ContinuousALM - Application Life Cycle Management for GxP

Nagesh Nama

01 May 2024

Chapters

00:00 Introduction and Background

04:59 Key Features of Xqual

09:59 Applying ALM to Various Industries

14:59 Eric Gavaldo's Journey in Testing and Validation

19:30 Programmability and Control with REST API

21:59 The Power of a Platform

24:19 Integration of Xtudio with Jira

27:44 Optimizing Performance and Control

30:14 Implementing and Customizing Systems

32:20 Distributed Architecture and Agents

34:06 X-Agent: Executing Tests on Specific Agents

34:34 Introduction to AI in Software Testing

35:02 AI's Role in Task and Document Management

35:21 Learning from Tools like Jira and Confluence

36:02 AI's Impact on Requirements and Test Descriptions

36:50 Introduction to AI in Software Testing

37:45 Automating Testing Processes with AI

38:12 Ensuring Completeness and Correctness of Requirements

38:46 The Future of AI in Software Testing and Quality Assurance

Transcript

0:00 welcome to continuous TV where we transform Life Sciences with Dialogue on AI and

applications so he is the founder of exqu like I said and you know he is uh

not new to this space uh he has worked in this play in this space for multiple

decades so I'm very glad to have him and before I start just to give you uh some

background here I have known Eric for more than eight years we have worked uh and we use xqu internally and now we

have continuous Alm which is which is built on exall and marketed by xlm so

that's a reason for having in the podcast to educate our audience on how a

good uh Alm package should work and what are the key features uh that xal has

which is very important to not just software development for software validation for um uh any any validation

project in fact we can use this exqu for not just for software but also any other

equipment validation facilities validation can be used and we will go into details very soon uh Eric very nice

to have you thanks for taking the time and if you don't mind can you please introduce yourself yes sure so thanks

thanks first to to invite me gu it's it's always a pleasure to speak with you uh you are such a valuable person in

this uh testing field so it's uh very very nice to speak with you introduction of myself yes so I'm so I'm Eric of

course I created xqu uh in 2010 I think

something like this so so it's start to be so big company that that with a long

life let's say uh before that I was as you said I was working on on testing uh

for my whole life so I was QA manager and uh I used to be a research and development uh directors and so a few a

few companies but mainly my life is is around testing so it's um that's that's

really what what interest me let's say and uh and today so today we we have

company called X exqu and we deliver a product that is called xtudio but probably we're going to to remove that

name because people remember about xqu and uh so with this software we are

helping our customers to better manage all their requirements specifications

test test campaign bugs etc etc so that's really what what we do today Eric

you know your name attracted me my company's name is xlm you are xqu now we

come together maybe maybe we getting to get married

soon no we already married right because we are marketing your product so you're already married you

know ER U one thing uh when I started looking at Alm packages when I started

xlm that's when I started looking at it because I wanted an engine to support my validation projects for the life

sciences I did evaluate over dozen products uh both SAS open source non SAS

product yeah then it was very clear to me when I looked at your product I built

a very nice grid comparing feature by feature which so that I could evaluate

for myself right because that's very important to me yes so I chose xqu in 2016 and since then it is our Alm

package at xlm and we've used it on multiple projects you know and we have built our own uh uh some of the

reporting uh tools also that extracts from xqu and builds a word Word documents which we can share with our

customers we have done that also to support our own implementation and recently I asked my

team to go back to the drawing board to without telling you we evaluated again to see if anything has changed in the

last seven eight years right so I said you you don't have to uh be partial to

xall I know xall is a good partner but don't have to be partial let's be very objective look at how the landscape is

today I told them if you think you find a better product than exqu you can you're free to select it we will switch

it okay they did their homework they did a very good homework you know we again

looked at about six to eight products we looked at they came back to me and tell me told me that n you know we to we

cannot find anything that's comparable to xqu and I'm not making it up I'm not exaggerating it and I would consider

myself as a small expert in this field because I do look at many products and we have our own software products to be

honest with you you your product is very feature Rich um and supports everything

that we want and more you know not just that you know so it's coming from me right uh so I want to hear it from you

there are so many Alm products right out there SAS non SAS you know open source like I said you tell me Eric you know

what is so unique about xqu okay uh so first of all I moved it's a nice testimony I will record this and and put

this on my own page yeah so so we have uh so you're right there are there are

many many many um Alm on the on the market uh many many uh with different

forms with different price with different so it's it's a huge maret so if you if you ask me what are the the

differentiator what makes us a little bit different what we made what what we make differently from others I would say

there at least six or seven maybe eight uh features that we are almost the only

one to to do the the the last one that we did which is we have excellent

feedback about that is the fact to to provideing in the version 12 that we've

done just right now um it's a visual graphic uh tracability metric so you can

see on the same screen the same page you have a graphics where you see um basically your releases your requirement

your spec your test your bugs and everything is related and and linked

together with some some some links and and now since last week we can also edit

the links from that page and that this is typically one one thing that people

love a lot because um that's very easy to use and it's it's going to to help

you to to gain some time in your life so this is one of um of course probably

we're going to discuss that there is also the digital signature that is

useful for for he science uh companies oriented companies uh one of the things

that we have from the very beginning is um our capacity to integrate with many

many launchers so we have many many launchers that allow us to to connect to some many many different different um

test automation framework so if you use test automation framework which is not

mandatory I mean a lot of people are going to do only manual testing but if you are going to do automated testing uh

then you are not limited to to choose one specific tation framework

among maybe 10 in our catalog we have n 90 different launchers so we have almost

everything on the market and uh and if one of a customer ask for for some that

we do not have yet we we do it for free for our customers so this is really open

to to any for for the automation part we are open to many many many uh things uh

another thing that we do also is um that we have is a contextual chat this is

very small thing but nobody does that and it's it's very useful so because

it's it knows for people who knows what who are the pre Amigos BP you have people working on the on the

on the business so business business analysts um the testers and the product owner and

developers Etc so this tool allows them to to communicate better uh from the

source of what they are going to develop and so on it it again it's something very very very useful because there is

less misunderstanding in between all these people and at the end theity is

much better yeah go go ah we have we have also maybe maybe two or three other

things uh we have also the the the exploratory testing uh so we we we do

that and uh and at the end we have something specific because once you have

done your exploratory testing with xio then at the end you can also generate

comp automatically your your test your scripted test at the end so you you

don't the time that you spend in exploratory testing is not lose it's must lost at the end you you get back

this time and evaluate it correctly uh and we have also the canb view which is

very um like by our customers because when you manage some requirements and

some test Etc these things have status priority severi these kind of things and

um it's always difficult with standard system to just select one change the

status or change the priority and then select another one and do the same Etc with the can man view that we propose

you have a a big screen with all your element and you just have to drag and drop from one column to the other one so

it's very nice and maybe the last I will stop after that but the last thing I

think is probably we there are there are some competitors that have this but we do it much better I think um this is the

the the Gin editor we have a edor OS

that is completely integrated in our system and I think that's very good point for yeah I think Eric you know you

bring up very good point with the fda's new CSA guidance right they are actually

uh encouraging companies not just to follow one framework uh they want to follow different Frameworks based on

criticality right if it's a very critical software yeah you need to go the traditional approach know uh plan

everything properly you know write your test scripts requirements specification test scripts and do it or it is not that

critical they're encouraging people to do more of the exploratory testing things of that nature which your software already supports you know um

and it supports bdd which uh which is very important like you mentioned the gkin model so they can take advantage of

that and also convert that over to real test scripts right over time so it's all in one package they don't have to go to

multiple packages now and considering the integration that you have you said

over 90 plus launchers that is amazing right it's what a database of launchers you have the audience may not be very

familiar with these launcher concept can you just explain how the launchers work

please yeah the IDE um by the way that that the very first ID that I that I had

when I started uh excor and the very beginning the goal was not to to do a MM

the goal at the very beginning was just to build a generic um test execution engine so and

that's what something that I just made for me at the very beginning I was I was managing a lot of different types of um

automated test and I was working on in some companies in working in Telecom and

uh unbeli softwares and such things and uh of course in this Hills you do not

have many many um test automation work so you need to build your own so so the

idea was to to try to be able to to to to drive all my different type of test

that I built on systems that I develop myself but to to to drive them with a

unique interface so I can I can mix test and say I want to do this and then and

then and then at the end I I get all the results and I can do a refa and um so

technically the thing was to separate what you get as an interface with the

end user and for the reporting from uh from the end test that you have and you

have some kind of module some kind of plugins let's say uh and each plugin can

handle one type of a test and so we have in our launcher catalog we have one

launcher for selenium one launcher for junit one launcher for something else

Etc and uh and and if you have one even if you have some propriatary test

framework that you develop your on your own because you need it to do something very specific for your very specific

product then you can even develop your own launcher yourself which is propriatary that you will be able to

plug in X studio and from X Studio it's transparent you can just say I want to

execute this and this and this on which agent with this configuration on which environment etc etc and at the end get

the results uh naturally so just to continue on the conversation right like what you are

discussing just now so I'm very familiar with X agent because we use your software so the idea here is that I

could use many other products for test automation it's very easy for me to use

your existing launchers to bring the data into xqu correct launch the test

and also bring the data back into xqu so that way xall becomes my single source

of truth right because that way I have all the audit Trails I have all the data I can do all the reporting where I can

use multiple one maybe multiple test automation Suites so that way it's nice

to have one console where I can bring it in so what you're telling me is that even if there is no launcher we can use

your API to to talk to other packages cor I can build my own connector or

launcher correct yes yes so if one customer comes to us tomorrow and say

hey I see that there is a new test framework automation framework on the market it seems very nice and I want to

use this then um they just tell us and we do the launcher for them you don't

have to take care of of this it's it's free it's it's part of the package if you are licensed with us we we do this

but if you have a proprietary uh system so of course this this is something

different because this is something that is very you for us after that but if you have a propery system and then you want

to make your own launcher you have SDK you can develop it yourself we can

review even your code if you want there is no problem with that we provide the SDK it's open source you you can you

have the build script everything is is is available and by the way all the

launchers even the ones that we provide in our catalog our open source so you have the source you can modify and

customize them that's wonderful LC not many companies do that first of all they don't have this many launchers 90 plus

and also you are so open to helping us helping the customer develop their own launchers right you'll develop it for

them or they can develop it but it'll be part of the catalog you know that's that's incredible so we are not just

talking about uh your uh launchers just for um test automation Suite right we

have for devops correct yes uh we have for Jenkins for example and we have for

standard package like jira to bring in the requirements correct uh so can you just in in in 15 seconds can you just

talk about other launches apart from test automation yeah so the launchers are just to integrate with some test

automation framework so so that's one thing but we we we saw we saw so a lot of customers that for for example for

the book tracking many of our customers when they came when they come to us they

already have a book tracker so correct we we provide the book tracking in next Studio but if you have a b tracker it's

very useful that we allow our customers to to be able to integrate with them so we have many many many connectors that

we provide that allows to to connect to third party system e for requirements management

offer bug tracking management so if you requirements in in some system and your bugs in some other system so for example

if you have your your requirement in Jaa and if you have your bug in bugzilla M

or in so you can integrate with these two systems so if you use xio as a main Alm

you will see in next Studio your requirements that are coming from Gyra you will see your bugs coming from gitl

and uh and both are in your interface and you can link these items to the rest of the the ecosystem so the test and and

so so so there is so there is there's this level of integration for POS requirements in the Bo and you have also

as you said uh integration at uh continuous integration level level so

that you can um there are many many ways to to to start and and run a test

campaign from X Studio you can run it manually or with a script or you can run

it from directly from your continuous integration system we have also integrated some schedule uh in next

studio so you can make sure that it's a session for example is executed every Monday and every Friday at 2: a.m. in

the morning and these kind of things so are many many ways but continuous Integrations is of course one one level

where we are interfacing and we also provide a rest API so if if you need to

to to drive and control and to create some elements or delete or move Etc

programmatically you can also use our W API to just uh interface with our system

more globally that's very nice you know um so from a integration point of view

it's very robust for requirements or specifications bugs right uh automated

testing and devops so that really covers a lot of different areas and bring everything back into xquad so that

becomes my source of Truth so talking about source of Truth what's important

for our life science customers and What attracted me to exol was one of the main features was your version control like

every object in your xqu like whether it's a requirement specification a test case you have very strong Version

Control and also you have audit Trail so I can tell who did what which is very

important for part compliance and security you have security which is very granular role based security which is

very important and lastly you have no two more items one is freeze and freeze

which is important like as a quality person I can probably freeze something so nobody can change it only a quality

person can unfreeze it that way it's locked so nobody can change it that way it becomes let's say one version point

to be locked nobody I don't want anybody to change it they can clone it if they want to but I can always freeze it I

know and nobody can change it that's important which is already you have it lastly the electronic signatures so

these are what's important to me from a part l so you already have that so it's

that's the reason we we chose to partner with you as well because you know it's not only robot from an Alm point of view

but also from a part Lev compliance point of view uh it's very important and you took it one step further with the

bdd and everything else which is now fdsa is also recommending that exploratory test cases and everything

else it's again industry standard right gkin model is industry standard PDD is industry standard and you already

Incorporated that so it makes it a very nice sophisticated platform which is

robust you know and integrates well with all the things so that is uh very good

but I just want to kind of lay it out for the audience so they understand how powerful of a platform this is it's not a software package to me it's a platform

right so you can integrate nicely and bring everything in and it gives you the nice Graphics like reports kpis

traceability everything is there they don't have to worry about it you do it once it's already

there that's the ID the ID is have something completely integrated and where uh your end user does not do not

have to go there and there try to the idea is to have something compile all the data are compiled in the

same interface and you have all your data in one place yeah so now what is

can you just give the audience an uh idea of one of your largest installs how can exqu scale up can you talk about it

pleas yeah so we have so if if we look at our customer Banks right now today we

have H very small customers uh with 10 people we have medium customers and big

customers and we have customers in almost all Fields so we have banking

Insurance Health Science retail Etc so it's very very large it's it's very hard

for us to to do a Persona which which is a problem salesmen are are saying oh I

cannot find my but anyway uh so we have so your

question was about the large um large business so we have a few

um uh we have a few very very very big users for example uh for example we have

Intel Intel uh Corporation we have airus which things

aircraft uh we have the French government we have with have the Belgium government as

well we have act Automotive which is doing a lot of things for for for

for automotive automobile industry have some instruments like s s is a very big

Insurance in Europe uh in terms of insurance we have also Pinacle in the US

which is quite quite a good insurance I think so we have we have more than 200

users we have um we have at least 15 uh customers that that have some large

scale like this in this in this term

um these these customers have usually May between 10,000 and 40,000 test cases

usually in their in their systems so it's quite quite a lot we have one of

our customer I I will not give the name because yeah it pretty strange but they

have I think they're using Gyra for for the requirements yeah and and um I

worked on on on on this uh recently to for for doing something specific for the

integration is Gyra and they have something like two 2,300 project in

Gyra I never saw that I never saw that in my life corre so so we needed to to

to integrate x studio with that we need to do some filtering to not get all this

stuff in exio so yeah yeah usually it's um it's in between 10,000 and 4,000 test

Cas and uh a little bit less in terms of requirements usually but that's not a

problem right from a scaling point of view it's not an issue for problem I I think we did we did some

test some performance and stress test yeah at one point I think we've we've

went to when we did that a lot it was something around 6 60,000 test Cas

something like this and we stopped because we said we never we will never find a customer that do

that and uh and now we see some people some customers with 14,000 so we say

maybe we're going to to redo this again to see where we can go but uh I don't

see any any major issue with that because so you know when we started to

to create a studio we spent a lot of time uh working on the basics the

foundation of the software and and we f a a lot and made a lot of experiments we

tried a lot of things Etc libraries etc etc and until we found something where

we we have the best performance and we have the best control on our software and we are not limited by a third party

libraries or a third party framework etc etc so today today we have we have

almost no Library usage in our software because we we made everything from

scratch even the even the the the the asynchronous event management thing we

did it ourself we we did not pick something on the market and implement it

and and so on because when we try these kinds of library or platforms we took a

lot of time just to to implement it we do not have the control on this and at the end when you look at it you have

something very very very big and at the end you are using 5% so it's it's a big

machine and you are using so so few corre of a machine and uh so we we we

decided to develop everything from scratch almost so it's yeah that's what

we have also found because we've been using your product now we're going to Market it uh as continuous Alf because

we trust your engine it is robust and we have never seen a time when your application has crashed I'm not joking

and we use it quite a bit you know we have three instances right now so we use it for where of to provide our manage

services and you know U we have we don't have 40,000 test cases but however we

use it and we also have integration our own integration also for reporting and things like that yes it works like a CH

it's never crashed it's always works and I really like your Cloud your Cloud also

now your Cloud option before we using the jnlp now we are using cloud and you

know it's really really robust and fast it's pretty fast yes and it's it's it's

very robust we we never had any problem the up time is is terrific and uh I must

say that if you look at the resources we have on each server is very low it's

very low uh it's it's some very small servers and that works perfectly so

which means that if at one point we have a customers that have a huge amount of

um C needs or Ram needs uh we just we can increase this without any problem

it's very simple because what we have today is a very small footprint on on on the images that's good though you are

optimized it right your code optimized on our Cloud we are using some Linux Linux kernel with no gooey nothing

it's and it's very very small wow that's very nice you know um I didn't know that

that's that's very nice um so Eric one more thing I have noticed is you know what also attracted me is you know

it supports various models right sdlc models like waterfall agile so can you

just throw some light there well it's it's pretty um it's what we have now is

pretty generic uh so so you can you can even if you have um whatever your system

that you're using if you're using waterful or if you are agile and when I say agile it can be scrum it can be T

down it link can be can be many many different systems anyway at the end the

way it is implemented it make it quite generic so you can you can without any

problem to adapt your model by using the preferences Etc so that it's it's used

by by any kind of team uh in that and yes yeah go ahead Sor sorry go ahe

you have reports you have reports also right to support it you know the built-in reports correct yes yes and you

can cust my report as well and the other um so we talked about a lot of different

things you know um and for me for the LIF sign customer what's important to me

and why we chose you as our partner was like I want to reiterate again is the

robustness and traceability I can go from Su the system under test all the

way through uh testing One Direction Not Just One Direction I can go by Direction

I can go the other way also this is good not only can I connect but also visually I can see it right it's very very nice

other thing is you know robust Version Control every object has versioning history and you know we have audit rails

we have U electronic signature freeze and freeze so we have evaluated your

product only and you know that's the reason we chose so it's very good for Life Sciences with adding that now you

have uh no you few years ago Incorporated the Gins model exploratory testing which is what FDA CSA is calling

for it's already there so in effect and we have built all the reporting that we need to export it to word and you know

and uh if somebody wants to take it into word and so that they they comply with their own templates header you know

templates you know so we can still use xqu to have the single source of truth but they can report it in word we have

built some of those so if you add all that together it is really a perfect package whether you're using manual

testing or automated testing right and with all the launchers you know it really fits nicely and we talked about

the robustness and scalability one thing we did not touch was using a distributed

architecture using agents can you just throw some light there where you have a distributed architecture and we can have

agents in on many servers and execute everything from One console yes uh yes

so that's one thing that I wanted to do from the very beginning of the creation of Expo uh as I said I was designing a

system uh when I was alone on on that on that that Adventure I was designing a

system so that you can interface with different um automation framework at the

same time and and so on and so on one of the thing that was very important at the

time was uh also um since I was working in some Network and telecommunication

and things like that uh to be able to um this the test was taking a lot of

resources and uh it was not possible that you run the test on your own computer so

the main idea was just to be able to say Okay I want to execute this test campaign that contain this test this

test this test executed on a staging environment that I built somewhere else

and on this staging environment I have an a network of nodes with many many

computers or maybe some VMS now we are doing a lot of Dockers

images yeah so you have some form of doctors uh running and yeah is to be

able to execute all the test on some specific agent pool on on this network

and that's what we do and that the notion of X agent so you can have a network with many many agents running on

some computers and and then when you are going to execute your test through exio

then uh they will be executed on agent and now doers and and with and with

Dockers now it's it's it's really excellent because you can get this image up only when you whenever you want and

need that and that's that's much better than what I was doing when I was young

and with some VMS you need to start the VM and so on and so this is very

efficient now from a computer point of view right yes yeah uh Eric last question for you can you please um I

would like to know your road map uh what is your road map for 2024 2025 and how will a

on this so that's interesting that you are you are speaking of AI because yes

this will be this would be one a big part of our road map in in the future I

don't want to rush on this because many many people are working on planning to do AI whenever they are not doing AI at

all correct but but for so the road map that we have today is uh quite uh quite

um large uh we have very big project in

the in the scope we have what what we want to do is in the next step will be

to implement task management uh in the way for example you can have in in Gyra

if you go to Gyra and you have some task and stuff like that so we want to have something a bit similar to this so that

we have in our in our system also more more um something related to project

management so we have test management of the task management of the projects so

the idea is that developers are going to to use more and more also ex toio to to

manage daily tasks we also want to have the document management a little bit

like if I if I pick the same example with atasan it would be more like

Confluence equivalent say uh and for AI um this will come a little bit later but

for AI we we want so we have many many ideas with AI um the main the main one I guess would

be to be able to to so you you design your requirements let's say and from the

requirements AI would help or would do directly all the description of all the

test to cover these requirements so that would be of course this is it is quite

this is a mid project it can difficult we need to check out the relability of

that kind of things because we need to check that of course but uh but that would be a very big

project to do also AI could be used in many many different space in in next for

example when you when you design requirement or test or test case or bug

even a bug could you could use the AI to to help you to ensure the the

completeness of what you are writing Al correctness of what you are writing maybe it can help by saying oh you are

you are creating a bug here but your bug is not well uh defined because you are

missing this and this and this if you do not do that the guy who are going to pick that bug and will be assigned that

bug will not understand what it is like like a copilot right like a like a

copilot on on that it will be copilot for for the generation of items it would

be more like a generative uh like like like an agent right yeah yes uh and so

we have we have a list of 15 different things that we could use AI in to to

improve the the experience of the tester or the developer or business in their

day today tasks Eric U that is very exciting I

cannot wait to see all that so uh we are out of time you know I really really

appreciate your time it was I enjoy talking to your whatx call and how it applies to life sciences and once again

all the best and thank you thank you very much NES it's I said that at the very beginning but it's always a

pleasure to speak with you again it's uh it was nice very very nice and thanks for the invite it was very nice to speak

you yep take care bye thank you byebye byebye thank you for watching continuous

TV if you have any further questions please shoot an email at info@ continuous validation.

Transcript

welcome to continuous TV where we transform Life Sciences with Dialogue on AI and

GXP today's episode is focused onm application life cycle Management in GXP

your host is nagesh Nama CEO at xlm the guest is Eric galdo CEO of xqu this is

nagesh Nama I the host of continuous TV uh this is episode number four I am very

glad to have a special guest today uh Eric Aldo is a CEO of exqu based in

France and he has probably spent most of his life building and testing software