0:26 Heat. Heat. [music] Yes.
1:25 Heat. Heat.
2:25 Hello and a very warm welcome to Front Page only on the A IM network India's definitive media organization and voice on everything AI and technology. It's such a pleasure to have you all with us. But before we begin with the headlines, I of course have to share some very exciting news with you. Cipher, yes, India's biggest AI summit and expo is back in all of its glory and it's taking place in October of this year at KTPO Bangalore. So if you are very very serious about AI, that's the place you need to be. And if you're all set now, here are the headlines. Galaxy I makes history with mission Dishi, India's largest privately built satellite launched on a SpaceX rocket [music] featuring world first optosar
3:12 technology that fuses optical and radar imagery for all weather global surveillance. [music] Indian IT faces a structural $10 billion revenue hit as AIdriven deflation forces a shift from billable hours to [music] outcomebased pricing decoupling revenue growth from traditional largecale headcount addition. India moves into the elite fifth generation stealth fighter club as DRDO secures 600 acres in [music] Andhra Pradesh to begin the production phase of the indigenous advanced medium combat aircraft AMCA. Karnataka cabinet approves high-speed rail corridors linking Bengaluru to
3:57 Chennai and Hyderabad at [music] 320 kilometers per hour aiming to rewire the southern economic triangle by slashing travel [music] times to under 2 hours. US semiconductor giant applied [music] materials secures 140 acres in Bengaluru to build advanced R&D and manufacturing [music] facilities strengthening India's position in the global chip equipment supply chain and fabrication ecosystem. Tim Cook declares Apple is over the moon about India after record March revenues driven by an unprecedented surge where over half of all Indian customers are firsttime buyers. Japanese toilet maker Toto seeks uh sees
4:43 a stock surge of 18% as investors revalue its semiconductor division which now of course generates half of company profits by producing critical components for the global [music] AI stack. The Pentagon excludes Anthropic [music] from a major AI defense framework following a dispute over autonomous weapon usage, [music] highlighting a growing tension between government military goals and corporate [music] AI safety principles. Banking leaders are urged to abandon death [music] by a thousand PC's as HFS research reveals [music] that experimental AI pilots fail to scale without shifting toward governed outcomebased production [music] deployments. And joining me for this
5:29 particular discussion is our senior tech journalist [music] Bala 5 years ago, a small team in Bengaluru was testing radar systems on drones, trying to solve a problem most people didn't even notice, which was clouds, darkness, blind spots in how we see the earth. Today that same team has launched a satellite into orbit and not just any satellite mind you the world's first optosar imaging satellite this is
6:17 mission dishrishti even prime minister Narendra Modi lauded mission dishrishti saying that this largest privately built satellite in India is a testament to our youth's passion for innovation and nation building has successfully launched mission dishi aboard a SpaceX fal Falcon 9 at 190 kg. It is India's largest privately built earth observation satellite and more importantly it carries the capability the industry has been chasing for years. Most satellites force a tradeoff. Optical satellites give you clear images but fail in clouds, fog or at night. SAR satellites work in all weather but of course produce grainy hard to interpret
7:03 data. Dishi completely changes all of that. It combines both optical plus SAR in a single pass on a single satellite. Meaning one image that is both clear and always available. Galaxy is a Bengaluru based startup. Optosar is its first satellite. What does that mean? Let's say on the ground. So during a cyclone when conventional imaging satellites go dark, Optos keeps transmitting. Defense operators get persistent surveillance. Disaster response teams get information when it is well most critical. Agricultural users get crop monitoring that does not wait for clear skies. The satellite transmits usable data through rain, cloud cover, and of course at
7:48 night. Galaxy has announced plans for a constellation of 10 satellites, substantially increasing both geographic coverage and how frequently any given location gets imaged. A manufacturing facility in Bengaluru is already in the pipeline to support that very expansion. The company sites strong early demand from defense, government and commercial buyers. The business model is a data platform. Customers log in, nominate a location and the satellite delivers the imagery. No ground infrastructure is required from the buyer's side. And now of course here is the front page take. India has a credible space program. What Galaxy is building is something distinct. A commercial product with paying customers aimed at global
8:35 procurement decisions. A constellation of 10 satellites is the business case, not so much the aspiration. It means recurring revenue, defensible positioning and a repeating argument that Indianbuilt hardware belongs in the market. Optosar is the first unit of that very line. The next nine will be the distinguishing proof. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen is front page by the IM network. Like, share, subscribe and always remember think AI, think a IM. Well, before we begin our story, I of course have to share some fascinatingly exciting news with all of you. Cipher, yes, India's biggest AI summit and expo
9:22 is back and it is going to be alive and kicking in October of this year at KDPO Bangalore. So if you're very serious about AI, you know where to be. That is Cipher 2026. Now of course for the front page story at Front Page, we've been closely tracking Indian IT. There's been good news and bad news, but one thing is evident. Indian IT can do better. Because right now something structural is taking shape. AI is effectively shrinking revenues and almost no one is talking about this. The headline AIdriven deflation could wipe out up to $10 billion worth from Indian IT revenues. But this would translate as a
10:09 one-year shock. But actually it will be a gradual 2 to threeyear structural reset. And here is how it will actually take shape. India's IT sector has crossed hundred billion in revenue. Deflation estimates are between$3 to3.5% annually. That's about $3.5 billion at risk every year. So over 2 years that amount jumps to 6 to7 billion in impact. In peak scenarios $10 billion plus territory is earlier. IT companies of course charged for what effort which was more people meant more hours which consequently amounted to more billing. Now AI is doing the same work a lot faster. So clients are asking a question
10:56 which is why should we pay the same a$undred billion $100 million deal is now becoming an $80 million deal. Same outcome lower price which is of course the definition of deflation. So most people think this will come from new deals. That's not where the shock is hiding. It is actually seated in the real pressure which is from existing contracts. Renewals are getting renegotiated. Same work, better productivity, lower pricing. Infosys already highlighted this as unusual pricing behavior, which means it is not a future risk. This is already taking place and we've covered this on front page. So please make sure to watch it if you haven't already. And of course this
11:43 is where it gets quite serious. Application managed services. The run well the business the layer which is maintenance support testing. This contributes to about 22 to 45% of revenues for large IT firms. And this is exactly what AI compresses first. That's why TCS, Vipro, HCL Tech are more exposed. Emphasis slightly less due to of course a stronger digital mix. Clearly, as you can see, this is a clear business model reset from effort-based billing to outcomebased pricing. Because when machines do the work, hours stop mattering. And here's what actually most people are missing. Revenue is growing
12:31 slower but productivity is rising and of course hiring is collapsing. Sector crossed hundred billion revenue but added only 140,000 jobs. Earlier average used to be around 600,000 jobs per year. This my friends is decoupling. Revenue is no longer tied to headcount. That changes everything about Indian IT as we know it. The market has already priced it in Nifty it down 20% YTD broader market down just 4 to 5%. There is of course a second half to this particular story. AI is also creating new demand legacy modernization agentic workflows AI governance data
13:17 infrastructures. So companies like persistent systems, tech mahindra and tier 2 Indian IT companies are better positioned to pivot way faster and over time of course some of this lost revenue comes back in new forms and here of course is the front page take. So here's the reality. The $10 billion number is definitely real, but it's not destruction. It's actually a transition. For the next 6 to 8 quarters, revenues will compress, deals will shrink, pricing will definitely reset. After that, a new IT model will emerge. The question though is can Indian IT move fast enough from peopleled delivery to AI outcomes? I would love for you to
14:05 please tell us what are your thoughts in the comments below. This ladies and gentlemen is front page by the IM network. Like, share, subscribe, and always remember, think AI, think aim. 600 acres of land in Andhra Pradesh, a fifth generation stealth fighter program and a country that has until now bought rather than built. That equation is changing. Ladies and gentlemen, the defense research and development organization has secured approximately 600 acres in Andhra Pradesh for the advanced medium combat aircraft program or as it's otherwise known AMCA is of course India's indigenous fifth generation stealth fighter designed
14:50 developed and intended to be produced entirely within the country. The club is small. The United States, China and Russia are currently the only countries operating fifth generation stealth fighters. India is now building the facility to join them. The Andra site is planned to support production, testing and sustained development, not mere prototype work. The land allocation signals that the program has moved past the drawing board and into physical commitment. And Pradesh has been positioning itself as we know as a defense manufacturing corridor. A facility of this scale brings industrial activity, supply chain development and private sector participation into a state that has been actively courting the sector. The broader program though
15:37 is built around stealth technology, AIdriven avionics and advanced composite materials. Basically capabilities India has historically imported rather than produced. AMCA will eventually be measured against the F-35 and China's J20. That is the competitive frame India has chosen to enter. Timelines and execution will of course determine whether that comparison is flattering. 600 acres is a number. What it represents is a decision about what kind of defense power India intends to be. The AMCA program has been debated for years. The land is now allocated. The mandate is set and the work is
16:23 structural rather than aspirational. Whether the execution make matches the ambition is actually the only question that remains and Andhra Pradesh is where that answer begins to take shape. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen is front page by the IM network. Like, share, subscribe, and always remember, thinky I think I am. Well, before we begin our front page story, I have some amazing news to share with all of you. Cipher, India's biggest AI summit and expo is back and it is going to be alive and kicking in October of this year at KTPO Bangalore. So if
17:08 you're very very serious about AI, well this is the platform where you need to be at. And now of course for the front page story. So I have a question for all of you. What if Bengaluru to Chennai took just 2 hours and Hyderabad felt like a daily commute? Well, that future just moved one step closer to becoming a reality. Karnataka's cabinet has done something this week that deserves more attention than it is getting. It has actually approved high-speed rail corridors connecting Bengaluru to Chennai and Hyderabad. The approval is through which is of course a very different thing from a proposal and as we know Karnataka has been at the proposal stage for long
17:54 enough which makes this approval stand out. The speeds being planned sit between 250 and 320 kilometers per hour. That is Japan's Shin Ken territory. That is actually the range Europe's faster intercity services operate in. Karnataka is not just building a faster version of what already exists. It is actually building something the south of India has never had which is here is exactly what that actually means for anyone who has made either journey. Bengaluru to Chennai currently takes between well five and 6 hours by rail under this plan that becomes under two. Bengaluru to Hyderabad which runs longer becomes a sub3hour trip. Honestly, these are of
18:41 course not just marginal improvements. And once you cut travel time that aggressively, the behavior changes. People stop thinking of Chennai as a destination and actually start thinking of it as just a commute. Businesses stop having to choose which city to headquarter in. Talent starts moving across the corridor rather than concentrating in one place. The three cities do not just get closer. They start functioning as one. We all know Bengaluru, Chennai and Hyderabad together account for an enormous share of India's technology output, its manufacturing base and its engineering talent which we are all very proud of. They are of course already among the most productive urban economies in the country. Connecting them at speed does
19:27 not just help commuters actually rewires how that entire economic region starts to operate. The honest caveat is this. Highspeed rail is extraordinarily difficult to build. Land acquisition, civil engineering, signaling systems, rolling stock procurement, every stage has its own timeline and honestly its own cost pressures and challenges. India has been well mid construction on the Mumbai Ahmedabad bullet train long enough to know exactly how complex this gets. Karnataka is entering that process with its eyes open. Whether the execution holds is well a different question from whether the ambition is right. What of course the cabinet has
20:12 approved is a direction. The delivery date, the funding structure and the execution timeline are still to come. But the decision itself matters because these corridors have been discussed for years. And what happened this week is that someone in Karnataka finally said enough discussion. We of course at Front Page want to be clear about what that signals. India is going beyond building fastest trains into making a considered argument about what kind of landscape it wants the southern part of the country to become. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen is front page by the Eim Network. Like, share, subscribe, and always remember, thinky I, thinky I am.
21:00 And now, of course, it's time to, well, get on with a wonderfully inspirational story. A major US chip giant is doubling down on India, and this move could reshape the country's semiconductor future because this does not seem like just another tech investment. Karnataka has cleared around 140 acres of land in Bengaluru. This land is for applied materials, one of the world's biggest semiconductor equipment companies. The company builds the machines that actually make the chips. Now, this is of course vital. Applied materials is not a chipmaker like Intel or TSMC. It resides deeper in the value chain. It supplies critical fabrication equipment used in
21:47 advanced nodes like 5nm and 3 nanometer which means this investment targets the very foundation of chip manufacturing. An interesting factor which of course distinguishes this development is this facility is expected to support R&D engineering and advanced manufacturing capabilities not just assembly uh outsourcing which of course makes this a high value semiconductor mechanism. moving into India and this of course realigns things because India has been trying to build a semiconductor ecosystem from design to fabrication to packaging and equipment players like applied materials are critical for that very puzzle without them fabs cannot scale the most significant aspect of
22:33 this is how it actually positions India countries like the United States Taiwan and South Korea dominate this space as we know companies like ASML and lamb research control key technologies. India entering this layer signals long-term ambition and of course serious capability building. But as al as always here is something to think about. The global chip supply chain is shifting. Geopolitics, China tensions and diversification strategies are driving investments into newer regions. India is actually emerging as a strong alternative and this move fits that global realignment. Now of course for the front page take
23:18 now for the big question will this translate into a full semiconductor manufacturing ecosystem? The answer to that is unfolding step by step. But one thing is clear India is not just participating in the chip race anymore. It is actually building the infrastructure to compete and this could define the next decade of technology growth. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen is front page by the IM network. Like, share, subscribe and always remember think AI. think yeah I am and of course right now we are going we are done actually with all of our Indiacentric news on front page and now
24:05 of course it's time that we find out what's happening on the global front Before we start in terms of what's cooking on the global front, I of course have some great news to share with you that Cipher is back. Yes. In all of its glory and it's going to be taking place in October of this year at KDPO Bangalore. Yes. So if you're very very serious about AI, you know where to be. It is at cipher 2026.
24:51 And now for the front page story. Tim Cook does not often use the phrase over the moon. He is not that kind of executive. We know that. So when he said it about India on Apple's latest earnings call, it was worth paying attention to. Apple's March quarter was its strongest on record. Revenue came in at uh $111.2 billion, profit at $29.6 6 billion. iPhone revenue at $57 billion. Growth came across multiple markets, but one market came up repeatedly in Cook's remarks. And it was not China or the United States. Cook's case for India is straightforward. It is already the second largest smartphone market in the world and the third largest for PCs.
25:39 Apple's share in both still remains small. More telling, of course, is this. across iPhone, Mac, iPad, and watch, more than half of Apple's customers in India right now are firsttime buyers. That combination is genuinely uncommon for a company at Apple's stage. So, in most of Apple's major markets, growth comes from Android switchers or existing users spending more on services. India is actually adding firsttime Apple customers across multiple product categories and that too simultaneously. So for a company at Apple scale finding a market that behaves like this is quite a rarity. Three things are moving at once. A growing tier of consumers who
26:25 can now afford premium hardware. An ecosystem pull which is first iPhone then AirPods then Mac then services. and of course enterprise adoption. That is further along than most people actually realize. Freshworks has actually deployed over 5,000 MacBooks for AI development work. That is Apple entering the tool stack inside Indian companies, not just Indian living rooms. Apple now has six retail stores in India. It is expanding local relevance, building enterprise relationships and growing its services base. And it is of course doing all of this before the market reaches saturation. The sequencing for sure matters as much as the investment. The
27:11 broader picture has some pressure built into it. iPhone demand is strong but supply remains tight. Memory and chip costs are rising. Margin compression is a real possibility in the coming quarters. And in September, let's not forget Tim Ko steps down. John Turnis takes over as chief executive which means the India strategy at the precise moment it is gaining momentum moves into its next phase under new leadership and here of course is the front page take Apple's position in India is not complicated high demand low share and a customer base that is expanding rather than churning what makes it genuinely significant is the timing Apple has
27:57 established lishing itself in this market while it is still early and the first time buyer numbers suggest there is considerable distance left to run. Whether John Turners carries this with the same conviction Cook has shown is of course the first real test of his tenure. India will be watching so will the market and guess what so will Front Page. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen as you know is front page by the IM network. Like, share, subscribe and always remember, think AI, think I am. At Front Page, we have witnessed this before. A shoe company says it's pivoting to AI. Stop goes up by 700%.
28:44 Now, a Japanese toilet maker is in the spotlight. And this time, the market is taking it very, very seriously. Toto has seen its stock jump 18% in a single day. The reason AI, but before anything else, here's the part that sounds strange at first. Toto announced a plan to scale its semiconductor component business. Markets reacted instantly because Toto is already the world's second largest producer of electrostatic chucks. These are, of course, critical components used in chip manufacturing, especially NAND memory. And right now AI demand is driving that entire ecosystem. So more than half of Toto's profits are coming from its semiconductor division, not so
29:32 much from toilets, but actually from chips. And the company is now investing $190 million to expand that very capability. So this is actually a doubling down or where the revenue already is because the AI stack is chips, memory, manufacturing equipment, materials, infrastructure and Toto finds itself inside that very layer. At the same time, its core toilet business is quite under pressure. Supply chain issues, material shortages, orders actually being paused. So the ideology changes, the legacy business slows down, the AI link business scales up. Well, honestly, this isn't something which is very new anymore. We saw it with a shoe company called All Birds. We are now of
30:19 course seeing it with a toilet company. The signal is quite clear. Anything connected to AI supply chains is being revalued sometimes rationally and sometimes aggressively. And here of course is the front page take. Here's actually where the case becomes a little different. Toto already had capability, market share, profit contribution. The market just noticed it. Now, what are your thoughts as far as uh this particular development goes? Please do let us know in the comments below. This ladies and gentlemen is front page by the IM network. Like, share, subscribe, and always remember, think AI, think a IM.
31:05 Well, before we get on with this very important development, I have to share with you an equally important development which is cipher. Yes, India's biggest AI summit and expo is taking place in October of this year at KTPO Bangalore. So, if you're very very serious about AI, that's the platform you need to be at. And now, of course, for the front page story. The Pentagon has signed AI deals with seven major tech companies, but one name is missing. Yes, you guessed it right, Anthropic. And this is of course a disagreement on how AI should be used in war. The US Department of Defense has onboarded OpenAI Google Microsoft Nvidia Amazon Web Services, Space X, Reflection
31:53 AI. These tools will now be used inside classified military networks. The Pentagon's goal is build an AI first fighting force. Anthropic of course has said a thumping no specifically on two things. No use of AI for mass surveillance. No fully autonomous weapons without human oversight. The Pentagon's position usage is their decision, not the vendors. That's actually where the deal dis got dismantled. This is actually not just about one contract. This is more so about control. Who decides how AI is used? The company that builds it or for that matter the government that deploys it. The Pentagon made its choice. Unconditional access. Entropic wasn't
32:40 just excluded. It was labeled a supply chain risk. That very label is usually used for companies linked to foreign adversaries. Anthropic challenged it in court. We know this. and a US federal judge stepped in. The ruling suggested the move may have been retaliation, not so much about security. Behind the scenes, things of course are changing. The White House has reopened discussions. Draft orders are being considered. Anthropic's new tech breakthroughs, the likes of Methos and others are changing the equation. So even now its models are still being used in parts of the system. The phrase they used is quite significant. lawful operational use which includes intelligence analysis, battlefield
33:26 decision support, target identification. But the definition of lawful is actually set by the Pentagon, not by the AI companies. And here of course is the front page take. This deal is not so much about the ecosystem. It's actually a stack. Nvidia is for hardware. AWS, Microsoft is for the infrastructure. Open AAI. Google is for the model. SpaceX is data plus connectivity and now all of them are aligned under one framework except one. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen is front page by the IM network. Like, share, subscribe and always remember think AI. Think I am.
34:11 Well, we'll take a short little break right now, but when we are back, we are back with our senior tech journalist Bala.
34:57 >> [music]
36:02 [music] >> Heat. Heat. >> [music]
37:22 [music] >> Heat. Heat.
39:16 Heat. Heat.
40:05 >> [music] >> The banking sector has a confession to make. It has spent the better part of 3 years running AI pilots that unfortunately go nowhere. The research now has a name for it, death by a thousand PC's. Today we are of course
40:51 going to find out what that actually costs and what the fix looks like. Bala who's our senior tech journalist is joining me on front page and he has of course been tracking very deeply the IT earnings and the enterprise AI adoption quite closely. Bala of course is somebody who has a very keen mind on going ahead and trying to understand what truly could be the problem. Welcome to front page Bala. >> Thank you for inviting me >> as always. It's a true pleasure. I want to start with this. The HFS research Bala says BFSI firms are stuck between proof of concept and proof of promise >> with little hope. >> Yeah. >> Right. Which is kind of depressing. >> Yeah. >> And how would you want to address this new terminology that has come into the
41:37 day a light of day which is the well death by a thousand PC's. What's really going on? >> Before I would uh start uh do you know what is the inside joke about PC's right now? I'd love for you to tell us. >> It's proof of chaos or confusion. So I'm >> kind of makes sense. >> Yeah, it makes sense. >> 3 years. It's not a short period of time. >> Yeah. And I'm nobody to undermine people's hard work, blood or sweat that goes behind PC's. >> But if you take a step back and see how PC's are performing, uh this is a clear picture you get. It's so much of Kos and a lot of money is going into PC's. uh BFSI companies are investing so much of money but they're not getting uh any uh as expected returns. So having said that
42:26 uh so what what to give you an anecdote uh to understand what exactly is happening uh for example if AIM is is a is trying to use AI into their operations uh we are contacting a tech company to give a P and uh let everything uh see how it goes. >> Sure. >> So the P works here. Imagine uh AIM is there in other geographies and AIM has a bigger employees, bigger clients, more more uh bigger structure. >> Sure, >> that's where it is not scaling. >> PC's are scaling here within a small team within a very sandboxed environment, but it doesn't scale at a in a >> so essentially they're not being able to
43:12 reach out to their internal uh larger footprint. >> Footprint. Yeah. And if you see JP Morgan CEO, he has told that uh the the bank has 600 US use cases. >> Wow. >> Yeah. If you just add three PC's for each uh use cases, just imagine the scale of uh >> Oh, yes. >> Uh the entire thing. So, so this is what exactly is happening. So, PC's are working in a very closed environment, uh predictable environment, but it's not scaling. So that's the reason why >> there is this entire >> entire moniker which has come up now which is death by a thousand PCs but pala which essentially makes me wonder
43:59 that clearly they understand that yes the importance of AI right and the usage of it which means obviously that's what they've been trying to do for the last 3 years which is to implement it like you said maybe it is good uh to go as far as a local site is concerned but when they talk about different geography raphies that's where it starts to collapse. I mean since you of course speak with so many experts and people who are the insiders what is it that they're saying that that this is not being able to uh take up that kind of scale because it's working. It's not like it's not working at all. >> It's not it's working. Yeah. >> So where is that gap coming in? >> It's it's also about uh the the kind of options you know AI is changing every day like the in the previous
44:44 conversation we spoke about. So every every other day, every other week there is new development happening. So everybody wants to you know get the latest update, latest uh tech into theirh system. So >> probably the clients the BFSA companies they want the best of the AI so they are experimenting with more with AI they're being very judicious in their spending. So this is where the gap is if you see the bipro thing the company has acknowledged that there's a deal ramp up delay happening. >> Sure. >> So that is also you know uh they are very selective in their spending they want the ROI to happen. >> So this is >> selective all right I get your point but
45:30 the fact is that whatever they're burning money on when it's not working >> is a complete waste. No, >> it's it's [snorts] we can't call it base because every every function for example fraud detection uh for example complaints for example underwriting everything has different use cases >> so so everything is uh the PC's are getting infused into all these use cases it's getting utilized but is it is there a system to stitch all these together >> okay so that's where the gap is >> yeah is there a system to stitch all all these together converge everything into a scalable AI deployment. That's where uh the entire uh problem is what I think based on the interactions
46:16 >> that you've had till now. Okay, fair enough. Which brings me to this. HFS frames that the core problem as bolting uh probabilistic AI systems onto operating models which are built for deterministic work. You and I were discussing this very briefly. Everyone seems to have this uh thought process which is true as well to a great extent which is that AI is probabilistic >> not deterministic. So is that why probably this is also kind of falling apart? >> Yeah that could also be a major reason because uh for example earlier the tech was still probabilistic but it was explainable. >> Sure. You for example to put it very simple u a a loan for example you wanted
47:04 a loan earlier metrics were like do you have a credit score do you uh do you have any default >> security perspective >> do you have the paying capacity these were the three metrics and if you satisfied all these three metrics it was yes or no >> yes >> so here when when we are including AI into the picture there are several other metrics that are getting uh captured So the the system uh >> but like what Bala sorry I'm cutting you but like what what are the other matrix that you're talking about are getting captured because to be honest >> it still remains foundationally true right >> my capacity to pay back have I had a a decent score do I have a security those
47:50 things will still remain a reality >> still remain reality but the the response AI gives you every time it keeps changing it. It might give you a maybe answer to to to a query which you're raising. So that is where exactly the problem is. So uh it is also capturing uh uh metrics like uh your uh uh your uh digital footprint. So these are the metrics extra that are that the AI is capturing and the regulators want something very deterministic. So the BFSI is one of the hugely uh regulated uh industries. So this is also a major constraint. >> So when for example let's say I want to
48:36 I want to drill a little deeper on this. >> Let's say the AI throws up uh not a yes or no not a black and white it throws up exact you know maybe a question mark. Do you not feel then that you know the the experienced uh members or the exper experienced people working at these firms could actually well step in and say okay wait a minute I've received whatever it is but from here on it will be my uh intervention and I will take it further. >> Yeah even the government the regulation they want human in the loop. Correct. Even the companies are mandating human in the loop because uh BFSI is a very sensitive sector. It involves lot of underwriting and insurance if there's a
49:22 claim denial. So it could go to uh it could seek judicial intervention. So so it's a very sensitive sector. So I think that's also a reason. >> Okay. So essentially the fact that I mean a lot of the companies which are obviously trying to implement this at scale have till now establish the fact that there has to be an human in the loop. It is a very heavy heavily regulated sector. But the fact that I'm trying to ask you now is that as far as the people who are creating these which are the IT companies which are creating these PC's are they not probably falling back on their experience and conveying this to their customers saying that okay listen these are the hurdles and this is
50:08 what we need to do and what we have done till now and this is probably a system which will help you as on date and we'll keep improving on it. Could that not be a way forward? >> Yeah, they are actually doing it. If you see the quarterly results of uh IT services companies, they are they are securing they are sustaining whatever they have. They're not building new business >> but acceleration is a huge trouble. Uh if you see a couple of companies like LTR they have uh there's a slight slump in the BFSI revenue and persistent but persistent has done better. So it depends on the the readiness of your system. Agree >> your way of doing things. So but generally speaking the acceleration
50:54 is not happening. Uh again uh >> which is a matter of concern to be honest that means something is not right. >> Yeah. If you see the Vipro thing they have acknowledged that deal ramp up. >> So that shows the client's spending is so selective. So they are they want something in return. Everybody is talking about outcome based models. So so this is how IT services companies are handling. >> I have to obviously ask you this Bala because I mean you know you and I have discussed this in the past as well. M >> uh there is I feel and I don't know if you agree with me but there is a lack of clarity right >> from from the companies themselves in terms of how are they >> for example using AI or what are they doing etc as and as investors as people
51:42 who are stakeholders for example >> what kind of an anxiety does that create in them that's something which I wanted you to shed light on because again in this case also they've not come out very clear in terms of what's going on Yeah. So, uh, ever since I started covering, uh, uh, the IT sector, so there's a lot of, uh, polishing that is happening. >> Frankly, honestly, you're you're as always you you shoot from the hip. So, I know you're very >> Everybody speaks about outcome based models, but how are you pricing it? They give a glimpse of the pricing model but are you why don't you just uh hold a presentation or just put it out put it
52:28 out and tell this is what is happening this is what can be done this is what is in the pipeline >> so that gives some kind of confidence to the people the investors the stakeholders so that's what is required so this is this is exactly what I have understood uh >> and also you To add to that point of yours Bala, I personally feel that the moment you come out clean and you tell everyone okay listen yes we're facing a slightly challenging time. These are the areas of concern, but this is how we are addressing it. So, what that does is it actually >> I feel it'll help their stock actually go up because I mean people will immediately >> uh applaud that honesty. They will say, "All right, here is a company. They know what the problem is. At least they've
53:13 acknowledged it and now they're trying to answer it. >> Answer it." But that's not the case. That's that's not that's not happening. But in quarterly results all the leadership teams they are trying to address this to a large extent. >> But is there anybody who's getting into specifics and and detailing put a put a clear >> which is the missing picture that's that I have not seen >> which is the missing link. Do you think it also has to do with a lot for the fact that you know AI pricing essentially now that it's it's come into uh the equation and is making a very strong presence uh of itself. Do you feel that a lot of the people don't even know right now in terms of how to price AI because for the longest time the model was simple right more people more hours more billing.
53:59 >> So do you think that could be a hurdle as well that they're not aware? Yeah, they are aware. If you see uh the recent uh interactions the IT services companies have had, they are talking about the pricing models. They are they're definitely time and material model has not gone. It's still there uh >> because of the existing contracts because of the existing revenue models uh the people based revenue model they have they are experimenting but >> is there anybody who's telling this is what it is I don't think so >> specifically I wanted to >> but to add to that it's also difficult to you know capture >> both AI how the way AI works also how
54:47 human also work together. How do you price it? >> I think that's a transitionary phase right now. >> A phase right now and it will take some time. >> I think that's a very valid point that you made. Bala, of course, you talked about persistent. I want to share with the viewers 24.3% yearonear as far as the BFI growth is concerned >> which makes it the outlier. So I want to ask you this. What do you think that they might be doing which a lot of other people who are probably facing the heat the other companies can maybe replicate or is it not replicatable? >> It I think it can be replicated but all these I mean all these companies don't work as a as a industry as a whole. They are competitors of each other and their
55:34 their way of functioning, their AI capabilities, their team uh the the the capability of their teams completely different probably persistent has cracked something really nicely. >> Maybe they had the you know foresight of of seeing this at a at a time and they were able to kind of pivot accordingly etc. >> Exactly. Right. Which brings me to this thought which is uh emphasis has appointed Richard Miller as global head of insurance >> explicit targeting AI le transformation tied to measurable business value which is wonderful. >> Now a little bit about Miller as you probably already know Miller himself used the phrase inflection point for insurance. So is the insurance vertical inside BFSI moving faster towards AI
56:20 adoption? Uh I personally think uh based on whatever interactions I've had, I think insurance is a bit structured. >> Okay. >> It's not so dynamic like the banking sector. Like banking has data, you know, coming very in a very dynamic way. So insurance could be a sector which adopts CI fast in terms of underwriting. It's already happening. Underwriting, fraud detection, compliance. It it could be one of the fastest growing sectors. >> Fair. >> Infusing AI. >> Fair enough. Which of course brings me to this thought very quickly. I you know I remember only about I think about a month or so back we' done a very aspirational story about how the pharma
57:05 industry had incorporated one of the bigger pharma companies had incorporated AI which would help in uh in reducing the time as far as cancer drugs were concerned. Right. And as you and I both know, the FDA has been very very strict about AI adoption in any pharma companies. It's the most regulated >> of all sectors. >> But this comes as a breath of fresh air which means that they would have done something for AI adoption. >> That kind of brings me to this point which is BFSI as you mentioned was very he heavily regulated. So if farmer can do it >> Yeah. >> you think maybe they can kind of you know take a page out of their book? I I don't I'm not sure if we can compare pharma which is also regulated >> comparing but I'm saying heavily
57:51 regulated. >> Yeah. Regulated. Yes. Um but when it comes to BFSI one thing we have to keep in mind is there is regulation governments across the across uh the world they are trying to uh make AI explainable. For example, if there's any claim that is getting uh rejected or they want the the companies to explain what exactly why exactly was the claim denied. >> If you see there there was a shooter in the US who killed a uh a CEO of insurance company. So that's how sensitive it is. So they don't want to take any risk uh when it comes to a sector like >> rightly so. Yeah. To be honest, rightly so.
58:36 >> Yeah. Exactly. and uh they want accountability, they want consistency, they want a yes or no answer, they want they want to know how an AI system uh arrived at a solution or a conclusion. That's the reason why the regulations and and the governments are are insisting humans are in the loop. M >> so this this I don't know if we can you know I I don't know how exactly pharma sector is working but but when it comes to regulation bank RBI is regulating is closely watching the AI space in India >> and of course recently there was that entire thing about the fact that how our finance minister had called upon an
59:22 urgent meeting >> when it comes to the entire bit of mythos remember so yes I mean everyone is on their toes and with with things like this happening probably you think that they get a little more circumspect >> circum yeah so yeah uh BFSI it it's it's it's highly regulated and again I can't I I don't know if I can comment about pharmacy >> okay final question which is that the HFS research concludes that scale comes not from expanding experimentation but from reducing it which is fewer governed and the aspect of repeatable deployments what is your thought on this? I want to understand this. >> Um uh can you can you please repeat the question? >> If you were advising a BFSI CIO today,
1:00:09 >> what is it that you would like to say? What is that one decision they should make this quarter to move from pilot to production? Maybe you will give them the magical elixir which will solve their three-year-old problem. No pressure of course but >> okay I don't think I'm I'm an expert domain expert to tell all these things but from whatever HFS has put out or from the interaction I have had in every in every function there are several uh use cases so just take a couple of them and fund it and let it scale >> fair point >> and instead of you know doing 100 PC's in one particular function that doesn't work >> so instead of that just let one or two one or two or three scale into the
1:00:55 production level >> and then keep expanding >> and then keep expanding. So that's that's what the industry tles. >> Fair point. Well, there you have it ladies and gentlemen. As far as BFSI is concerned, well, the IT companies are holding ground as of right now. They're not accelerating. And of course they have been right now lumbering under the failed PO narrative for quite some time. Three years actually to be precise. And I feel I agree with Bala what Bala is saying. And as far as finally to sum it up in our front page take the HFS finding is not that complicated. Banks actually that keep the funding new pilots without closing old ones are not actually building AI capability. They're on the other hand building a catalog of
1:01:42 things that almost work. The IT earnings data across TCS, BRO, Infosys and LTI Mind Tree tells kind of the same story from maybe a different angle. BFSI spend is not disappearing. That's not happening. It is actually starting to concentrate. The programs with clear ownership, defined outcomes and governance are getting funded. Mind you, the rest are facing scrutiny because yes, let's face it, it is one of the most heavily regulated sectors. Bala of course has made a very valid point which I agree with, which is that let us not squander away the entire thought of a P with 5,000 different angles. Let's keep it focused. Try a couple of cases, implement it to uh its fullest capacity
1:02:29 and then start to expand. Maybe, just maybe, that might be the mantra moving forward. Please do let us know what are your thoughts in the comments below. This ladies and gentlemen is front page by the AM network. Like, share, subscribe and always remember think I think I am
