Rajeev: Certain. I feel today none of those conversations may be full with out speaking about AI and gen AI. We began this early exploratory part early into the sport, particularly on this a part of the world. However for us, the bottom line is approaching this based mostly on the client’s ache factors and enterprise wants after which we work backward to determine what kind of AI is greatest appropriate or related to us. In Cathay, at the moment, we deal with three important sorts of AI. One is in fact conversational AI. Basically, it’s a type of an inner and exterior chatbot. Our chatbot, we name it Vera, serves clients straight and might deal with about 50% of the inquiries efficiently. And nearly two weeks again, we upgraded the LLM with a brand new mannequin, the chatbot with a brand new mannequin, which is ready to be extra environment friendly and far more responsive by way of the human work. In order that’s one a part of the AI that we closely invested on.Â
Second is RPA, or robotic course of automation, particularly what you are seeing is throughout the pandemic and post-Covid period, there’s restricted assets obtainable, particularly in Hong Kong and throughout our provide chain. So RPA or the robotic processes helps to automate mundane repetitive duties, which does not solely fill the useful resource hole, but it surely additionally straight enhances the worker expertise. And to date in Cathay, we’ve got a few hundred bots in manufacturing serving varied enterprise items, serving roughly 30,000 hours yearly of human exercise. So that is the second half.Â
The third one is round ML and it is the gen AI. So like our digital workforce or the info science workforce has developed about 70-plus ML fashions in Cathay that turned the group knowledge into insights or actionable gadgets. These fashions assist us to make a greater resolution. For instance, what meals to be loaded into the plane and particular routes, by way of what amount and what sort of product provides we promote to clients, and together with the fare loading and the pricing of our passenger in addition to a cargo bay area. There’s quite a lot of exploration that’s being performed on this area as effectively. And a few examples I may relate is if you happen to ever occur to return to Hong Kong, subsequent time on the airport, you might hear the general public announcement system and that’s additionally AI-powered not too long ago. Prior to now, our workers used to manually make these bulletins and now it has been moved away and has been moved into AI-powered voice expertise in order that we may very well be constant in our announcement.Â
Megan: Oh, incredible. I am going to should pay attention for it subsequent time I am at Hong Kong airport. And you’ve got talked about this subject a few occasions within the dialog. Look, after we’re speaking about cloud modernization, cybersecurity generally is a roadblock to agility, I suppose, if it is not managed successfully. So may you additionally inform us in a little bit extra element how Cathay Pacific has built-in safety into its digital transformation journey, significantly with the adoption of growth safety operations practices that you have talked about?Â
Rajeev: Yeah, that is an attention-grabbing one. I take care of cybersecurity in addition to the infrastructure providers. With each of those essential capabilities round my hand, I have to be aware of each elements, proper? Sure, it is an attention-grabbing one and it has modified over the time period, and I absolutely perceive why cybersecurity practices must be inflexible as a result of there’s quite a lot of compliance and it’s a extremely regulated operate, but when one thing goes improper, as a CISO we’re held accountable for these faults. I can perceive why the workforce is so inflexible of their practices. And I additionally perceive from a enterprise perspective it may very well be perceived as a highway blocker to agility.Â
One of many key elements that we’ve got performed in Cathay is we’ve got been following DevOps for fairly quite a lot of years, and not too long ago, I feel within the final two years, we began implementing DevSecOps into our STLC [software testing life cycle]. And what it basically means is relatively than the core cybersecurity workforce being accountable for most of the safety testing and people types of elements, we need to shift left a few of these capabilities into the builders in order that the individuals who develop the code now are held accountable for the testing and the standard of the output. And so they’re additionally enabled by way of the cybersecurity course of. Proper?Â
After all, after we began off this journey, there was an enormous resistance on the safety workforce itself as a result of they do not actually belief the builders making an attempt to do the testing or the testing outputs. However over a time period with the introduction of assorted instruments and automation that’s put in place, that is now getting right into a matured stage whereby it’s now enabling the upfront groups to handle all of the elements of safety, like menace modeling, code scanning, and the vulnerability testing. However on the finish, the safety groups can be nonetheless validating and act as a type of a gatekeeper, however in a really mild and inbuilt processes. And this fashion we will be certain that our cloud functions are safe by design and by default they’ll ship them sooner and extra reliably to our clients. And on this complete course of, proper?Â
Prior to now, safety has been all the time perceived as an accountability of the cybersecurity workforce. And by enabling the builders of the safety elements, now you will have a greater possession within the group with regards to cybersecurity and it’s constructing a greater cybersecurity tradition inside the group. And that, to me, is a key as a result of from a safety facet, we all the time say that persons are your first line of protection and sometimes they’re additionally the final line of protection. I am glad that by these processes we’re capable of enhance that maturity within the group.Â