavior as well as the non-functional.
George
Behavior along with the compatibility of a solution across multiple mobile devices.
George
So and a bot will actually internally a bot will actually have a crawling engine which.
George
Basically understands the data that has been given to the bot for uh for.
George
Testing I will show you an example of some of the data that needs to be paid in you'll.
George
Also have the capability to do the uix action so the the the the botan.
George
Will need to understand the different UI elements if you likely have a mechanism to report what.
George
You call and monitor the test as well as same time it also needs to understand the hierarchy.
George
Of your application in the sense like if that is also basic understanding around however applications being built what are the.
George
Screens are directly available and then it also needs to have an action engine which.
George
Basically looks at what are the different types of actions.
George
That needs to be performed on the application if you want to actually do a functional test.
George
So basically what happens the only thing that you need to do is you have to give some basic data to your application.
George
You need to update your you need to provide your mobile app APK file if it's Android you.
George
Need to actually tell desired upon the mobile device so that you should be using for your testing and then within.
George
A couple of minutes you're actually getting feedback about your product quality.
George
So that is a very interesting feature as well as the capability that you need to have you're actually dealing with a.
George
Mobile app kind of scenario where you're actually having frequent releases and the development team is very keen on getting the the the updates.
George
Around the application quality I'll just show you a small demo around this […]
George
So what you're seeing is basically the pre cloudy certifier module.
George
So you're actually uploading an APK file here which is basically a flight application.
George
So you're basically selecting the APK file on which you want to actually apply.
George
Testing or you want to where you want to do testing correctly.
George
Using that and then you're actually providing the username and password that needs to be actually used in order to run your test.
George
We're basically leaving the email address and the password and then you are also giving us one more data.
George
So you're actually putting labels there there is one more field.
George
So you need to understand what flight you need to book.
George
So you're actually giving the flight to from and to information you're actually booking Delhi and one more location.
George
And then you're actually telling the bot to run the test.
George
So the bot you can either make the report public or private depending on the need that you have.
George
So the session is active now and once the session is the complete.
George
You basically within a few minutes you actually get a report uh generated.
George
And you'll actually get an email also with the details of the report on how we should be downloading the report now the testing.
George
Has completed so you're actually getting a report of your testing you're actually going to the platform.
George
So what you see here is basically there are around seven devices on which the test has been run on an average it's.
George
Showing you the the CPU usage memory usage Network usage and battery usage for this particular app you're.
George
Also seeing across each of those devices how are the performance of.
George
Your application again from a non-functional standpoint there's this Benchmark against which you are comparing it you.
George
Can also get some insights towards the portion of the dashboard around.
George
The various things that were done there is also a crash test that you have done which basically as well as the acceleratory test.
George
So crash is nothing but you're actually testing these apps and opening every screen in the app and then you're performing testing.
George
So it's not done in a very structured way randomly opening screens it's something like a monkey testing that you do.
George
But you also at the same time you also measure the performance of the app and you can actually see this results across.
George
All these devices that you use for testing in this case there are seven devices.
George
So across or seven devices are your application has actually behaved.
George
Something that you can see there are detailed error logs if there are failures happening you can actually get specific screenshots.
George
And also you actually see each of those screens and then I click covered as part of the testing.
George
So the beauty of this is you're not spending any effort to do this testing everything happens within the background.
George
So that's the main advantage here you can actually see all the screens that were there.
George
So this is what you likely see within a crash test and similar to this and there's also a detailed log file that.
George
You see now there's something called exploratory tests also which.
George
Is again what you say about fracture test I would say this is actually done as part of workflow.
George
So you basically look at what you call seeing the the workflow of your application here.
George
So basically looking at all the screens that you have and you're actually seeing the detail report.
George
So the exponent testing is basically created as part of a workflow you basically go from your logging screen to the next screen home.
George
Screen maybe booking reply to a done and structured manner where are the crash test or the board testers.
George
The initial test that is done is done in a random yeah the other important use cases around visual testing.
George
So there's a lot of focus around user experience in uh especially.
George
With digital applications and it is very important to make sure that your applications look and feel exactly good across various platforms.
George
So if you're dealing with banking you may have multiple channels on which your customers may be actually performing transactions.
George
So it can be a mobile application it can be internet banking.
George
It can be ATM it can be a kiosk it can be a pause.
George
So there are various tunnels that are available and it's also important for as a testing team for us to actually provide.
George
The same experience users across all these digital channels but if you want to validate the look and feel of maybe a website across.
George
Maybe different what you call configurations it may be very tough job because if you just want.
George
To validate a single field that is available inside a web page you may need to look at maybe at least five different factors.
George
You may need to ensure that the field is visible you.
George
May need to actually make sure that the X and Y coordinates or the position of the field is correct you.
George
May need to look at the height and what you call width of the what.
George
You call particular object of field you may also need to look at the background color.
George
So these are basic five fields that you five things that you may need to actually check on a particular field that's.
George
Available on a page and then similar to this you may have 20 different controls.
George
So overall it's 20 into five you're actually looking at 100 different checkpoints that need to be validated against a single page.
George
So doing this manually is impossible I would say because.
George
You may need to actually spend a lot of time in order to do this kind of manual verification.
George
And it may not be accurate also because you might have seen a lot of spam mails.
George
Coming to you where it tells you to actually find out maybe.
George
The differences between two images you may struggle to even find the basic difference it's very similar to that it's.
George
Very difficult for us to actually go to every page.
George
And then find those things that are not working based on the visual verification.
George
That needs to be done now you want to automate it also in the example that I mentioned we are talking about 20 different verification points or assertions for a single page […]
George
You're talking about 100 different assertions that needs to be made on a single.
George
Page and the same page needs to validate across maybe two different mobile operating systems four different uh devices each or.
George
You may have to actually do it with portrait as well as landscape.
George
Modes they are actually talking about maybe like more than 32 different combinations on which.
George
So 32 combinations on which a single page needs to be validated.
George
So it's impossible to do this even using an automated solution that is when a solution like aptly.
George
Tools would actually come into pixel where it became where the platform offers you carefully to do visual verification using an image recognition algorithm.
George
So basically uh you can either use the apply tools capabilities assets or what you can do is.
George
You can actually use the capabilities of a test automation tool like Cyprus selenium.
George
Or appium you can use the SDK of aptly tools use the SDK within your framework look at calling.
George
The verifications or the capabilities of apply tools from the framework and once you call those capabilities once.
George
You invoke those capabilities what happens is the apply tools SDK.
George
Actually has the capability to extract the Dome of your page or the document object model of your page.
George
And it actually has the capabilities of practically using the the capabilities.
George
Of the visual grid solution it actually has the capabilities to actually go ahead and create snapshots.
George
Or screenshots of your page based on the test environment configuration that you're using.
George
And these configurations are actually taken or compared or these screenshots are basically compared with the Baseline.
George
Images that were actually created for the same test configuration.
George
So you're basically doing an image to image comparison for each test configuration that you might have a test configuration can be maybe a.
George
An environment that you're using for maybe iPad or an iOS environment with iPhone.
George
And you're trying to do the testing on the new build to see whether the look can field has changed.
George
So you compare that iPad what you call screenshot that you're actually captured earlier which is basically called a baseline with a new.
George
Uh what you call run now the new screenshot that you have captured as part of your test cycle and then you're trying to find out the matches the mismatches between these two.
George
So you're actually doing visual what you call comparison of each and everything a.
George
Field within your web page and then you're trying to report that as an anomaly.
George
So basically it's not just apply tool that actually has the capable PLC.
George
And other capability other tools also have similar capabilities but it's a very good use case from a compatibility testing.
George
Standpoint you can actually save a lot of time around writing.
George
This or designing tests to verify the visual Behavior it.
George
Can actually bring a lot of accuracy and then also look at improving.
George
The overall what I would say the the quality of testing as well as the experience that you are providing to your users.
George
The next is created around the unit testing so you.
George
Might have I think people have been working our Engineers have been working in software testing for the last two.
George
Decades or more one of the interesting things that you might have seen us the.
George
Level of unit testing has not improved a great action I've not seen visible improvements in the coverage of.
George
Unit testing and that's the reason why you actually find 1998.
George
Or 99 percent of the issues uh during your black box testing page because the development testing is not very complete.
George
I don't blame Developers for it because in the current model of working especially.
George
With agile and devops developers hardly get maybe two to two weeks of time to actually do.
George
Uh to make it to actually generate an understanding or to where.
George
You could understand the user story what they have given what have they given to them they.
George
Also have to do design they have to actually write the code they.
George
Had to do what they call unit testing and then they have to hand over the code to the developer to the testers on a daily basis to actually do testing.
George
So a lot of work that needs to be done on the design and coding side.
George
So developers do not get a lot of time to.
George
Actually design unit test case and that's the primary reason why the development test coverage is low.
George
Now someone has actually looked at like this as a very unique problem and a company like called.
George
Who has actually come up with a solution so strictly on Java based what we call code where.
George
You can actually look at generating unit test cases using AI based solution.
George
So instead of actually a developer writing unit test what do you basically do is there's.
George
A solution that can actually look at the code look.
George
At the rules within your code look at the logic between their code and then look at creating tests that are intelligent.
George
Tests which can actually help you to improve the unit testing coverage.
George
So the main challenge is not on executing these tests uh even though it's not a whatever unless.
George
You actually have thousands of tests if you have a very small code you.
George
Can easily generate test using the solution then run it as part of your devops pipeline.
George
So they prove such a solution that actually has the capability to go ahead and run this based on the changes that.
George
Your development team has actually done or you can actually generate tests unit test cases in an automated manner without actually the.
George
Developer spending time it also has capabilities to do the coverage analysis it tells you exactly what portion of the code has been tested […]
George
The developer might have created it also has refactoring capabilities where it also tells you how that you can actually optimize your code.
George
So that it can improve on a coverage as well as testability and it also has a module which which actually gives you insights.
George
Around the type of test that if you're running for example I would have written 5000 unit tests for my entire product.
George
So based on the change that your developer is actually making.
George
Uh it's similar similar to what you call a solution like test analytics dipflow.
George
Also the capable to recommend the types of the number of unit tests that you should be actually running in order to validate.
George
The change uh that is actually going into production so it's a very simple tool it basically works from a development.
George
Environment you can the best thing is to basically use the IntelliJ the.
George
Demo even though it's right clips that you can use to IntelliJ also you may need to actually go ahead and add.
George
The plugins related to div clue and then you have to actually pick the code base that.
George
You want for which you want to generate the unit test.
George
So initial step was directly add the plugins of diff Loop by going to the path where you have installed it.
George
And once you're done with that you have to actually go and fix the working directory periodically.
George
Point to the solution that you're trying to so what you say that you're trying to uh uh test.
George
And there are certain arguments that you may need to pass for example you may need to actually pass on some standard arguments.
George
Which are required for naming the tests that you're creating and also these uh you're.
George
Also looking at parameters like whether you want to overwrite your existing test or you won't append to the existing test.
George
So there are a lot of small parameters in our standard.
George
Things you don't have to worry too much about it but once you give these parameters once.
George
You're done with the arguments what you have to do you have to actually save this and then you are directly refresh your environment.
George
So you know you will get access to the code base with HD unit tested.
George
So there might be multiple it can be Java package with multiple classes.
George
And it can actually go and generate a unit tests for all of those classes.
George
That are available within your code it actually looks at the method level and then generated.
George
Test cases now the code is there it's basically a mobile a core banking code that is there where you're actually going ahead.
George
And seeing what you call a package called account there's an account Java.
George
File there are around four classes in it and now we are trying to use the deploy.
George
To generate unit test cases for this Java code that is there.
George
So exactly the starter the generation it is actually going through the code.
George
And then exactly trying to find out uh or generate the unit test cases for these four classes that it was actually discovered within.
George
This Java file so we look at every class and all the methods within the class and then generate unit test cases for.
George
Each of those methods so this is a beautiful way of improving your unit test coverage especially if you're a Java shop.
George
Uh it's easy for you for the developers to go and use the solution not a very costly solution there's an open source.
George
Or a free version or a community reached national using.
George
Which you can go ahead and create these tests now whatever you're seeing is basically the tests that are actually created.
George
For the account class it is actually generated account test.java.
George
File which actually includes all the tests that are applicable to that particular file.
George
And then you can actually run it as a junit test.
George
Now so I think exactly running those just now it also tells you what the coverage that is there and it also tells you what portion of the code is being covered by.
George
These tests there by these methods that are not covered but on an average it actually generates somewhere between 50.
George
To 60 percent of tests for your code and then you may need to do further optimization to get better coverage.
George
The last solution that I would like to cover is around test analytics.
George
And this is again one more pain point that you might be having where.
George
Your developer might be making major changes to the code or in some cases developer might be fixing a couple of bucks.
George
And then you're actually putting this bill for testing so you really don't know what you should be testing because it's.
George
Very tough to understand the change that has actually gone into the code.
George
So you basically end up doing running too many tests or in some cases you only run a small subset.
George
Of tests by taking some assumptions and you can also leak defects as part of this.
George
So you're trying to solve this problem around the number of tests.
George
That you want to execute or the number of tests that should be executed as part of a regression.
George
Test cycle we're actually applying test analytics where a solution like C lights will.
George
Have the capability to actually recommend the types of tests that you're actually that you should be running as part of your regression.
George
Cycle or as part of your test cycle so how this works is basically you may need to execute all your tests at least once.
George
So what happens during that time is your C like solution okay listen to the tests that are getting executed by to Bob those tests to a particular […]
George
Within your application code you don't have to give the code it basically does byte cut by twice comparison.
George
It extracts the code from the byte code and then it actually ties to do the functional mapping.
George
So CLI is now knows once you complete the Baseline run C lights know that what portion of what.
George
SK is actually testing what kind of what you call functionality or capability within your application.
George
Now you're done with that Baseline run now we have a new build that a development.
George
Team has actually given you at that point of time you basically run an analysis using C right.
George
So C lights you actually go and find out what your developer.
George
Has actually changed there are three things that your developer would have done one.
George
Is basically developers would have added new methods or new portion of code or new classes to your what you call application.
George
So you can actually go and find out it tells you that for this portion of the test that this.
George
Version of course is actually added by the development you don't have any kind of test that is available you.
George
May need to add new test or it may tell you that some of this code has been removed.
George
And you don't have to execute these tests so you can actually remove some of your tests.
George
And now it'll also tell you that the developers actually changed maybe 15 different methods.
George
As part of the what they call build that has been given to you.
George
So you need to actually run Tesla belonging to these methods and then a couple of additional tests.
George
So it can be any type of test it can be a manual test it can be an API level test it.
George
Can be a UI based workload as any of those tests.
George
So it basically looks at your previous Baseline run it compares.
George
The Baseline run with the code that has been or the new build that has been given to you and it actually tells.
George
You what kind of test you should be running against.
George
Each of those what you call a build that has been given to use basically a solution.
George
That actually helps you to cut down on the number of tests that have been given to you it.
George
Actually allows you to reduce the number of tests and also save a lot of time and cost around running these.
George
Tests that's basically a good solution but if you actually have an application that actually has more than like four thousand.
George
Or five thousand tests this works very well because you might be spending days together as well as there will be multiple resources.
George
Who are actually engaged would be engaged to test these Solutions.
George
So basically another good example of how analytics can be used in order to cut down in a cycle time.
George
So these are few examples there are many more examples but overall we.
George
Are trying to implement any kind of AI solution in an organization.
George
Couple of things that you need to keep in mind is it is you.
George
Need to stay set the real ex or you need to set realistic expectations with your stakeholders.
George
It may not be good it may not it may not be what you call practical to get results on day one.
George
So you said that very practical what you call uh expectations with the stakeholder.
George
So that there is no what you call questions asked if there are delays in getting implementations.
George
Done it's not something like the first day itself after the AI solution is actually implemented you actually see start getting the.
George
Results it will take definitely some time uh to actually get the get to the results.
George
So similar to this when you're setting a deadline for implementation also you need to make sure it's really a realistic timeline because.
George
The accuracy of the solutions will only improve on a period of time you are now actually going to get maybe 100 accuracy.
George
On day one or even 70 accuracy on day one especially.
George
When you're trying to do it with a what do you say a custom solution it's actually a periodic improve on the right rate Improvement that you can expect so you […]
George
Now the second thing is basically most of these Solutions will require enough data to be put in to get accuracy.
George
So the quality of data that needs to be given.
George
To your development team or to your IIT for implementation should be pretty good and they should also be giving the bit of data.
George
So if your data volume and the data quality is not good do not expect a lot of accuracy around the date the.
George
Solution it's very important to keep keep that also in mind and then none of these Solutions I would say 80 Solutions.
George
Will definitely require customization it is not like one size would actually fit all.
George
So you mean you may need to actually customize solution you may need to actually rework on the model to see the required results.
George
So it's not it definitely takes time to actually implement.
George
But once it is implemented it will actually run forever but that one should actually focus on continuously improving the quality as.
George
Well as show set some realistic timelines as well as set the right expectations if you're actually getting into a state of implementing.
George
AI in your organization so I'll take a pause now I'm done with my slides now I'm open for.
Moderator
Oh thank you so much for that uh insightful presentation.
Moderator
Um so we've received several questions from our attendees um so we'll do our best to answer as.
Moderator
Many as possible in the time that we have left.
Moderator
Um so here's the first question for you so how would a self-healing function in.
Moderator
The case of defects does the AI model need to learn and comprehend changes in the UIL workflow similarly.
Moderator
How do humans learn and understand changes in an application.
George
Okay so the AI solution would like definitely have the capoe.
George
To understand product defects for example you're actually getting run time errors you're actually not able to go to the next screen.
George
So that level of intelligence should be there so it should be able should be able to understand these.
George
Type of issues for example one of the one of the uh reasons for uh failure can be a product defect.
George
So the AI solution should definitely understand the different types of errors that actually comes in as part of the execution with JavaScript.
George
Error it can be a page not found kind of an error.
George
So it's not going to actually invoke the helic engine to make any of those scenarios work.
George
So it can be a synchronization issue this environment will be having slow or a UI change that.
George
It sees basically it can be UI property change or.
George
A change in screen position it actually has the capability to go and rectify.
George
It otherwise it may not act on it because it will be very risky to actually ignore.
George
Those things and whenever it's actually making it change it.
George
Asks the capability to go and fix it or record it as part of the log files.
George
So that uh what do you say it is easy for people to actually understand why a specific saying is actually implemented.
George
So that's why works but definitely every healing solution that you're trying to implement is the commercial loan or open source.
George
One should actually have the ability to differentiate between a product error it can be basically as an.
George
HTTP error a JavaScript or anything that you see on the screen.
George
And maybe uh fix it and it's not going to fix anything that is actually what you call like.
George
A business rule or a functional validation for example you're actually seeing a wrong account balance.
George
So it's nothing like self-feeling would be required to correct those these.
George
Are not valid depicts and your step will fail and proceed that's how it works thank you.
Moderator
Okay so here's the next question um so how do you foresee the role of AI in test data generation.
Moderator
And its impact on testing coverage so what benefits and challenges might arise from leveraging AI in this area.
George
Okay so test data is one key where I think there.
George
Are use cases that you have seen where AI based Solutions can be used.
George
So let's actually look at maybe testing an aib solution or maybe an RPA based solution which basically may.
George
Be uh using a set of PDF files as an input.
George
So you may need to actually generate thousands of PDF files using various permutation combinations of data.
George
So it definitely need a solution which can actually create that kind of a document.
George
So it's not the structured data you may also need to generate unstructured.
George
Data it can be images you're actually testing something around images you may need to actually what.
George
You say look at a solution that actually has the capability to generate the images.
George
So generation data using AI capabilities is something about structural manufacturer.
George
Is something that will actually be needed in the future because there are no too many solutions coming into the market exactly.
George
As generative AI capabilities or traditional AI capabilities and there can be other use cases as well for example you.
George
Can even develop Bots tab Bots using which your testing team can actually provision.
George
Data or create data based on when when they need it.
George
So basically an on-demand or self service kind of model where AI can be again applied to actually generate.
George
A solution that actually can provision data so these are very simple use case that I can think about one other bot.
George
That you can create using AI to provision data second is basically using synthetic data temperature tools.
George
That can actually create both structured and unstructured data for your testing needs and if you can actually and one of the challenges.
George
That you might face within data generation using synthetic what you call capabilities is around data models.
George
So if we can actually have some kind of intelligence.
George
Where your synthetic data platform can listen to a database or analyze a database and then come up with data models automatically.
George
I think that will be very very uh interesting kind of a solution.
Moderator
So uh here's one final question for you George so uh do you have uh any.
Moderator
Um suggestions for someone who's switching their career from a functional or automation tester to a AI tester.
George
Okay so I think I don't think you need to there is unless you're actually going to test uh something around AI.
George
Uh in the sense like if you're the two ways in which you can apply or.
George
You can actually be involved in a Oneness you're going to test an application.
George
That actually has AI capabilities for which you need to actually have background around data.
George
Science data models and all of those things which may require some.
George
Kind of training and upskill but if you want to use an application that actually has AI capabilities.
George
For example let's take an example of maybe uh a c likes you know you have to and just understand this.
George
Understand the solution but be very clear on the problem that the tool can actually solve.
George
But that doesn't require a lot of upscaling you only have to do upscaling.
George
With respect to the tool but if you're going with the former where you're going to test an AI application.
George
Then you need to actually learn the fundamentals of AI data.
George
Science the different algorithms and all of those things and the types of testing.
George
But very diff very very different from what you've done earlier but using.
George
An AI based solution to only requires knowledge around the solution around the platform or the tool which can be actually easily be easily.
George
Be what you call a uh I think you can acquire the knowledge in a pretty fast.
George
Manner I think in a couple of days of time it is something that I think everyone should be actually looking at.
George
So you should need to look at what are the different solutions that are available.
George
And then look at what how is that you can apply those solutions to solve the problems that you might be having that's only intelligence.
George
That you may need to apply but learning a solution like C lights or.
George
Maybe diff blue is not going to take more than two weeks of time.
Moderator
All right all right so looks like we are out of time for today's session.
Moderator
So in case audience if you didn't didn't get your questions answered uh no worries.
Moderator
Uh so we'll reach out to you directly with the uh with those questions answered from George.
Moderator
So we love seeing your engagement and enthusiasm during the webinar.
Moderator
So Before We Say Goodbye uh remember that we'll send the recording of This webinar to all attendees.
Moderator
So feel free to watch it again or share it with your friends and colleagues.
Moderator
And a big thank you from both Pcloudy and guest house for joining us today we.
Moderator
Hope you enjoyed it and learn uh something new um.
Moderator
So stay tuned for more future events and collaborations between Pcloudy and test house.
Moderator
So thanks again for being with us today and have an awesome day thank you everyone have a great day bye.
