[Host]

[00015.76 - 00021.56] Hello and welcome to the enterprise tech tech forecast. I am your host, Saumitra Kalikar.

[00021.56 - 00027.92] The retail industry has already gone through transformation through e-commerce, mobile

[00027.92 - 00035.52] apps, marketplaces and personalization. But the next wave will be even more significant because

[00035.52 - 00042.96] agentic systems are now increasingly becoming capable of doing product research, comparing

[00043.04 - 00049.52] products, interpreting customer preferences and recommendations, and even assisting customers

[00049.52 - 00055.40] in completing their buying journeys and tracking orders. And that raises an important question

[00055.40 - 00062.32] as to are we simply improving the convenience of digital commerce or are we creating an entirely

[00062.32 - 00068.92] new commercial model? To explore this fascinating topic, I am joined today by Charisma Glassman.

[00068.92 - 00078.16] Charisma is currently the VP and Global Head of AI Advisory for Retail and CPG at Genpact.

[00078.16 - 00085.64] And she brings multiple recognitions, including the many I need to read those through. But

[00085.64 - 00097.44] she has been named as top 50 American Asians within business, top 25 AI consultants, top

[00097.44 - 00105.52] retail consultant in 2026. She has been a TEDx speaker and she is also an industry advisor

[00105.52 - 00111.76] for Retail at Cornell Tech. Charisma, welcome to the podcast. Great to have you here.

[Guest]

[00111.76 - 00117.74] Thank you so much so much for being here from across the world. I really appreciate the

[00117.74 - 00122.48] opportunity of talking to your audience and bringing my experience in agenda commerce.

[Host]

[00123.48 - 00129.28] Charisma, you indeed have a very impressive resume. So maybe we can start there. What do

[00129.28 - 00134.32] you mind providing a brief overview about your professional journey and how you land it into

[00134.32 - 00140.32] the world of AI? 

[Guest]

Absolutely. Thank you for all the good things that you said. I will not

[00140.32 - 00148.08] repeat them, but I'll simplify it. My journey really started in what I say is helping organizations

[00148.16 - 00154.88] rethink their top line and finding cost efficiencies. I think that's how most modern businesses create

[00154.88 - 00162.00] value today. I started in my career in traditional advertising when I do digital advertising and

[00162.00 - 00169.52] then what we call agenetic commerce today. What 23 plus years across continents have had different

[00169.52 - 00176.88] roles, personas across different industries, I've worked in retail, of course, CPG, luxury,

[00176.88 - 00183.92] hospitality and travel have also worked in financial services, if you will. And across all the gamut,

[00183.92 - 00190.72] I find the same thing that customers are looking to be served better and companies are looking to

[00190.72 - 00198.64] reduce cost to serve. 

[Host]

Right, right. Okay, so let's start with a broader discussion on the broader

[00198.72 - 00207.04] industry shift Charisma. As you said, agentic commerce is a widely used term now in the industry,

[00207.04 - 00213.20] right? And it has multiple interpretations because at one end, you can think of agentic

[00213.20 - 00219.92] commerce as a as a shopping assistant which can provide recommendations to customers to the

[00219.92 - 00226.32] other end. It can be an entirely autonomous agentic system which can make decisions

[00226.32 - 00233.28] and complete transactions, right? You have been in this industry. How do you see agentic commerce

[00234.16 - 00241.20] emerging as the from a technology concept to a broader industry shift in retail industry now?

[Guest]

[00242.48 - 00246.56] Absolutely. So I'm going to just simplify it. I think they use these buzzwords and every year

[00246.56 - 00253.60] we come up with new ones. Agentic commerce is nothing but an autonomous way of how customers

[00253.60 - 00259.92] shop in the new agentic world. That's it. It's very simple. And then the whole story,

[00259.92 - 00266.88] we will take about like how it combines loyalty, one view of the customer, more if you say CDP,

[00266.88 - 00274.32] which is customer data profile, and then matching all the way up to PDPs, product display pages,

[00274.32 - 00278.96] and of course, AEOs, and I'm giving you all these abbreviations that are floating around

[00278.96 - 00286.32] in how agents are shopping today. And not just customer could be an agentic bot or a human,

[00286.32 - 00292.00] but at the end of the day, agentic commerce is how we how we serve our customers in the modern

[00292.00 - 00300.00] day of agentic commerce. 

[Host]

Yeah, yeah. And I came across a report from McKinsey charisma. We said that

[00300.56 - 00307.12] agentic systems can mediate somewhere between three to five trillion US dollars

[00308.64 - 00317.84] across the global commercial consumer commerce by 2030. I wanted to get your thoughts on that as to

[00317.84 - 00323.20] because you must be interacting with multiple leaders in retail industry, how they are seeing

[00323.20 - 00329.76] these as a potential commercial opportunity. 

[Guest]

Absolutely. And then multiple researchers

[00329.76 - 00335.04] jumped back to its own. We did it recently with HFS, which is what we call the enterprise debt

[00335.04 - 00341.68] research, in which we say there are people process technology and data are the four areas where you

[00341.68 - 00347.36] need to concentrate to create value. So coming back to your point, yes, agentic commerce is a huge

[00347.36 - 00353.28] opportunity because in the modern way, in the modern world, the agents, this is the first,

[00353.28 - 00357.76] we said last top black Friday was the first black Friday with agents were doing shopping,

[00357.76 - 00362.96] and this was the first holiday season that was really put brands and retailers to test

[00362.96 - 00369.36] of how of how their bots will actually shop. So it could be like, hey, I've created an agent

[00369.36 - 00375.04] and it will give me is my product display page or really buddy for the way and always

[00375.04 - 00380.80] really written the way a bot can shop for now today. For example, bots don't shop in the

[00380.80 - 00385.84] post while way of how metadata was written. It doesn't it doesn't have keyboard, it used to have

[00385.84 - 00393.68] keyboards in the past. But now the way people search on a chat board or in ChatGPT or any other

[00393.68 - 00400.24] element you use is like, they look for a wife. It's like, can you show me a dress that can help me

[00400.24 - 00407.44] look very good or at a gala or can or a man will shop like, can you show me the options for

[00408.00 - 00416.24] black try, black tie event. So the way PDPs are coded today are very different than how shoppers,

[00416.24 - 00423.12] bots are actually searching for products. So let's see what this opportunity really comes down to,

[00423.12 - 00427.92] but there's a huge opportunity because it's just not humans and the bots will be shopping.

[00427.92 - 00434.80] So think of like an autonomous cart, think of like your dinner planning and all that is completely

[00434.80 - 00440.16] automated and delivered at your door, at your doorstep in minutes.

[Host]

[00441.28 - 00448.80] Yeah, yeah. That's a good point. And because again, you are talking to men leaders in the industry,

[00450.56 - 00458.64] where do you see the organizations, the retail companies are thinking of implementing their

[00458.64 - 00466.88] agentic use cases first? Is it more in the product selection, product discovery

[00466.88 - 00474.48] journeys? Is it in the buying journeys or is it something back and like fulfillment journeys or

[00475.36 - 00479.36] customer service journeys? Do you see any pattern or priorities emerging there?

[Guest]

[00480.72 - 00485.84] Well, the industry is still struggling from, hey, how can we get cost efficiencies from AI?

[00485.92 - 00492.72] I think where we are today, right? But to your point, is there a connected spine out there?

[00492.72 - 00497.68] So some, some companies, whether it's retailers or brands have started to really think about how we

[00497.68 - 00505.28] can connect this entire what is a unified commerce profile from loyalty to planning to one data profile

[00505.28 - 00513.12] to intent and discovery. How am I stitching that and having one view end to end? So entropic

[00513.12 - 00518.48] just released two agents in which is very autonomous way of shopping. They actually packaged it and

[00518.48 - 00525.36] just made it very openly available for public for anybody. In fact, on my website, chrismaglossment.com.

[00526.64 - 00532.80] And what I say is pseudo shop that I have created. So they're two agents, merchant agent,

[00532.80 - 00537.76] down a shopping agent. So shopping agent is front end customer facing in which it helps you

[00538.40 - 00543.52] assist in your shopping journey or customer journey. And at the back end, they have a merchant agent

[00543.52 - 00550.32] where it's giving in a merchant an autonomous way of giving the entire 360 view of what merchandise

[00550.32 - 00556.48] is actually available. So in my view, it is not a siloed way of thinking through ecosystems or

[00556.48 - 00563.60] journeys. If you will, more you can connect those dots and those data and those journeys together

[00563.60 - 00567.52] would be the way you will succeed in 2027 and beyond.

[Host]

[00568.24 - 00574.56] Yeah, yeah. And I want to unpack this more in more details later with you, Kariswah. What are

[00574.56 - 00579.84] the technology foundations company organizations should be thinking of? And also, what is the

[00579.84 - 00586.56] implication of that from the retail operating model perspective as well? But before we go there,

[00586.56 - 00594.88] I want to get your thoughts on implications of the customer experience from brand perspective.

[00594.88 - 00603.28] Because historically, the retail companies have invested into websites, personalization through

[00603.28 - 00610.24] mobile apps and websites, as you say, loyalty programs, search rankings and so on and so forth.

[00610.24 - 00617.28] Right? And those have been the ways customers engage with the brands.

[00617.92 - 00626.00] Now, with agentic AI, your agents do not necessarily need to rely on websites,

[00626.00 - 00634.08] do not need to rely on mobile apps. They need to have more access to the data at the end of the day.

[00634.96 - 00641.92] How do you see brands that think of their customer experience, which is no longer just,

[00641.92 - 00647.36] they don't need to just consider a personal person-based customer experience. But they now need to think

[00647.36 - 00650.24] of how agents are going to interact with their brands.

[Guest]

[00650.24 - 00659.66] It's a soft line between customer experience and retail media networks or

[00659.66 - 00666.62] agenda commerce, if you will. It is a soft line. But it also depends on what kind of customer

[00666.62 - 00670.86] you're targeting and what geo. So they became very different for me. For example,

[00670.86 - 00676.54] my luxury retail clients, they want AI to be invisible. They still want their butler to be human.

[00677.58 - 00681.26] But my grocery clients, very different, what we say,

[00681.26 - 00686.38] k-economic segment. There's something called k-economy. We can talk about that later.

[00687.34 - 00692.78] And then they want everything to be self-soured. I want my Instacart. When I shop with Instacart

[00692.78 - 00698.86] at Costco or BJ's or any of these Sam's Club, with Sam's Club at Walmart, I want my cart to

[00698.86 - 00705.42] be auto-populated from the last order I had created. And even if it doesn't help me look

[00705.42 - 00710.38] through my dinner now, that I want at least the last order to be pre-populated. So it's a very

[00710.38 - 00719.74] different experience on which segment you're trying to target. For example, a travel retailer

[00719.74 - 00727.58] would behave very differently at airport versus a travel retailer outside the airport in Uber

[00727.58 - 00733.34] as a thing. So it depends on the implications of customer experience,

[00734.22 - 00738.94] very depending. And you have to target the moments, a moment when I'm in Uber,

[00739.82 - 00746.22] a much more patient watching an ad or an airline versus I'm on a computer because I want to get

[00746.22 - 00752.06] shopping done or just select what I need to do and get on the way. So you have to find the right

[00752.06 - 00759.66] moments and then target them there. 

[Host]

Yeah, yeah. And one thing I wanted to understand was

[00759.66 - 00766.14] implications of the depreciation strategy from branding for brand specific because historically

[00766.94 - 00772.94] organizations have invested into customer experience as a means for differentiation as well, right?

[00773.74 - 00782.14] With agents, I wanted to understand how, what is the implication of that? Does that

[00782.14 - 00787.82] negate or dilute the emphasis on differentiation if agents are doing the work on behalf of the

[00787.82 - 00794.78] customers?

[Guest]

Actually not, actually, it makes it slightly tougher because the window of showing

[00795.50 - 00800.54] in the other lens search is smaller. So you can, it only recommends like one, two or three products

[00801.50 - 00809.98] and just like even lesser than Google ads would paid search in our wild days. So I do not think

[00809.98 - 00814.94] that the customer experience is going to get worse. I think the number of products competing

[00814.94 - 00821.98] for space is more. So you need to be just better to be able to show up when a bot searches and they

[00822.06 - 00827.98] have different. So chat GPD bot searches very different than how they would show up in a cloud,

[00827.98 - 00832.94] how they show up in a grok. So are you really, your PDPs are really optimized across the board?

[Host]

[00833.82 - 00839.66] Yeah. And when, when probably a side question I wanted to get your views on is,

[00840.94 - 00844.62] is there a, do you, are you seeing potentially some sort of even

[00846.30 - 00851.26] competitive threat or even the threat to their existing business models for some of the brides

[00851.26 - 00860.30] and the reason I ask you is, we recently heard this example of Amazon blocking news from, from

[00861.98 - 00870.30] Meta. And is this just one offer? Do you see this becoming a precedent as well in the future where

[00871.02 - 00876.14] some of the organizations, if not all, potentially pushing back for party agents?

[Guest]

[00876.94 - 00884.86] Yes. Unboxed was just launched at Amazon and Amazon has blocked news and some other agents from

[00885.50 - 00892.06] searching their platform. I think it is an evolution. Everything is always evolving at the time.

[00892.06 - 00897.26] They will be new ecosystems, new partnerships, new ways of working that will evolve. And what I

[00897.26 - 00902.62] say is agent, me, the agent, more is begin because everybody has an agent, Instacart has their own

[00902.70 - 00910.22] agent Clementine. We have Sparky and Marti at many more at Walmart. And then Amazon has its own,

[00910.94 - 00918.94] own agents as well. So with the scale, Amazon, Walmart, Costco and all these other, even in fashion

[00918.94 - 00925.18] retail, others have. I would just say that it's an evolution. These ecosystem partnerships

[00925.34 - 00931.98] has evolved over time. 

[Host]

Yeah. I like the term you use that agent agent wall will begin.

[00934.06 - 00941.34] And at this stage, I want to get your thoughts more on the technology readiness that retail

[00941.34 - 00947.18] brands need to consider. And you mentioned a few things. You mentioned about, you mentioned about

[00947.18 - 00955.90] the CDPs and PDPs of the world. But in general, for agent moving away from moving beyond just

[00955.90 - 00963.98] agent co-pilots in certain use cases to make agent commerce at scale for organizations,

[00964.70 - 00970.30] what kind of technology foundations, whether it's in the data, whether it's platforms, integrations,

[00970.30 - 00976.62] et cetera, what kind of readiness is you would advise brands to consider?

[Guest]

[00978.06 - 00981.58] Yeah. I mean, I think something's been talking about this for a while, our organizations have

[00981.58 - 00985.26] been discussing and we need to have one single source of truth. How really it happens is your

[00985.26 - 00989.82] marketing data connected to your technology data. Is it connected to even within marketing? There's

[00989.82 - 00996.22] so many silos, retail media has its own data, loyalty has its own data, and then front-end

[00997.18 - 01001.90] POS has its own data. So like, how do you kind of store it as its own data? So how do they collect

[01001.90 - 01007.66] that data and do one data leak is number one thing. I think governance and data is a product.

[01007.66 - 01012.94] I'm not throwing buzzwords up there. It's real. You need to have at least a basic

[01012.94 - 01019.26] common way of working and have one single source of truth with a customer who's shopping. Where is

[01019.74 - 01030.94] shopping across the web store, POS, whether it's across the different channels, wherever retailers

[01030.94 - 01036.22] or brands have, expose themselves. And in many cases, we have seen the case of clean rooms,

[01037.10 - 01041.50] in which not only they're sharing data within their own ecosystem, but also with others.

[01041.50 - 01046.70] If something is shopping with another brand or retailer or multi-brand outlet,

[01046.70 - 01052.86] so can we share those data? At least having a customer behavior profile is the basic

[01052.86 - 01059.02] technology readiness. For that, governance is very important, putting Guardrails, not exposing

[01059.02 - 01065.02] PII, if you will, or any other credit card information. So like, hygiene, basics, back to

[01065.02 - 01071.98] the premium basics, I would say. And then we can talk about what is the data lineage, how often

[01071.98 - 01077.50] be fresh, who do we share with, what parts of the data profile are we sharing? And so the data

[01077.50 - 01082.38] is the basic foundation, but also in within data, there's so many other complexities with governance

[01082.38 - 01088.94] and the agent data as a product itself.

[Host]

Yeah, yeah. And some of the points that you mentioned,

[01088.94 - 01096.06] which have, as you said, there have been classic readiness or building blocks that

[01096.06 - 01101.26] organizations should have had even few years back, I would say, because some of the points you said

[01101.26 - 01106.46] about, just unified customer profile, et cetera, they were required even when we're talking about

[01106.46 - 01114.06] omnichannel experience. Yes. Right. But is there something on top of that,

[01114.06 - 01120.46] that agent commerce requires in particular? So for example, if you are launching agents

[01120.54 - 01127.34] for particular transaction, particular use cases, agents need to have better context, any unified

[01127.34 - 01131.58] context of what the customer or the process, et cetera. Right. Is that something?

[Guest]

[01131.58 - 01136.62] So I think there are multiple protocols that have come up like for them, for models and MCPs and

[01136.62 - 01142.38] UCPs, which is unified customer profile. I think you have to follow the latest protocols,

[01142.38 - 01148.62] you have to expose to different APIs. So if you're, if your data or your profiles are exposed to

[01148.62 - 01154.14] different models, you have to enable MCPs. So those are the basic things that are coming up,

[01154.14 - 01160.54] and how do you really expose your endpoints? And then that exposes a security layer, because

[01161.66 - 01168.46] as soon as you start sharing customer data outside, then you have to really control the

[01168.46 - 01174.22] guardrails. Who are you exposing that to? So what are the security risks? What are the guardrails

[01174.22 - 01180.46] even put there? So that there is, you know, our governance layer is like, what is, what is exposed?

[Host]

[01181.66 - 01187.98] Yeah, yeah. And particularly for machine readability, because the agent system is that you need to

[01187.98 - 01195.34] make your core data more machine readable, right? Are any kind of practices or standards emerging

[01195.34 - 01202.78] as to how you want to model product information or pricing information, offers, et cetera, which are

[01202.78 - 01208.30] typically, in today's world, they are modeled in databases, catalogs, and then exposed to APIs

[01209.90 - 01215.58] for machines or agents to use make use of those consistently. Do we need to model those in a

[01215.58 - 01221.42] different standards as we know? 

[Guest]

There are standards. I'll tell you, it's again back to the basics.

[01221.42 - 01227.42] The way people are coding their product display pages is not through key words or SEO anymore.

[01227.42 - 01232.62] It is through English language. So how your PDP is written, if they're written through the

[01232.62 - 01238.30] metadata, it's written through the English, basic English language, the way we speak is the, again,

[01238.30 - 01244.62] it's the brilliant basics. It's not, you're not going to solve. I mean, it is nothing very different

[01244.62 - 01252.30] because the way the agents behave are very similar to how humans talk. So if you're designing your

[01252.30 - 01256.06] product display pages according to the English language, that will be the basic thing to go back

[Host]

[01256.86 - 01265.58] to. At this stage, I want to get your thoughts on how the accountability would work and how it

[01265.58 - 01274.62] should be fixed as you, as organizations start scaling agent to commerce. Because, I mean,

[01274.62 - 01279.58] you have one use set of use cases where agents just provide recommendations, and it is still okay

[01279.58 - 01284.62] in the sense that at the end of the day, customer has a choice. You are just making a recommendation,

[01284.70 - 01290.62] or it could be an internal employee has a choice based on whatever AI is recommending there.

[01290.62 - 01297.42] But as you start going more into the autonomous decisioning and autonomous actions scenarios,

[01297.42 - 01304.70] where agents can make pricing decisions, agents can make eligibility decisions, etc., that might

[01304.70 - 01311.82] have implications on financial, legal, brand reputation, etc., for the companies, right.

[01312.78 - 01319.90] And if an agent is fulfilling any of those use cases, and there is a mistake, how do you fix

[01319.90 - 01327.26] the accountability there? Are there any thoughts emerging in terms of who owns those kind of mistakes?

[Guest]

[01328.86 - 01334.54] Yes, if ultimate accountability is with a human, every agent must be marked to a human.

[01334.54 - 01340.14] And the companies are getting sued in their future, right, left and center, if they do not take

[01340.14 - 01345.66] the accountability. So they're please read the blueprint or black print, whatever, fine print,

[01345.66 - 01352.54] if you will, ultimate responsibility of any agent action is with the firm, is with the brand,

[01352.54 - 01358.38] is with the retailer, is with you as a human being. So that's again back to the brilliant basics.

[01359.10 - 01364.30] And the way to hold this, like every agent that's developed must be marked to a human being.

[01365.98 - 01369.82] And that's what we say human in the loop and not or on the loop, if you will.

[Host]

[01370.78 - 01378.70] Yeah, yeah. And when you talk about, if you want to expand that further in terms of the overall

[01378.70 - 01385.90] operating model within retail business, because in retail, you would have multiple business units,

[01385.90 - 01393.42] right? You have merchandising, you have marketing sales, fulfillment, etc. And I would imagine typically

[01393.42 - 01399.58] those business units would work a little bit in siloing in the sales that they will have their own KPIs,

[01399.58 - 01407.42] they will have their own systems and decision cycles, etc. But when you think of agent commerce

[01407.42 - 01415.58] again, it needs to connect all these part of the businesses together, right? Do you see

[01416.22 - 01423.34] potentially a need for having a relook of the current ways of working current operating models

[01423.34 - 01432.30] within retail businesses to enable this?

[Guest]

Absolutely. Most retailers, or even other companies,

[01432.30 - 01437.42] are organising departments and research, a lot of research, and also practical, like they're

[01437.42 - 01442.94] really helping them, reaching the operating model, and firstly, connection into in customer

[01442.94 - 01452.46] journeys, right from planning, sourcing, making, buying, and to the store, to the feedback loop,

[01452.46 - 01457.26] everything else, all that, how does it connect? Can we have the end to it? So that's number one.

[01457.26 - 01463.18] Number two is really designing work around the outcomes, just not going to be siloed departments

[01463.18 - 01472.06] that HR is working separately, then marketing, then technology or operations. Those lines are

[01472.06 - 01476.78] getting blurred. In fact, I've written an article recently, and there's a lot of research even

[01476.86 - 01483.42] by McKinsey about how to design around outcomes. So giving more authority towards where decisions

[01483.42 - 01489.42] are made, it's not just authorities at the top, and it's not a pyramid anymore. So authority,

[01489.42 - 01494.38] where the decisions are being made, whether it's on the floor, whether it's in the factories,

[01494.94 - 01501.82] wherever, or in the case of CPGs, in the manufacturing side, and then enabling

[01502.38 - 01508.46] those, like, laughing organization around the outcomes and notches around human beings.

[01509.34 - 01516.94] So there'll be less full-time human work, less FTE work, more and more, and more outcome-based

[01518.14 - 01526.30] solutions. And there is very clear research and work around that we are doing with organizations

[01526.30 - 01529.26] and how work should be redesigned, what we call the future of work.

[Host]

[01529.98 - 01538.38] Yeah. And in that, do you see any particular business operations with the marketing sales,

[01538.38 - 01544.46] fulfillment, and purchasing, any particular business operations getting more impacted?

[01546.14 - 01553.02] And is there a pattern emerging that some operations should take a lead in terms of rolling out

[01553.02 - 01562.38] agent use cases versus others? 

[Guest]

Yes. Any departments or organizations in the organization

[01562.38 - 01568.22] that are less exposed to the customer directly go first. So I would say back office is always a

[01568.22 - 01575.18] finance business, finance HR, those are less exposed to the front and top line. I would say

[01575.18 - 01580.38] marketing sales are the last ones to go because you need to directly touch the customer. So there

[01580.38 - 01585.42] is more risk exposed. So the organizations always start from back office, middle office,

[01585.42 - 01590.38] and front office, and in that order, typically. Customer service also we see, even though it

[01590.38 - 01598.06] touches the customer, is getting highly automated very fast because, again, it depends on industries,

[01598.06 - 01604.22] people really don't mind, they really don't want to hold on to a telephone line for very long time,

[01604.22 - 01610.30] so they're very happy talking to a chat or an agent and getting their work done low value queries.

[01610.30 - 01615.98] But high value significance impactful queries, they still want to talk to a human being.

[Host]

[01617.02 - 01624.38] Yeah, yeah, which makes sense. You want to go start with more being conservative test the use cases

[01624.38 - 01630.38] in a which are non-disruptive in the sense, right? And then once you build the capabilities,

[01630.38 - 01634.38] build the confidence you start rolling out to the more customer facing operations,

[01634.38 - 01641.66] that's what it makes sense. Building on that, I wanted to get your thoughts on

[01643.34 - 01650.62] at the end of whatever use cases we are building, because organizations want to make sure they're

[01650.62 - 01657.98] delivering value, right? And I would imagine in retail industry, there are plenty of use cases,

[01658.94 - 01666.46] for agent agent use cases. But how organizations are not thinking,

[01666.46 - 01672.30] or what has been your advice to your clients as to how they should be measuring value

[01673.34 - 01679.58] of agent role-outs agent use cases, is it more focused on productivity,

[01680.38 - 01687.10] or is it more focused on revenue, or is it, how do you, what kind of advice you are offering to your

[01687.10 - 01694.78] clients? 

[Guest]

Well, productivity is very enticing to be attracted to, because that's easy,

[01695.82 - 01701.90] and that's, but don't get stuck there, start there, but don't get stuck there, right? And measure the

[01701.90 - 01707.10] overall impact in terms of what is moving the needle in your balance sheet, what's making an impact,

[01707.66 - 01712.62] is helping me reduce costs, or is helping me make more money. It's not an

[01713.26 - 01718.70] or situation, either, or I think people over index in productivity, because it's saving costs,

[01718.70 - 01724.86] and they can really show that impact quicker. But I would say revenue is equally important,

[01724.86 - 01731.34] and, or risk, in some cases, if it is helping you solve very regulatory issues, you should get

[01731.34 - 01738.70] prioritized. I advise clients, whatever is making you the fastest impact on your P&L

[01738.70 - 01745.66] should go first, typically, unless it's a risk-based use case, or a priority that has emerged,

[01746.38 - 01751.18] but people should not, executives should not get stuck in productivity use cases,

[01751.98 - 01758.46] and get some very enticing to go there first. 

[Host]

Yeah, and are organizations thinking of

[01759.50 - 01765.90] taking a holistic AI strategy and an execution approach across the organization,

[01765.90 - 01772.38] or using those are more emerging or constrained within respect to business operations, and 

[Guest]

they

[01772.38 - 01778.70] are thinking of these use cases within their operations. So it is kind of better over the

[01778.70 - 01785.02] last two years, I would say. It's starting with pilots, pilots are scaling in many organizations,

[01785.02 - 01789.82] a center of excellence is already there. So there is some kind of method to the madness,

[01789.90 - 01796.22] if you will, and at least, you know, folks have started hiring, organizations have started hiring

[01796.22 - 01802.54] chief, AI officers, chief agent officers, we have our own, by the way, and he's awesome.

[01803.82 - 01810.06] I would say that it's getting better, but human beings, right now we are kind of a transition

[01810.06 - 01814.62] period, right, and human beings are creatures of habits. So they keep going back to their old ways

[01814.62 - 01822.54] of working. And even though they are AI strategies, sometimes they live on very pretty powerpoint

[01822.54 - 01828.86] decks, and it has taken time to operationalize them as holistically, but it is getting better

[01828.86 - 01835.42] day by day of things, especially as soon as we start getting a chief AI officer in the

[01835.42 - 01841.82] boardroom, it will become a mandate. Right now, chief AI officer, like chief technology officer

[01841.82 - 01847.50] is off the boardroom. You know, there's always like, they treat it as a cost center, they never,

[01848.30 - 01853.02] never treat them, have a seat on the board table. I think that visibility is changing in many

[01853.02 - 01859.66] organizations, especially digital main organizations. As soon as we start exposing them into Francie,

[01860.62 - 01868.38] hosted threats will start coming across. 

[Host]

Yeah, I agree. Look, as we start wrapping up,

[01868.38 - 01877.18] Charisma, I want you to get your thoughts on some specific advices to senior leaders,

[01877.18 - 01885.42] CIO, CTOs, etc. Because if an organization is thinking of getting onto this journey,

[01885.42 - 01893.10] of implementing on the agency commerce, it's not always clear to many leaders as to what should

[01893.10 - 01899.66] be the starting point, right? Whether they should be focusing on customer facing pilots,

[01899.66 - 01904.94] whether they should be focusing on to a point in the back office use cases, or it's more about

[01904.94 - 01910.54] getting the foundations ready, like getting the data foundations, other technology foundations,

[01911.10 - 01917.74] or even whether there is a need to re-assess and re-engineer the business processes.

[01917.74 - 01924.54] Because one thing important to notice in my view, you don't want to just automate the existing

[01924.54 - 01930.86] ways of doing the processes. You might use your opportunity to re-engineer processes at the same

[01930.86 - 01938.46] time, right? So, there are many ways organizations and leaders can approach this. What is your advice

[01938.46 - 01945.02] to the leaders who are thinking of investing into this? How should they approach it?

[Guest]

[01945.90 - 01954.38] I would say stop approaching it as a tool and start approaching it as an enabler. Firstly,

[01954.38 - 01960.78] it should be a top-down mandate, CIO, of adopting any new technology, any new general purpose technology,

[01960.78 - 01966.62] whether it's AI or quantum in the future or whatever that is. It has to be top-down mandate

[01967.50 - 01973.34] led by the C-suite, in this case, as a CEO, most of the times. Unless we're talking about that,

[01974.30 - 01980.46] then the change doesn't happen. Second is you should not chase AI. What tool, what platform,

[01980.46 - 01985.66] re-engineering processes, just concentrating on the business itself. How am I going to change

[01985.66 - 01992.22] the business? What it does take to change the business from X to Y? And what tools and technologies

[01992.22 - 01998.22] in AI do I need to get there? As long as we're focusing on the brilliant basics, and I think that's

[01998.30 - 02004.94] coming up the theme of this episode, we should be good. What we start doing is getting to this

[02004.94 - 02010.78] ego of, oh, I launched this because my surprise will go up. I partnered with this platform because

[02011.50 - 02018.30] I couldn't tell my board as long as we keep doing our job, creating value for the shareholders,

[02019.02 - 02026.38] understanding what will impact the P&L. And use AI technology or any other automation as an

[02026.46 - 02038.22] Igor is something that I talk to CEOs on anti-basics.

[Host]

Do you see some leaders or CIOs making

[02038.22 - 02043.66] any mistakes with which they should be avoiding? Are they new strategic mistakes? They should be

[02043.66 - 02051.74] avoiding?

[Guest]

Many. And one of them is getting enticed by new Chinese toys, or less. And then

[02051.82 - 02058.38] it's likely to make mistakes as long as you acknowledge them. It is, you know, life in a public company

[02058.38 - 02063.74] is very quarterly driven, and it is very stock price-diver. There's so much noise in the system.

[02064.62 - 02069.74] And people are making a mistake on like, hey, the new model is launched. Let me go do one

[02069.74 - 02076.70] partnership versus another. If we start distinguishing between what is a single and the noise,

[02077.58 - 02085.82] the technology, what platform, which new model, high school sonnet, Gemini 4, Amazon ads, all these

[02085.82 - 02091.90] are, there's a lock wing on. Come back to your brilliant basics, which is your balance sheet.

[02092.54 - 02100.46] Focus there, and you will rock it. 

[Host] 

Yeah. No, that's well said. And there is always a temptation

[02100.46 - 02106.14] of getting carried away with new technologies, as you said. But at the end of the day, what matters

[02106.14 - 02113.98] is how you create value using the technology. Don't heal it in a rush. Just make sure,

[02113.98 - 02124.62] as you said, building the basics, the brilliant basics, and then test and then learn and then build

[02124.62 - 02131.50] further. So on that note, Charisma, Thank you. Thanks for your time. For me, the key takeaway

[02131.50 - 02139.26] is that agentic commerce is is definitely a promising technology. But the key, as you said,

[02139.26 - 02144.62] the key theme of this whole discussion has been around, you don't necessarily need to rush and

[02144.62 - 02152.46] get carried away. But build the basics, which whether it is in the understanding how you are

[02152.46 - 02157.82] going to measure value, what are the technology readiness unit to consider, what is the operating

[02157.82 - 02164.78] model you need to think of to make it successful before you scale it and and reap the benefits.

[02165.58 - 02170.30] So, thanks again. Thanks for your time. It was a really pleasure hosting it.

[Guest]

[02171.10 - 02174.78] Thank you so much, Rav. Thanks everybody for having us here today.

[Host]

[02174.78 - 02179.10] If you found this discussion valuable, please follow us as a step to enterprise tech talk.

[02179.74 - 02187.66] And thanks for listening. I look forward to seeing you in the next episode.