What Does Microsoft Know That You Do Not - Microsoft Director of Customer Success‍

‍

Ayman Husain leads an exceptional team of Cloud Solution Architects and Customer Engineers who create Data, Analytics and AI solutions powered by Microsoft Azure. The key focus is to solve customer challenges with data to enable them to achieve their outcome and do more with data and analytics.
Solving data warehousing, predictive analytics, low-code and no-code applications, enabling artificial intelligence, transforming how to be data first, cloud first, mobile first as an organization is where Ayman exceeds as a leader, architect, solution creator and technical leader.

Ayman serves as the Customer Success Director for Microsoft’s South Region. Ayman manages a team of Cloud Solution Architects (CSA) who lead and deliver digital transformation by leveraging Azure Cloud Solutions. Ayman also serves as a Cloud Strategy Advisor and supports several regional Microsoft customer’s digital transformation and intelligent cloud journey.

Ayman excels in his capacity as a business problem solver by leveraging innovative Azure solutions for AI, ML, IOT, Big Data and Hyper-Scale Computing. Enabling cloud adoption and reducing IT costs is how Ayman provides customer success and value.

‍

Host: 

Hello, how are you?

Ayman Husain:
I'm doing well.

Host:
This is Ayman Husain, and right now he's a director of customer success, data AI, and advanced analytics at Azure Intelligence Cloud at Microsoft. He's going to talk to us today about every single one of those things. So, I'm looking forward to hearing it, and I hope the audience is too. Thank you for being able to join us.

Ayman Husain:
Hey, it's a great pleasure.

Host:
Let's talk a little bit about your previous experience, and I think you'll give some context on sort of how you got where you are today. So, if I write it right, you worked at long Consulting before Microsoft, and you were a senior engineer at Target, which is quite interesting, and I think everyone recognizes that name. But before that, it was—alongside that, you were also working as a manager at Inside. So, if you could talk about any or all of those things and talk about how that has sort of colored your experience and brought you where you are today.

Ayman Husain:
Absolutely. So, if you think of my career journey and where it started and where it is now, one of the things I wanted to do, and this comes from a mentor and guidance that I had from someone who was looking out for how I should grow my professional life, is one of the things I did early in my technology career was just have contract work, staff augmentation type roles.

And often, I—and this is over a few decades—what I did in those early days is, you know, implement solutions, technology, hardware, software, whatever you name it. And I often was struggling with the fact that these things were not designed right or probably thought out. And I was always wondering to myself, "Who comes up with these ideas?" Some of these are not the right way to do it. Well, it turns out there is somebody out there doing these decision-making. Someone is coming up with the solution—the consultants, the strategists. So, I wanted to work upstream into that realm.

So, when I started my career, it was essential contract work, staff augmentation. And then what I did next was try to figure out where those decisions were coming from, and then a company like Inside, where it was a validated reseller at that point (and still are), they were helping large corporations get hardware and acquire software to go do initiatives that they had on their own. So, I was trying to build on top of that.

So, I joined Inside to see how that magic worked. I learned a lot of things from my decision-making on enterprise-level software acquisition and hardware acquisitions and implementation goals behind it. But I still found there was a gap of who's making these decisions. And so, what I did after that, I joined Target. Target was, at that time, one of the upcoming big-box retailers, and they still are today. But at that time, I was competing against Walmart in a different way. And I thought about how they’re trying to get a competitive advantage, and they were using software and hardware and solutions to build the competitive advancement of their brand, their way out.

They’re doing supply chain and their way of doing inventory management. All these things became relevant, so I was understanding how business outcomes are connected to revenue generation and how hardware, software, and building on top of that were essential. But when I was still not quite there, I was still not at the top decision-making spot of what these things are being done and designed.
Then I realized there's strategy consulting firms like Accenture, Deloitte, KPMG, and McKinsey doing a lot of the design behind it. So I said, "Hey, let me figure out how those organizations should do it." And so I was talking to my mentor at that time, and my mentor actually said, "Look, if you want to be a part of those large, dynastic consulting organizations, you should have done it right after you finish your college graduation and education because that's where they get recruited from. That's where they start their journey."

I was a few years behind it, so I couldn’t just jump right into it. So I figured out a way to find that organization that was mature enough but yet not looking for those early project graduates to start their workforce. And foreign consulting was just that. At that time, when I joined, it was a very small firm but doing a lot of strategic consulting and doing these business outcomes that customers were asking for, clients were asking for.

I had the opportunity to bring my experience of Target and Inside together to say, "Okay, let me go to this firm." And I had the opportunity to join them and build on top of that. And that's where I started realizing decisions were made through strategies and consultants based on business outcomes these customers and clients were asking for. So, a lot about how to create the business value journey of an organization to figure out what they need for technology and hardware. And after I did that, I realized, you know, even behind the scenes. There's still another journey left, which is this initiative. These product companies like Google's, Microsoft's, Amazon's, and IBM's—there is that last journey part of it. So, as I charted my career journey, I took that learning and then ended up at Microsoft. If you think about how I evolved here, I wanted to see how things were getting done at the trench level, at the field, but you had to kind of work its way upstream to figure out what those journeys were and how they were being created. And that's where I am today. That's why I'm at Microsoft. I took the learning from Slalom, where all the strategy consulting was happening, before that it was Inside, which is essentially an evaluated reseller adding a lot of those components of the district. And then, very early on in my career, that was the contract work, staff augmentation—essentially doing other people's plans and doing it right for that outcome. But ultimately, the conversation started with the business value. No one's going to do any kind of acquisition or implementation unless it brings direct business value to their organization. And us in offline cases, which is a profit organization at the bottom line, the revenue—that's gonna make them differentiate from their competition. Fantastic.

Host:
So, let's go back a little bit to, um, sort of your previous experience with Slalom and your previous experience with Target. And could you tell me, uh, first of all, with Target, you talked a little bit about some of the differences they were trying to make with their technology to become competitive at that time. So, can you tell me how technology enabled them to sort of enter that market space or acquire market share, especially with something like Walmart? Walmart is deeply entrenched as a provider in its space, and to be able to sort of acquire market share there, you have to have quite the advantage or quite the access, or both. So, could you talk a little bit about what you discovered there?

Ayman Husain:
Absolutely. So, it is, you know, several years ago, right? So, the dynamics of retail have changed since, but at that time, one of the things that they realized was, you know, one of your big-box retailers—your cost of profit and acquisition—it's very thin. It's not a hugely profitable organization, so they had to keep their costs down and make it so, more so, that they could maximize their reward on their investment. And the way they were doing that was using technology—essentially augmenting human tasks and capability using technology. Either it be the ERP, either it be the supply chain productivity—whatever they needed to do—they were getting ahead of that. At that time when I was working there, their e-commerce platform was not something they built. They said, "Hey, look, why build something when we can go get something that's readily available at that time in the industry?" And so, even their Target.com website was not powered by their own proprietary solutions. They set up this platform called Guest Services. Guest Services was the essential CRM for knowing who you're selling to, who your Target customers and buyers are. They used that natively, and they created the analytics on top of that. At that time, analytics was not a big powerful capability natively available like we have today in e-commerce with internet-facing solutions. So, they were building these things, and technology was the way they were differentiating.

If you think of that time, Walmart was a large big-box retailer, and they were doing volume. They were maximizing their vendor relationships by pushing them to acquire solutions, and Target decided to do a little differently. Target said, "Hey, look, if we're going to be the niche, we have to have unique products. I need to have unique products that require me to work with supply chain providers." They were small—they didn't have to have the dynastic capability to supply thousands of stores. They didn't need to go in demographic and geography capability, so they were using technology to drive that conversation. They said, "Hey, look, if this is the West Coast versus the East Coast, there are certain product subtleties that will be different, and we will work on maximizing those things." But they were also learning from the fact that Walmart had already negotiated large contracts with larger suppliers, you know, like the commercial goods suppliers—like the toothpaste or the shampoos—and so they didn't need to have the same thing recreated from the ground up. They just maximized that value proposition by saying, "Hey, let's do it better, quicker, and faster," while uniquely positioning their brand at that point.
So, when you think of a company like Target at that time, and that's a slice of time, they were trying to do more with less and using technology, software to do just that—primarily optimizing their operation cost within the same sector. That increases their margin. It's sort of what they were trying to facilitate and then doubling down on e-commerce platforms, bringing them in-house and optimizing them. Well, or optimizing them, of course, but at that time, there wasn't a competitor where people had their own e-commerce platforms.

Host:
But another question I would have also is, why don't other people do that? Why don’t other retailers leverage technology in this way? Can you talk to me a little bit about what you learned regarding leadership and because you talked a little bit about being in at a certain point and then trying to go deeper down the pipeline and see how decisions are made and why decisions are made so moving on to some of the other places you were at, what did you find there that was a little bit that really was the issue with especially with Consulting being in different places why things are done incorrectly or why things were done in a way that you didn't seem to feel would actually be you know successful in the long run?

Ayman Husain: 

Yeah, so this dynamic nature of consulting or solutions or implementation has rapidly changed. So think of everything that happened before COVID as just academic in nature. Those are lessons learned, those are not applicable. What I mean by that is what has happened in consulting and leadership all the way to now is you have to make decisions very fast and you have to make decisions based on data. You can't do it on a gut feel, can't do it on a gut check. You know, those leadership qualities of, "Hey, I'm good at this because I just feel, you know, spy sense a tingle," it's a bullcrap and anybody who's doing that kind of decision making is really not doing decision making, they're just going with the gut feel. Right? You have to use data. Luck has, you know, in my opinion, like it's just, you know, right place right time, not really a lot, you know. So if you think about the decision-making quality, it's all based on data and if you don't have data, you're making wrong decisions. Right? And so the leadership skills that I've learned is even those people that are making those strategy decisions and if they were incorrectly made, it wasn't because somebody was just not good at their job, they made a decision based on that data they had at that time in that spot and that decision led them to a place where it may have been incorrect and they couldn't pivot fast enough to make the change. So some of these decisions have to be made very fast, very quick.

So if you think of leadership skills that are applicable today, it’s being agile, being able to fast pivot on anything, including software engineering development. You know, the whole ability to write software for exploits or zero-day exploits that need to be corrected has to be done very fast, very quick. You can’t sit in a room and have a waterfall project management approach over time and days and weeks. You have to do it very fast, very quickly. If you think of the skills required, the developers and the mindset has to pivot into a technology framework that requires you to be able to compartmentalize those things and develop against it and then deploy against it in a way that is going to maximize those decision-making points capability.

If you use COVID as an example, what happened during the COVID? Did it shut down a lockdown? Supply chain took a humongous hit because the traditional supply chain was, you looked at your demand, you ordered to your suppliers, the suppliers made the stuff, put it in a boat, and shipped it across, right? That took a huge hit because everybody was trying to stay home and work from home and they were going to approach, let’s do the grocery stores and they needed supply of products right there and then. It wasn’t that everybody just woke up one morning and made the bad decisions. The organization, the world at that time was designing solutions or products on a longevity of scale that did not account for a pandemic. They did not think of it that way. They did not think supply chain would take a hit. And so when you think of all those solutions, all the platforms, and all the capabilities had a bit hard was based on the fact that we needed real-time data, real now. And how do you get real-time data if you’re a supply chain and or factory floor? You may have to use IoT sensors to make sure that you have data points and distribute it from every segment of it. You have to have digitized the capability and get rid of analog systems. You have to have that data set in a place which is readily accessible, that’s what the cloud solutions so I kick it right. You're a web-based web interface-based application, so there’s no need to develop apps that have a standardized protocol and capability behind it. So web interface capability, just like you and I are talking on a digital recording platform that’s web-based, we didn’t have to install fat clients on our systems and solutions to have this conversation for recording purposes. It was done because these capabilities of developing solutions in the traditional software engineering model was not going to scale.

And so when you think of that, that’s what it really identifies for good leadership and skills. And so you have to make decisions based on actual data, and if you don’t have the right data, you’re going to make poor decisions. And you cannot wait on data all the time. You have to make sure that if data is a choke point, build systems that will get the data to you fast and quick. AI integration strategies and responsible AI practices are essential here to ensure that the data used is accurate and reliable. In addition, the adoption of AI-powered productivity tools can assist organizations in managing and analyzing vast data sets more efficiently, allowing businesses to make real-time decisions. Enterprises must consider machine learning in enterprises to keep up with these rapid changes and ensure that their software solutions are AI-optimized software for scalability and performance.

‍

Host: 

Got it, so let's talk about customer. So customer success, while it seems like a new disciplinary career and trajectory, it's really not. The idea of customer success is in the words that suck—you have to have a customer and you have to make sure their success is highlighted. In the world of Microsoft, where I live and do the customer success role, it's identifying my customers, the solutions they have purchased from Microsoft, and making sure that it is useful for their success. And this comes in a variety of forms and solutions. It is highlighting customer success in the world of cloud solutions because that's where the technologies are all pivoted towards. So if you think of acquisition of tools and technologies that are SaaS-oriented, I can't have a customer buy things just because it looks good. It's not about bundle pricing; it's a standard enterprise agreement. They have purpose for those things that they’re acquiring. I need to connect that dot.

Ayman Husain: 

So customer success in any context is about the customer's success. And some of these customers have customers of their own. So sometimes, you have to go two layers deep. You have to find out their customer's success to figure out what solutions and success they would have, so that they would use the technologies and solutions you're trying to put in front of them. The discipline is around understanding the business value of every customer you're working with. No one's buying solutions and technology or consulting or hiring for anything but their success. Even in the non-profit world, if you're, let's say, a healthcare provider, your intent might be to save lives. That's their success criteria. They're not looking at how they're saving lives; they have to make sure that they have a criteria there. So when you create metrics and dashboards and the capability of return on those investments you're making, you have to connect the dotted line: I saved lives because I bought these technologies and solutions. If you can connect that customer success journey, now you've successfully identified what it would take for you to position that advocacy of solutions and technologies they may want or may not want.

So, when you think of the complete landscape, why would somebody move away from product A and go to product B? It's not because they had a good salesperson with great discipline or great pricing. Yeah, those make a difference. Ultimately, a decision may be made based on price because that's how sometimes people are going to get the better return or better investment model. But before that journey happens, there's purpose for their success, and somebody was taking the time to connect the dots. Like, "Hey, if you invest in my product, I can guarantee a level of success for you and your customers because this is how you're going to get there." So, that business value tie-in. So, customer success roles have two purposes: one is understanding the product and portfolio and solutions that you're trying to pedal, but also understanding your customer's domain and discipline. If you're in healthcare, if you're in education, if you're in supply chain, if you're manufacturing, if you're in energy, whatever it may be, you have to know those things so you cannot bring purpose to the solutions and services you're trying to bring forward, including consulting. Even consulting organizations now have customer success because a consulting organization will not get hired if you cannot tie the purpose of your consultation to their customers.

Ayman Husain: 

If you're a strategist, and that's what you do—business value strategy—you have to understand that customer needs business value. Maybe there are data points that show that your strategy is always off, so you need to pivot. That's how leaders get hired. That's how people are making C-suite advancements, because they’re saying, "Hey, we need this or all this kind of capability." So, whenever you think of customer success in any discipline and technology, you have to understand: it is my purpose or your customer's customer or your customer success plan for that business value they need to return on those investments they're going to make sooner or later.

Host: 

When we're talking about customer success, is there what is the range of the responsibility? So, what I mean by that is, when you have customer success, you're communicating to the customer that there's going to be an outcome here, and regardless, based on our relationship, is that outcome something that you may manage and maintain throughout the integration process? Or will you rely on a team that's external to you that maintains that?

Ayman Husain: 

So, relative to what side of that coin you're on, how do you make sure that the outcome actually hits? Because some of these outcomes may happen, you know, a year, two years, and especially with a company like Microsoft, the organizations you're dealing with have long cycles for anything in general because they have five to ten-year plans. So, how do you manage that and meet those expectations, and obviously, again, relative to how much control you have over there?

Yeah, so customer success as a discipline or a place of career has evolved and it's continuing to evolve. But yes, you have to stay connected to the entire life cycle of that journey. So, there are some pillars that have come into the bear in the context of customer success. If you went into the internet, searched what it looked like, customer success roles, they'll tell you there's some pillars of discipline, but there is a core component—there’s an essential baseline, a foundational capability. One of the most important parts of this is enablement on onboarding. Right? No one's going to implement a tool or solution unless you've got them there in the proper way, and you will not get there unless you knew the outcome exit strategy.

So as a customer success manager, and they could be several of these in an organization due to the nature of the products and solutions, like you mentioned, it takes years and decades to get there. You will have to stay the journey start once and done. Now, there are products and solutions that could be once and done, but those are consumer-oriented products. Perhaps there may be very, very solution-solvent in the way that they're going to be there. I'll use an example that's becoming a recent phrase right with the GRTBI solution. A lot of people are creating AI-powered productivity tools like headshots for their profiles or just photographs that they want to use for headshots and their business purposes, right?

There's a good chance that's a once-and-done thing, right? So the customer success role behind that paid product—many of them are free—but if you were a great product, say, if you're going to go get a digitized AI-infused headshot, you're going to buy, say, like 20 pictures. So your customer success is about, "What is the outcome you want? You want good pictures?" Right? So I'm going to sell you that license or that one-use case, and we're done with you. You probably won't come back for a while unless it was that awesome, right? So the customer success journey is going to vary. It'll be smaller for that. But then the customer success journey is a large enterprise and corporations that will take years. So you will have to know the exit strategy, how you, the outcome, will be recognized as successful.

So, if you think of C-suite roles and what he says, "I want to have an efficient supply chain," that board-level decision, or that decision is being made with the outcome that they want. Like, "Hey, in five years, I want to be profitable with my supply chain by reducing costs and whatever," right? They come up with an idea. That idea could be generated by that leader or come from a strategy consulting firm, like, you know, McKenzie or somebody like that. They come up with that. So, as a product provider in the world of Microsoft, what I'm doing in this position in the Azure cloud and the data analytics that go on it, I need to connect that dot with AI integration strategies. Machine learning in enterprises is one of the major ways that businesses are enhancing decision-making capabilities and improving the efficiency of their operations. So as we're talking about the strategies for AI in enterprise applications, how can we scale that effectively and integrate these systems without sacrificing the long-term vision?

Once we’ve established responsible AI practices, it’s crucial to adopt AI-optimized software that can adapt to the needs of each enterprise while maintaining high ethical standards. This way, businesses can feel confident in their long-term AI solutions, ensuring they're driving toward their goals of cost reduction, efficiency, and profitability.

That if you implement my data and AI strategy, my intelligent platform strategy, my Azure platforms, you will be able to have a better supply chain and decision-making capability there for you or reaching the outcome you want. So, that customer success manager role, or customer success role, will have to make sure that you have that exit strategy in mind. And it could take years to get there, so you have to understand those investments that are being made are focused on that. So yes, the journey can take years, it can be months, it could take minutes, depending on what kind of connected solution you're highlighting and positioning.

So, you alluded to and continued on the theme of AI, which is sort of the main purpose of this meeting and podcast and episode. So, I'd love you to expand on that a little bit. So, just to give you a framework, we've talked about how important getting an understanding of what's going on is essential for decision-making. Essential decision-making essentially has been left in the dust as more and more information is introduced into the market and more and more opportunities to get an understanding of customer behaviors in the market. So, pivots are happening faster and faster.

Host: 

So, one question from that is: how is this analytics, or but there's also, how is AI helping with that? 

Ayman Husain: 

So, and then as AI continues to evolve, and it seems to almost exponentially accelerate, and also explanatory accelerate to the degree to which it can be integrated with, in a way as its interface with reality, the surface area of that isn't getting higher and higher, with all these sorts of graphs going in that direction. How are customer success people, and specifically you, dealing with that? However, customer expectations are changing. How are they affecting the way that you're dealing with them and then a big part of the decision is, uh, how do you find out more deeply what these different sectors need? As it's quite tough to do that very quickly, how are you managing all that in your room from the perspective of AI and solutions that are built on it?

What we call the artificial intelligence—it's not new, it's been around for a while. What is becoming more relevant is the use case capability for what that solution space was doing. So, the hype cycle that we are in right now is the generative AI hype cycle. Right, and it's not hype cycle; this solution is capability on it—large language models. When you think of the use case, it is a cycle, but it's not new. These large language models have been around for a while in different capabilities.

So, when you're thinking of how AI is influencing decision-making, it's augmenting humanity. The idea of a person doing more of what they need to do with decision points and solutions that have augmented them to be faster at those decision-making points. So, those AI models are nothing but trained models of human activities and construct, and that's the way you want to start looking at it.

So, anything that in any discipline, any industry, doesn't have any tie-in of any kind. If you are a medical professional, a doctor, and you want to make some caregiving decisions, you can do it one of two ways: you can go to school for decades and become a professional expert because you have acquired the knowledge and put it in your brain. So, you can look at these different capabilities and say, "Oh, this is what's wrong with you; it's the element here is a pharmaceutical medicine that will fix it." Or, you can leverage the data points that are actually digitally available. All those academic textbooks and medical journals are available, and the internet, to some fashion, if they're not, they’re getting digitized very fast right now.

Now, with machine learning in enterprises, you have a large language model that is trained against that set of data. So now you can use the same capability of decision-making by prompting an encyclopedia of data that has been categorized using efficient data storage systems and databases and connected data points. AI integration strategies are key here, with AI saying, "Hey, you're asking the same thing a human would be asked if you were trying to do medical science or some kind of healthcare caregiving. I'm just doing it faster." The way I’m doing it faster is using a large language model that's trying to understand your prompt, that's trained on a lot of GPUs and chipsets that are going to return the value of what you're asking in a very efficient, quick manner.

So, the augmentation there is of a human purpose and discipline. You're not magically creating jobs and roles for people that didn’t exist. Like, if you are using AI-powered productivity tools to replace decision-making and solutions, you're essentially just replacing a human that previously was training that model. You can’t just artificially create a genre or existence because it has to be trained now.

So, leveraging AI-optimized software and following responsible AI practices ensures that this process is efficient and aligned with ethical guidelines.

‍

Yeah, there's a lot of people they're worried that it's going to take their job away like coders and software developers are pretty concerned that using ChatGPT type capability you can augment uh software development absolutely so the reason we add so many software developers because you needed that human brain to create software in large amounts because you couldn't do that now yes ChatGPT now may be able to train give you the code base be able to translate code and do it faster what what did you do you just replaced humans with the technology and skills required to do that because you have a baseline trained index or capability that's going into the the decision making process so anything AI is not new AI is just a faster capable to make those decisions on the data if you have bad data you'll make bad decisions so guess what you have to do in the world of AI if you want to make good decisions make sure your data points are there and the way those language models training capabilities done or capability to make those things get done right a software developer let's say wants to uh I move from c-sharp to Java coding right and they say I'm going to use ChatGPT to train myself to go do that very good awesome but if he did not have code samples of chat uh c-sharp or Java to be trained against that software developer is never going to learn the skill of how to convert code so that codex or that the engine that's behind these a lot of popular software developer platform that says that you can translate code you're going to ask chatgpt like questions to it well it's pointless if you didn't have the Baseline data point to trade it it gets it so you still have to collect the data and that data acquisition will not happen automatically it can be digitized so think of sensor data iot devices collecting all the data points and storing it in a data system is likely going to be in the cloud because that's what the volume and scale capability is available but you still have to train a guest and so these language models have to be trained over and over again that's why chat three 2.5 and judge Equity 4 and charge 55 they're using you know tokens and uh where multiple billions of parameters capability because you are getting more data don't have to train it faster and quicker and the way you train fashion quicker is having silicon based GPU chips available in large quantities so you can run these as a multi-processing capabilities to trade that so ultimately it is Data driven and the domain that it comes for is going to be based on the domain that the data has translated to uh think of wrong right if you see those pictures of lawyers and their Law Firm portraits sitting in front of a bank of bots that looks like encyclopedias on the shelf behind that well those are cases that have been documented and printed and published and so when they have a case law conversation they're looking through all the historical information there's a librarian and a law firm that knows all the cases and all the different states and jurisdictions those have been done well if you digitize all that that same lawyer can create a brief by asking the right question to the repository of data and get it done but what if those case law was not digitized well then you still have to hit the books you still have to flip the page and look through it word by word and so if you think of acquisition with data point you have to have enough data in the system in the world in the internet to be able to confidently make the right decisions or you will continue to have bad decision making so no matter how amazing that AI solution is.

Host: 

So when we talk about this, I mean, um, you made a point to talk about the different ways AI can augment people's experiences, but then, uh, the integration part is where there's a lot of complexity. Because, first of all, there are pre-existing systems or pre-existing understandings of what kind of interface they're going to have to have to be able to be functional, and with AI, the interface is—I'm going to say fundamentally different—but it is, uh, the resolution is not as high. It doesn't have to be; that's an advantage of AI, at least for now, especially specifically with language models. The resolution is human language.

There's also accuracy and security. And, you know, because the resolution is not as high and because it's almost a stats model, you're not exactly sure what will come out—whether for better or for worse, instead of the advantage, but you're not exactly sure, which brings in security concerns. These concerns are significant, especially when implementing responsible AI practices to ensure that the integration is ethical and fair. So there are those two parts: those parts are new, but they are unique.

And then, obviously, once you scale all this, it becomes a lot more complicated. Now, is it still extremely valuable to bring in? Absolutely. It's almost inevitable that you must. So, um, you know, whether you have to sell it or not is one thing, but being able to actually have a successful integration with an existing organization—it's a complex task. So how do you go about getting that even started? Well, it begins with carefully considering AI integration strategies that align with both business goals and operational workflows.

For example, if you're developing AI-powered productivity tools for your team, it's essential to identify how these tools can streamline tasks without disrupting the existing systems. A careful approach is needed to incorporate machine learning in enterprises to optimize decision-making processes while ensuring the technology complements the organization's core processes. Additionally, to make sure your systems remain efficient and scalable, you might want to focus on adopting AI-optimized software to ensure that your existing infrastructure can handle the complexity of AI without compromising performance.

‍

Ayman Husain: 

Um justifying that well justifying the cost is definitely one thing and for some people it may not be as valuable but but how do you even approach that like how do you even say okay look um and this is even before you start talking about hey if you don't have a whole data infrastructure that can feed this model information that you're comfortable is objectively valuable that this is after you've made the assumption that this is the actual in this information is the information you actually need relative to your market and can also that is is wide enough that uh and when your Market shifts you're still getting that information right yeah as you've talked about rotations and markets are happening at a faster and faster rate so so when we think about this architecture uh that you sort of drawn out over the last few minutes 

Host: 

Um how are you able to ensure and feel make a a customer uh feel like they will have some success like with all this?

Ayman Husain: 

Yeah, so it's a great question. Uh, AI is not going to help you do anything unless you understand what the choke point is. I will use Enterprise Commercial software as an example, but let's say you have a CRM system. I don't even have to talk about a brand; it can be any brand. A CRM system's purpose is, you know, customer data for whatever it is, whether you resell, retail, or energy distribution. You have a database of customer information, and you are a salesperson today. What you have to do is open up your CRM, look at the customer profile, their address, their location, their buying habits—all of those data points are in this system. And what that person is doing is using their sales prowess or sales experience, the decades of experience they have amassed, to say, "Hey, maybe this person will like to buy this," or "Maybe this person is right for...," maybe a car dealership is a great example. If you have a good CRM in a car dealership, you may know exactly when this person may want to buy another car, what kind of model they want, what they bought previously, and what they had good success with, how did they achieve success with that car they picked up.

So, you have this database environment where this information, the CRM, has stored the capability addressed. Now, let's put Microsoft's AI services on top of that. Now you say, okay, what would Microsoft's AI services bring to me in this space? Well, what AI will do for you is make that decision-making faster, right? You're using machine learning and AI models to give you those prompts, so you can say, "Hey, look at my CRM and tell me who's ready to buy a car." You're using English language to prompt your CRM to give you an answer. Well, you can use indexes, you can use search parameters, you can write SQL queries, and do exactly the same thing.

‍

Host: 

Well, if you were to do it today in the traditional model, you need somebody that knows SQL for maybe the database environment you're using. So maybe that's a technology, so I need a query to do table joins and this and this and this because the data point I want is how many people are going to buy a car and when they're ready to buy the car. So you have a lot of tasks and capabilities to get there. But the purpose of that intent and the exercises that I want to find out who's going to buy more cars and when they're going to buy it, and I have a lot of decision-making points. Oh, by the way, I need it now, I need it faster, really quick. That's where AI comes into place because you're not able to bolt on AI solutions. So, before you build on AI solutions, that CRM platform might be ready for AI plug-in. So, if that CRM platform maker did not create an AI plug-in capability solution, then you are going to be at a disadvantage. Your AI solution will need an intermediary place to go before you make the decision. So you'll take all the data out of the CRM and put it in a data lake of some sort, then bolt on AI on top of that to build the intelligence for you to have that prompt by conversation to say who's going to buy a car, when, and how.

Ayman Husain: 

So if you're a software developer and you're a software manufacturer, you're making software and platforms, you need to have that AI plug-in built-in capability designed, ready now, or work towards creating a way to do that tomorrow. Microsoft, I use the example that we use Bing as our search engine for the internet, that now has chat GPT capability built on top of it. Well, what is that? It's that the software developers that write the search engine for Bing have developed over years and decades, realized that for them to influence the GPT AI bolt-on on top of that would need X number of parameters, X number of developments. So they start working on it. And the reason they got faster and quicker at it is because over the decade of software development, they wanted an agile mindset. They went into this scrum mindset where they make small changes flashing quicker, so the software engineering group had to, over years, pivot away from waterfall approaches into an agile approach to develop software. So, when AI became more fast, and it became a real necessity, the software developers sat in a room during a decision-making conversation and said, "Hey, we need to bolt on top of this, what are the different APIs we need to think about? How do we secure it and govern it? How do you make sure it's the right amount of data? How do we create the packages of intelligence just to go crunch it against the CPU?" They created that framework very relatively quickly and then implemented it because they're software developers, and they're using agile frameworks. But you still needed to bolt on AI on top of that to make that capability.

So, CRM or internet search engines or, let's say, a healthcare provider trying to figure out healthcare patient data, will need that interface built in some fashion. So, if you're commercializing AI for the purpose of it, the first thing you need to do is how do I bring Microsoft's AI services to the decision-making that I'm making today? Say, is it human-based, is it machine-based, is it data-point-based? All of those things still require you to design the product and solution so that you can put AI on top of that. So, the journey of "Am I going to make the right investment here?" is based on, do you need to make decisions fast and quick? Weather models are a great example. Weather data is where AI really got its refinement and age, like Weather Channel or any of those people that are breaking what the decision is. Weather is a volatile creature; if you don't make the right prediction, you could have damages and outcomes to crops and, you know, health and life. People might die if you couldn't predict a tornado or a hurricane coming down someplace, right? So, if you think of that, I need to make the decision now, very fast, and very quick. So how am I going to get that? Well, all the data captured by satellites and weather stations across the world and global zip codes have to be aggregated in a place where these models can be crunched. And the models have to change. If you ever look at the Weather Channel or any weather really use it, you'll notice when they're predicting a hurricane path, they have different models. Yeah, they have the European model, they have the American model, they have the, you know, the German model because all these different models are using data points to say, "Hey, my model says this weather path of that hurricane is going to be." That's why they show us quickly lines on those maps, because they can't really predict where it's going to go.

You know, it's going to happen because the data has to change fast to show you what that journey looks like for that hurricane during that period of a weather event. So when you think of that, you're creating an interface for those meteorologists, people, to be able to take the data, put it in, and get the response fast and quick. Same thing with healthcare, same thing with inventory measurement, same thing with energy and oil and gas. You have to have a decision-making point that is accelerated because of the data you already have and data acquisition. If you don't bring them together in your software engineering platform, you won't succeed.

Incorporating Microsoft's AI services into your system can greatly enhance your ability to make quick, data-driven decisions. By leveraging the power of Microsoft's AI services, businesses can streamline operations, improve prediction models, and stay ahead of the competition in fast-paced industries like healthcare, energy, and weather forecasting.

‍

Let me use an example of what happens in digital photography. The context of digital photography is no different than traditional photography. You still have to have a camera with a lens that takes a picture that's going to be converted into pixels or a chemical-based photographic grain, and then you look at the image. But what has happened over time is digital cameras became very popular, they became better, digital software became popular, and became better. So the traditional camera-based and chemical-based photography has fallen on the wayside. But guess what? The photographer is still applicable because you still need to know what you're looking at, you still have to frame the picture, you still need to know colors, you still need to know how to correct imperfections in pictures. They’re removed digitally, but you're just using digital technology to do that.

If the color red doesn't look like red, that picture is no good, because your perception is still a human-based element. What have you just made? That digital photographer became faster at digitizing those images and editing those images faster and quicker. And that's what AI is doing. So AI is now saying, "Hey, this color of the sky doesn't look right; maybe it should be a little bit more blue or a little bit less grayish," based on the fact that a photographer, a human, actually says this is what a blue sky looks like, this is what a gray sky looks like, this is what a red sky looks like. And they've trained it, so now that AI intelligence behind Adobe or whatever that photograph digitizer experience is, it's using that model to say that. So if you can't create the model, you're just really coming from humans. That digital photographic experience will not look right. And that's why you have these contexts of hallucinations when an AI model doesn't know what they're doing. They just create makeup stuff. Sometimes they're cool, sometimes they're just off the wall.

So that decision you're making is exactly how you would want to get there. And so if you're writing software, developing code, developing something that requires you to inject AI or machine learning into it, make sure that you write software and code with the capability of that. And by the way, in the world of economics and commercialization, you have to make sure that data is private. So privacy matters, security matters. So you have to now inject the same principles that you would have done in traditional software development to not become more relevant in that space and create those points of design to be applicable faster.

If you are a software engineer working for a software company that is making money based on their software, you want AI/ML, machine learning, governance, security—all to be a part of your big picture. Because if you're not ready for it, you may miss out on the ability to pivot fast and quick. Then the competitive advantage is lost if the outcome is reached, but the customer is not happy with its value proposition. It could be because it simply didn’t give the expected result, or the customer’s customers didn’t find it as useful as your original customer would have hoped.

Host: 

So I guess my real question is, where have you seen that the most mismatched? Or what kind of person or organization does this happen to the most?

Ayman Husain: 

It happens all the time. So what happens is, that's the product feedback loop, right? You wanted an outcome, we agreed on the outcome, I provided you the outcome, but you, as a customer’s customer, did not get that. So you'll come back to me and say, "Oh, by the way, you know what? If it could do this, that would be more valuable." So now I'm going to take that to my product engineers and develop that product feature for it, right? If that's important. Or I can just say, "Well, you know, too bad. Don't buy it again."

It's the same thing that happens when you go to a restaurant. If you see a picture of a nice meal and you order it, and it doesn't satisfy you in the same way, you're not going to buy it again. You're not going to buy it again, right? But guess what? It may not be the problem with the restaurant. It might be the problem with the menu item. There are restaurants that do well—you go to, you know, Cheesecake Factory, Olive Garden, whatever, you go to those places all the time, but not all the dishes on the menu are good, right? So you might just say, "Hey, you know what? If you go to Cheesecake Factory, don’t buy this, because that sucks, but still go to Cheesecake Factory because they have 10 other things that you like."

So the same thing applies to your products, right? You will have a customer that says, "I thought it was going to achieve this, but it didn’t." So you have two points to do. You can take that decision point back to your product engineers—maybe it's a product efficiency like, "Oh, it’d be nice to add this feature," so let me go talk to developers and product engineers to see if they can create the additional feature. Because if I did that, that customer now would be satisfied because it achieved the success they wanted even though they knew it wasn't going to do that. They just... oh, look, I didn't get the reward I wanted, but if you had this part, it would have been nicer. And so that decision-making of buying again will come back from that. Now, it could also be that you don't want this. This is not what you want, you just decided and abandoned the product. What will happen is, if it's a competitive landscape, the person that has that feature and product capability will now win the day.

Leveraging Microsoft's AI services can help optimize the product feedback loop by providing insights into customer preferences, making it easier to identify areas for improvement. By incorporating Microsoft's AI services into the process, product engineers can quickly address customer feedback and integrate the desired features more efficiently, leading to higher customer satisfaction and product success.

‍

Host: 

But okay, so one of the things I'm thinking is, if it happens a lot, then a lot of customer success is how to pivot. Whether it does happen is that fair to say?

Ayman Husain: 

It is happening a lot. It happens all the time. What do you think? There are so many solutions and products out there because the people that are not pivoting, they're not happening, they're losing out. I'm thinking of Microsoft itself. There was a time when Microsoft products were the best, and then there were times when they're like, "Oh, this is the worst of the product." But then you go back up to it again. It's a cyclical journey. A large company always has cyclical journeys. There was a time when IBM was on top, IBM is not. Nvidia is on top, Nvidia is not. AMD's on top, AMD's not. All these things have a cycle, and it's customer success-oriented ultimately at that last bit. That's why every company is investing in customer controls. If you don't get that product feedback loop established faster, you may lose that journey at that point.

Host: 

Right. So, how do you deal with that? Because it seems like customer success is pivotal, and organizational understanding of where it needs to go in the future. Hence, the sort of focus on agile systems when it comes to building things and getting an understanding of that. But how do you see that reality? And how do you preserve that relationship? Because there are different points, right? You can build a whole new product, or you can talk to the customer and say, "Here's where you might need to have put it right, with the outcome still being reached." But you know, this thing at the end of these are tools, right? It is how you use them. And now that you are at step one of your understanding of the tool, you have reached an outcome, and now you have the information necessary to make the decision for step two, right? Our tool can still do that. The tool is the tool. How do—what kind of conversations do you have at that point?

Ayman Husain: 

Uh, it depends on the discipline in the industry. Not every conversation has to happen. And what I mean by that is, your outcome is based on the outcome at that point. Like, you don't wake up. Software companies don't create answers for all the software a customer needs. That's why we have variety in the mix, right? There's ERP software, CRM software, productivity software. There's not one software that does it all. Now, there are companies that are achieving or trying to achieve that, but they're still not there yet. Microsoft being one of those examples. We want to do it all, but we won't be able to do it all. We can't do it all because we have to create this product and also understand the reward of the product in the journey of that conversation. We won't have it all, right? We don't need it because we won't make money on it all.

And so when you have a customer in that same boat, that product feedback loop is important so that you can create the next version and iteration of it. And if the customer and you have a good relationship and they believe that you can achieve that, that's what the buying relation really stays like. If I say, "Look, you know, I don't think this is gonna do what you want, but give me six months, I'll create a product feedback loop, and our product engineers will drop a feature in six months," and if you believe that I can do that because I have a track record of doing that, you will stay, and you'll persevere. And you'll say, "You know what? I'm not going to change your decision just because one feature is missing. I'll wait until that product comes back." That's why your product roadmaps—almost all companies have a product roadmap. So that's where you talk about what's going to come in six months, a year, or a journey, because you're buying that stability. It's like, yes, this company can pull this off even if the product doesn't exist.

Right. I'm going to go to them, I'll tell them the feedback, we're going to do user testing, I'm going to beta release, we're going to do all that, and then we're going to provide the feedback. They'll go correct those things, and then they will have the version two of it or the pre-release version that will fix that. That's the relationship of a customer's manager understanding that, getting that feedback to the right place, and also being honest about it. I might take it back to product engineering. Product engineering will be like, "Nope, we're never going to do that. It's not our core focus. Better go tell the customer." So, I'll go back to the customer and say, "Hey, we can't do that, but I have a way to integrate a third-party product to do that." That's why we have bolt-ons, that's why we have third-party products, ecosystems across the world. Like this, people are saying, "We don't have that, but if you added this, and I'm going to have an API that works together, I'm going to support a structure that works together, I'm going to have a way to make sure that the Lego pieces work and connect together without having a detrimental impact."

In addition, Microsoft's AI services have proven to be extremely helpful in accelerating the product feedback loop, enabling better decision-making based on data insights. By integrating Microsoft's AI services, the team can optimize processes faster, making it easier to iterate on the product based on real-time feedback. Microsoft's AI services provide powerful tools for scaling and improving the product development cycle.

‍

If we do that conversation, that's what customer success is. So, make sure your outcome is aligned with the gaps you're finding, finding a solution for that. If there's a gap that is not addressable, being clear about it so the customer knows that they'll not get that outcome they want. The last journey, the point is you're not having a fabrication now. The misses of customer success are usually when there's a hype cycle or a marketing cycle that actually it makes that messaging very challenging. Like, oh, marketing says we can do that, and then customer service managers are sitting there like, "No, it doesn't." So, who are you going to blame? Are you going to blame marketing, product, or the salesperson for committing to something that's not going to be done? Well, if you have a good relationship, you can go to the customer and say, "Hey, by the way, you bought this, right? Let me tell you—you can only get A and B, you won't get C." If the customer flips out, that's a different problem, but if the customer says, "Yeah, I kind of figured A and B would be there, but C won't," you document it, you focus on the outcome they should achieve with A and B, and then you say C is a gap. But guess what? I'm going to take C, go back to product engineering, and have them start working on it. So now we can also do C soon, a bit later.

Host: Got it. Ayman, thank you so much for joining us, really appreciate it.

Ayman Husain: Absolutely, yeah. This has been great, thank you for having me.

Host: Thank you so much. 

‍