Valliamai Senthilnathan

Electrolux

Chatbots and AI: Valliamai Senthilnathan and the Future of Testing

Gek Yeo: Hello, everyone. Welcome to Eurostar Podcast Studio. This is Gek. I’m the lead manager for QA with AICPA. Today, I very lucky to have our guest Vali to talk about testing and what testing will be like in the five years down the road. So would you like to give our audience a quick introduction about yourself?

Valliammai Senthilnathan: Hi everyone, I’m Valli. I have 17 plus years of experience in testing. Currently, I work as an engineering manager for a QA and release team in Electrolux. So I manage a team of 60 plus resources.

Gek Yeo: Wow, that is a huge team resources. So I guess in with handling such a huge team, right? So, how do you encourage, new tester into your team then?

Valliammai Senthilnathan: I think for testing more than the competency as a tester, the soft skills are very important because that makes a difference. I always look for a person who is very collaborative. And who has good communication skill and also has the tendency to learn.

So that is what I look for when I hire a new person in my team. This is a core, skill set that I am looking for other than the technical or testing competency.

Gek Yeo: And what about a tester that has been in your team for a long time, how do you encourage them, growth in quality testing?

Valliammai Senthilnathan: So in testing as a carrier, there are like, either you can get into technical depth being an automation tester, Or you can get into a business or domain depth so you can be a good, functional tester or you can also the other thing is on the strategy as a tester as a test process or test governance if you are mastering those things.

So that is a bit of a different career. You need to choose in which part you want to grow. So whichever make is your passion or what is interesting for you. So that can decide how you want to.

Gek Yeo: Yeah, this remind me of one of the, key speaker, one of his, points talking about testing. If you don’t enjoy testing, something must be wrong.

Yeah. So I think that’s the key point about testing that we need to enjoy it and we need to find fun in it. So, what do you think about this year, Eurostar conference ?

Valliammai Senthilnathan: Yeah, I love Eurostar conference. This is my second time being here. I really like the talks. It’s a lot of source of information that we get access to.

And it’s a lot of tools, like from the expo as well. We get to learn about new tools. How do we implement in our landscape? How do we optimize? So it’s, a lot of opportunities at different places when it comes to testing. You have like from the strategy perspective, from the tooling perspective.

So they equip you from different kind of needs, from a testing perspective. So I think it’s a wonderful place to be. And, it’s a lot of, communication, like collaboration, networking, connections that you make. Uh, it’s a big community. Difference. I really enjoy being here last year. I was part of Eurostar conference, so I get to meet a lot of people and it gave me a new energy.

Gek Yeo: Yeah. That’s how we met each other.

Valliammai Senthilnathan: Yeah, exactly. And also I really want to be a speaker one day. I’m looking forward for that, but I get to participate in one of the virtual Eurostar huddle, where we did a presentation on how to test our AI chatbot.

Gek Yeo: Would you like to give the audience a bit insight about that one then?

Valliammai Senthilnathan: Yeah, we have an AI chatbot and voicebot in our landscape. So testing this voicebot is different from testing the other things. Because when you test, these kind of, AI products, there is something called NLP testing, which is very key. NLP is nothing but To find out what is the understanding of this AI?

How is this AI trained? Is the training good enough that it understands very well? Because a same question can be asked in multiple different ways. Yes, but the chatbot should understand and respond you properly. So that’s a key for a chatbot. So it is something different that we do. For this AI related thing, which is, and the other thing we also do is emotion context testing because we are in the service and repair industry of Electrolux, we have this AI chatbot.

So when a consumer is already frustrated, if something is not working, they should not get more frustrated . So the bot needs to always be polite. However, the customer is.

Gek Yeo: Yes, because after when you need to go to a help desk and help out, you must be in, in some sort of very frustration situation. Yeah.

Valliammai Senthilnathan: And people don’t want to talk to chatbots. They just directly want to talk to human agents. True, true human touch.

Gek Yeo: Isn’t it?

Valliammai Senthilnathan:Yeah. It’s quite important.

Valliammai Senthilnathan: There is another side towards it because the agents, when we, when there were no chatbots, when the agents were answering the calls, it was a lot of amount of calls.

It was difficult for the agent to respond. And in the same time for the consumer, it is too much of waiting time, which is frustrating from their perspective as well. So we want to help the consumer quickly. And we also want the agent to enjoy his job. True. So this is a win win. So we introduced this chatbot.

Currently in our use case, the purpose of chatbot is like mainly three things. One is like it will direct the call to the right agent. So you get the necessary information. It needs to always get repetitive information from the consumers. What is their product, their PNC model. So it collects those repetitive information.

So the agent doesn’t need to ask those information. So the call handling time comes down. That is where we use the chatbot today. And there are some problems like if water is leaking from a dishwasher, there is a troubleshooting article where you can refer, you can fix yourself. You don’t need to wait over a day for an agent to answer that, how to fix it.

So you can fix it quickly. It’s mainly to help the consumer. Maybe like we are also constantly improving our chatbot, the experience and many things that we wanted to do.

Gek Yeo: So when you first learned that you have to test this AI chatbot, right? How do you convince your team to take on to this challenge of testing it?

Valliammai Senthilnathan: For me, it was completely new. A couple of years when I joined there, you have an AI chatbot, you have to test it. But the good thing is we had an engineering manager on the AI side. She had a PhD from AI and she had a lot of experience working with this kind of things. So she was already helping us.

How do we test? What are the different things that we need to focus on? So it gave us a lot of insights

Gek Yeo: and confidence.

Valliammai Senthilnathan: Confidence. And the main thing is community. I think you always go back to the testing community. There is a lot of source of information that you can refer to and benefit out of it. So I would definitely vouch for this testing community where you get a lot of information which will help you.

Gek Yeo: And, since you mentioned, your teams now get into AI testing. What kind of, Tools or, learning. Do you advise somebody who are new to AI testing, to get them a rolling because they may not have the privilege like you have a, in house, expertise of AI.

Valliammai Senthilnathan: Definitely. I think they, in the huddle community of Eurostar, we have done a presentation. What are the important things that they need to consider for AI testing? So that is a good source of information for anybody starting new. To look at what they have to start looking into and what they have to focus.

That is a great source of information. And other than that, you have Pluralsight and lot of education material that you can go through and learn yourself.

Gek Yeo: Okay. So forth. Look out for her talk. And. Where do you see us in five years down the road in terms of testing?

I mean, you see the last two years, the Eurostar conference, the content and the talks and tutorial. Yeah.

Valliammai Senthilnathan: Yeah. I think the metaverse of testing.

So maybe like the AI or everything can support us to do our job more efficiently and more quickly. But I don’t think it is going to replace testers or testing.

So there is some niche still we need testers, testing team to focus on to deliver and everything. But the metaverse where I see, for example, we have the e commerce business where we sell our consumer electronics in our website. So how do I see it changing in the five years is maybe they implement augmented AI reality.

So where you can spin around the product, touch and feel. How do we test these kind of new use cases when the technology advancement happens? That is where I am thinking will be a metaverse of testing.

Gek Yeo: Yeah. Metaverse of testing and managing all those testing environments we have to come up with and the production environment before we push it live.

Valliammai Senthilnathan: And, today also we have The AI chatbots are based on our own language models, but they’re trying to introduce a generic AI, gen AI. So it’s a challenge for us to, how do we test a gen AI? Because for a tester, it is always, they have to look for an expected outcome. In a gen AI, there is no expected outcome.

Every time the outcome keeps on changing. So how do you test it? It’s a, it’s going to be a challenge. When you implement more gen AI stuff. Maybe we will focus on more non functional aspects rather than functional aspects.

Gek Yeo: That’s quite true. And it’s also quite a challenge for folks that being so used to, years and years of, conventional testing you have to introduce them to new ways of testing.

It can be like you say, you need the confidence, you need an insight, you know, overcome the fear and able to ramp up very quickly and do very quickly. Yeah. How long did it take for you to test for your AI chatbot then?

Valliammai Senthilnathan: It goes quickly. So we have a couple of ways that we test. Because every time when an AI, when you chat with an AI, when you ask a question, when it responds, it’s ideally an API.

So we can test it in seconds.

Gek Yeo: Good.

Valliammai Senthilnathan: So that’s how we test it. But from the end to end test perspective, we have another tool, which we use for testing as well. To see it’s like IVR testing we do, where we have the voice bots, and being Electrolux is in 120 different markets, the bot needs to talk in the local language.

True. Yeah, yeah. The tester cannot master all the languages in the world. So we take help from this tool, which help us in this language translations. So it makes us our job easier.

Gek Yeo: Yeah, I guess all your team members must be very proud when you went live with all the testing, the effort you guys put in.

Valliammai Senthilnathan: Yeah, they are, they know like some of the words in the new languages, at least the welcome messages.

Yes, true.

Gek Yeo: So, in this case, what would you like to, um, last thing that you would like to say to our audience.

Valliammai Senthilnathan: Yeah. I feel the people who are successful in the testing world should have the passion for testing.

If they don’t have that passion, then they don’t stay in that field. I really see people who are very passionate on testing would stay in that field and really. Make a difference because when you see outside for a new person to start a career in I. T. Everybody choose testing because it’s easier to start a carrier as a tester.

But to sustain that carrier and to grow in that carrier, you really need a passion. Otherwise, it’s impossible to grow.

Gek Yeo: True. Because quality testing. We are the one that always suffered the most stress. Yeah, we are the gatekeeper and, we need to be able to say exactly. We

Valliammai Senthilnathan: just want to prevent the fire.

We don’t want to do firefighting, but unfortunately today many of the teams, it is not prevention of fire. It’s more like firefighting.

Gek Yeo: Yeah. Yeah. Fixing the rail while the railway train is running. Yeah. I do agree. Yeah.

Valliammai Senthilnathan: Thank you so much for this opportunity. It was really nice being part of your podcast.

Gek Yeo: Thank you very much. And I’m so glad I get to interview you. You know, you are like one of the first few friends I met in Eurostar conference last year. All right. Thank you, everyone. Thank you. Bye. Thank you.