Building an A.I.-first organization

Everyone is excited about A.I. and the idea of using technology to improve business. But that excitement has also led to confusion. There are many definitions and applications of A.I., and few have been able to truly optimize their A.I. strategy. That’s where Ashwin Mittal comes in. Ashwin is the CEO of Course5, a transformative intelligence company helping companies improve their businesses using technology like A.I. and machine learning. On this episode of Up Next in Commerce, Ashwin explains where and how people have gone wrong with A.I. in the past, and the steps an Ecommerce company needs to take in order to be able to get the most out of A.I. It all starts with understanding and controlling your data, and includes retraining your employees to rely less on their guts, and more on the analysis A.I. provides. He reveals how to do exactly that on this podcast.

 

3 Takeaways:

• It’s important to build a structured data foundation from the start so you can move to a place of making decisions with data

• You must teach your employees how to interpret and analyze data and make sure they are comfortable using the data to make decisions. Individuals must be taught to lead with data and then let their gut take them the final 20% of the way toward a business decision

• Understanding the entire customer journey and where there are points of failure will help you create a strategy on optimizing your data and building an Ecommerce experience that is successful from start to finish

 

For an in-depth look at this episode, check out the full transcript below. Quotes have been edited for clarity and length.

---

Up Next in Commerce is brought to you by Salesforce Commerce Cloud. Respond quickly to changing customer needs with flexible Ecommerce connected to marketing, sales, and service. Deliver intelligent commerce experiences your customers can trust, across every channel. Together, we’re ready for what’s next in commerce. Learn more at salesforce.com/commerce

---

Transcript:

Stephanie:

Hey everyone, and welcome back to Up Next in Commerce. This is your host Stephanie Postles, and today we're joined by Ashwin Mittal, the CEO of Course5 Intelligence. Ashwin, thanks for coming on the show.

Ashwin:

Thanks for having me, Stephanie. It's a pleasure to be here.

Stephanie:

Yeah, I'm really excited to have you on today because throughout a lot of our previous interviews that we've had so far, everyone has mentioned AI in some form or fashion, and when I saw that you were coming on the show and we were going to have a deep dive conversation into AI, I was very excited.

Ashwin:

It's a deep subject to talk about, so yeah, it should be fun, hopefully.

Stephanie:

It will be. So, what brought you into the world of AI? What got you excited about that and building a company around AI and analytics and all that?

Ashwin:

So, we've always been in the business of delivering insights for sales and marketing to customers from data and information. But about five or six years back, I realized that we're in this perhaps future technology wave of transformation because of the onset of artificial intelligence technology, and from a big picture perspective, if you look back the last 30 years, 40 years, we've perhaps seen two or three waves of substantial change value creation through technology. We've seen the PC wave, and we've seen the internet wave. You can say, okay, maybe there's a separate mobile internet and social media wave, but it's part of that. Both of these have completely changed our personal and professional lives, and one was built on the other. Because we had PCs, so the internet was [inaudible 00:02:03].

Ashwin:

So, my belief is that with AI it's going to be pretty much the same. It will take 15-20 years to play out, but its impact is going to be as deep and profound as the internet. And it's going to be built on top of the fact that we all have computers, that we're all connected, all devices are connected, and for us being in the data business of driving insights from data, we're really in the eye of the storm. Because the root of AI comes from data science. So, on the one hand while this presents a challenge to our business, and possibilities of us getting disrupted, but at the same time, it presents even larger opportunities.

Ashwin:

So, at that time, I decided that we needed to really double down on the AI wave, and started to completely reorient the business and the company towards riding this wave. So, that's really how [inaudible 00:03:12].

Stephanie:

That's great. So, how would you explain Course5 Intelligence? What do you all do? What kind of clients do you help? What problems do they come to you with?

Ashwin:

So, we help our clients make the best decisions from data and information for sales marketing customer-related problems. And within that, one of our largest focus areas is digital and Ecommerce. We typically work with large corporates, Fortune 500, Fortune 1,000 companies, and in some cases, we might be working with the Ecommerce group, their digital group, with the CMO's office, sometimes with IT or some combination of all of these.

Stephanie:

Very cool, and you guys have over 1,000 people that you employ, right?

Ashwin:

Yes, we have 1,200 people.

Stephanie:

Wow, that's crazy. When did you all start, and how long did it take to get to that 1,000-person mark, and what was the change like within your company? That's a lot of people to manage.

Ashwin:

Sure, sure. Yes. So, we've been in business for, what? 16, 17 years, and we've also changed and evolved along the way. The way you run a company that is 50 people doesn't really work when you're running a company that's 200 people, and you need to run a company differently when it's 500 people, and certainly when it's 1,000. So, we've also had to learn, change, evolve along the way. How we run the company, what kind of talent we bring in. How we set the organization structure, and today we're pretty much a global business. We're present in seven, eight different countries, have customers all over the world, so we're sort of a micro multinational on our own.

Stephanie:

That's great. Is there any struggles that you face when it comes to working with your teams around the world that you had to learn along the way?

Ashwin:

Sure. Yes, yes. No, absolutely. Absolutely. So, there's one challenge, which is a more difficult challenge, which is culture. There's a second challenge, which is also important, but perhaps not that difficult, which is regulation. Culture is a really tough one. It's not as tough today as it used to be maybe seven, eight years back. Today because of just the globalization and easy access to information everywhere, people are a lot more aware of different cultures, different people, and a lot more sensitive and things like that.

Ashwin:

But in our early years when we were globalizing ourselves, we had to learn about different approaches, different cultures. We had to be very open minded, and I think we managed okay. We're still learning, it's important to be sensitive. Culture is an important issue. In terms of regulation, that's somewhat easier. You just have to follow certain processes, protocols and make sure that they're completely compliant in every job that you're offering. And especially in a business like ours where you're dealing with data and information, and there's a lot of regulation around that, around data privacy and sensitivity around that and things like that. So, we need to just ensure that we are compliant in every geography that we operate.

Stephanie:

That makes sense. So, everyone talks about AI. Like I mentioned, a lot of people mentioned in our previous interviews, and oftentimes I think people if they're using AI or they're referring to something like it's AI and it's actually not. So, how would you describe AI? And specifically, how does AI help Ecommerce, or could it help Ecommerce if people aren't already looking into it?

Ashwin:

So, first point you made is spot on, right? What is AI?

Stephanie:

Yeah.

Ashwin:

So, this really means different things to different people, and honestly, there is no one perfect watertight definition. So, AI is essentially an umbrella term that covers a bunch of different technologies, which are all meant to convey computers getting intelligence, which is not typically expected from computers, and which is more human-like. Which is the common sense definition, which we would all accept. But then what falls in that bucket, and what doesn't fall in that bucket is different for different people, and I'll give you a simple example. Take a very old technology, something like OCR, optical character recognition. Nobody would call that AI today, but when it first came out, maybe some people thought it was actually computers becoming intelligent, and it was kind of AI.

Ashwin:

So, the definition is also keeping on evolving as our expectations from technology keep changing. But I would say that amongst the different technologies that encompass AI, the core one, the most important one, at least to me, is what you call learning or machine learning. And machine learning is essentially the ability of computers to learn, and that is really a game-changing technology. And then that supports all the other technologies that are typically and the umbrella term of AI. Where a computer can run a certain process or program and get better at it every time it runs, which is what humans are good at doing.

Ashwin:

So, that for me is one of the most important technologies. And sorry, you had a second question you asked about AI and Ecommerce?

Stephanie:

Yes. Yeah, how it impacts Ecommerce right now. Do you think people are using AI efficiently or what do you think the future could look like if they do start implementing this in a better way?

Ashwin:

So, it's already here. It is being used extensively, more by some companies, less so by some. Companies like Amazon are, of course, using it in a very, very mature and sophisticated manner. But various types of companies are using it, many of our clients are using it, and its footprint is increasing. It's used in automating various tasks, in virtualization, in campaigns, even at the back end in supply chain, inventory management. We all know that chat bots are used for customer support in A.I. In Ecommerce. Physical robots are used in warehouses, which also could have AI technology. If it helps, I can give you a couple of examples-

Stephanie:

Yeah, I think exactly. That'd be great.

Ashwin:

... of a couple of our platforms-

Stephanie:

Yes, please.

Ashwin:

... which truly you would consider as AI. So, we have this one platform called Compete. Now what Compete does is it enables our customers, our clients to gather competitive intelligence in the market, and respond in quick time. So, it has technologies and a bunch of bots that go out and search all competitive websites of a brand to collect and synthesize all the information on what the competitors are doing in terms of the five Ps. So, you traditionally have four Ps, product, placement, promotion and price, and we have a fifth, people reviews.

Ashwin:

And so, this keeps collecting this information in rapid time, and on different product SKUs and different combinations, how competitors are doing, what they're placing, how they're pricing. And then this is updated in a dashboard along with analytics to help the Ecommerce players monitor competitive moves and respond quickly so they can optimize their revenue, their profits. This platform's also used extensively during these major selling seasons, like Black Friday, Cyber Monday, and brands expect to make substantial sales in a very short period of time.

Ashwin:

So, they have to be very responsive to what the competitors are doing and what's happening in the market. So, this is one example, one platform that we have. Another one, which we call Adomate, this is really taking it to the next level where we're using AI to optimize creativity?

Stephanie:

What?

Ashwin:

Creativity, exactly. So, AI is not getting creative, just to be clear, and I'll explain to you how it works.

Stephanie:

It's not going to take over the world, do all the artwork in the future, write all the books.

Ashwin:

No, no, no. Well, there's some of that talk going on, but I think that you'll always want to read the book written by the author, you'll always want to know about the author. You'll want to see the artwork, which you want to know the painter's story. While those things might happen, but I don't think that's going to... Robots can play sports, right? But you want to watch humans play a football match. You don't want to watch robots play, they might play better than the humans. So, it's the same thing with art.

Ashwin:

But in terms of this platform called Adomate, so what we're doing is we're actually optimizing the process of content creation for either campaigns or advertising. So, when companies typically do campaigns online for Ecommerce, they will create, let's say, 100 different creatives, 100 different images, and put them into a system. And then the system will do live optimization depending on who's responding to what. Change which creative is being shown to whom. But you might have missed something fundamental, right? You don't know why something is working, why something is not working. Maybe you missed something in those creatives. Maybe there's some combination of both that you didn't know about.

Ashwin:

So, what our AI system does is using computer vision it actually reads every creative. So, it actually looks at the creative and then distills it into structured data. Who are the protagonists in the creative or what's their ethnicity? What's their emotion? What the background colors? Is the brand shown? Is the logo shown? Where is it shown? If it's a video, then it will break it into frames and read each frame. And so, you've got now structured data against each creative, and then you have the data of how people responded to that. Who clicked, who bought which segment, and you can combine all of that and actually then get prescriptive.

Ashwin:

So, depending on what kind of campaign you want to run, what segment you want to target, you can actually advise, "Okay, use a dog in this or use a Latino woman or bring the brand in or use red color instead of purple or whatever."

Stephanie:

Wow, that's interesting. Are a lot of companies using that today where they're actually personalizing the image for the product based on who comes in using AI behind the scenes to come up with that in real time?

Ashwin:

So, this is an early stage platform. It doesn't change the image in real time, though that's also possible. And my AI head tells me that we should be doing that next. But it doesn't change the image right now, what it does is it prescribes. So, for the next campaign, when you want to create something, then you would enter into the system what you're trying to achieve and the system might advise you what to do. Or if you have a creative you can submit it to the system and it will give you a score of how successful it's likely to be.

Ashwin:

But even that is an early stage platform. This is one of our exciting newer products, and we have maybe three or four customers using it right now, but it's certainly something that we're betting on.

Stephanie:

Got it. So, when these customers come to you, one thing I think about is when you have a problem. Back in my old days at prior companies, the teams I worked with who were focused on machine learning always told you like, "Well, you can apply machine learning to a lot of things, you just have to know the problem and if it's worth solving with AI and machine learning." How would a company know if the problem that they're maybe encountering is something that could be tackled with AI? Or how do they start thinking about that, especially if they're a smaller company, maybe a B2C company right now, and they are feeling certain pinches in areas, but they're not really sure how to handle that? How do you tell someone to start thinking about this or how do you have a conversation with someone so they can get this on their radar?

Ashwin:

Sure, sure. No, great question. So, it has to go step by step. That's what we always tell them, and first, honestly, a data foundation needs to be in place before you should even be thinking and talking AI. In terms of technology, what data you're collecting, how your metadata is defined, what data sources you're using, what are the connections between all of those, and are you able to establish a single source of truth. You can't put the cart before the horse. You need to make sure you're collecting the right data and it's reliable data.

Ashwin:

Then the other, I would say, even more important piece than that is organization culture. So, technology is all available, and getting your people aligned to data and technology to drive their actions and decisions is very good. If you're not able to achieve that, then all else will fail. So, we tell people that even before you think about advanced analytics or you think about AI, first get your teams into the culture of making decisions through data. We've all made decisions through gut, and gut is nothing but some kind of big data swirling in our heads. How do we move from there to letting data take us 80% of the way, and then still the top 20% can come through our gut? Because there are things we know that we may not have put that data into that system. We may not have been able to capture everything which is in our heads.

Ashwin:

So, we need to first get that culture in place, otherwise the entire analytics and the agenda won't [inaudible 00:18:12]. Once we have that in place then we can start driving different types of analytics programs and business outcome led programs for higher sales, higher profits for customer. Getting better customer experience, and it's not necessary that every problem, as you said, needs AI. Depending on the size of the company, the complexity of the problem, sometimes you might just be able to use a prepackaged solution. And there are prepackaged solutions out there, which maybe can solve your problem instead of developing something custom.

Ashwin:

But yes, if your problem is complex, your data size is large, then yes, you can have substantial rewards by deploying an AI solution, which we've seen with many of our [inaudible] customers.

Stephanie:

Cool. Yeah, great answer. So, when it comes to the companies that are coming to you, there has to be a core theme that many companies are struggling with, where you keep hearing the same recurring problems or the same thing that they want help with. What is that theme or problem if you could kind of group it together, and what did it look like after they implemented some of your analytics and AI got assistance from you guys? Kind of like a case study if you have one.

Ashwin:

Sure, sure. So, yeah, it depends on the company and what stage of the data and analytics maturity they're at, and what their business objectives are. So, sometimes we're asked to help with foundational data issues of assessing data quality of living data infrastructure. We sometimes work with our customers to create access to data and to create across sources, and just provide them with reports they can consume to run their business.

Ashwin:

So, in these cases we might start with a discussion of their business, and what data and metrics they need to make decisions and on what frequency. Sometimes we might get asked to choose or implement a specific data technology. And then for customers who achieve data maturity around data and metrics, then we get asked to drive business outcomes, which in the Ecommerce world it could be conversion rate optimization, it could be upsell, cross-sell, customer churn reduction, personalization.

Ashwin:

Essentially for many clients, we work with them through this life cycle, and this typically takes years. We first build their data foundation, we provide them with key metrics and intelligence to run the business, and then start driving sophisticated analytics programs, and then start leveraging AI in a more sophisticated way. So, it's really a journey, and it keeps evolving. So, it's not something that you come in, you do and you finish.

Stephanie:

Mm-hmm (affirmative), makes sense. So, what are some of the metrics that you provide them. You said you provide metrics around the business. Is there a core set of metrics that you think are really important for every company to look at or how do you think about that?

Ashwin:

Sure. So, again, it depends on the way they view the business and their needs, and typically, this will start with a conversation between us and them of how they run their business, what are key metrics they want to look at? Maybe we might have a point of view what metrics they should look at. But on the tactical side, we might help people optimize metrics or measure metrics like basic things like nuclear revenue, margin, campaign incrementality, lift. So, lift is the extent to which a campaign drove sales over and above the regular run rate of the business.

Ashwin:

These high-level metrics could be broken down to... Within these metrics you might have some metrics like average order value, units per transaction and others like that. Conversion rate typically tends to be a metric of focus, and then these could be compared to past periods, could be segmented by channel, by device, by geo, by transaction

Stephanie:

When it comes to the metrics, have they ever led a company the wrong way where you saw someone look at a certain metric or data point and they were making decisions off that where it was actually giving them bad information? Or you had to advise them like, "You guys shouldn't be looking at this because this isn't helpful. Maybe you should be looking at this instead."

Ashwin:

No, absolutely. So, that happens all the time, and honestly, data if it's garbage in, it's garbage out. It's actually a great question because so often we've seen companies and very large sophisticated companies where different business units, geographies and departments have built their own data systems and their own infrastructure. And then in that process they've gone about defining their own data and their own way. They define a certain metric. Every metric needs to be defined what that metric essentially means, how it's calculated.

Ashwin:

And then if you have different geos, different views defining metrics in different ways and then when you put it all together and you're trying to look at it, it makes no sense. So, we've often seen that happen. It's very important for companies to really have a very clear data foundation and a data strategy, and to have a metadata layer, right? Define the data beforehand. And often sometimes we have to come in and just do that. Sometimes we'll end up just doing their house cleaning with our customers.

Stephanie:

So, if you were to start an Ecommerce company today, would you tell them getting the data aspect right from the start is a priority? Should they make sure they have their data dictionary, and they're talking about how they're actually putting their metrics and collecting the right data? How should someone think about this if they're brand new, and set themselves up for success?

Ashwin:

Absolutely. Absolutely spot on. So, what we call it is a data strategy. So, they need to have a clear data strategy in place as in what data they need to collect, how they're going to define that data, what technology they'll use to collect that data, and what business outcomes they'll drive from that data. How they'll use that data to drive certain business outcomes. And of course, that will evolve because business is dynamic, the business changes, the market changes, and what you track, how you think about things, it has already changed a lot in last few, and it will continue to change further.

Ashwin:

But it's very important to have a data strategy, and it's important to keep reviewing it and enhancing it as you go.

Stephanie:

Are there any data points that you recommend people to collect that maybe they're not already, because I'm sure a lot of platforms that especially newer brands are on probably collect some level of data, but I don't know if it's the right kind of data or what they really need to help them with that longer term data strategy. Are there any key data points where you're like, "Make sure you get this, this and this from the start to really be able to help build towards the future"?

Ashwin:

I mean, it's all these ones that I've mentioned like conversion rate. Obviously, traffic and obviously conversion rate. Different points of failure where drop-off has taken place. Campaign effectiveness, campaign effectiveness by segment. All of these definitely would recommend that people collect. One thing we've discovered is even in today's day and age, one of the biggest failure points which we've talked about for a long time in Ecommerce, but it still holds true even today, is just the checkout process. So, just the customer is willing to give the brand his or her money, but somewhere something doesn't work, something doesn't render, some option doesn't come up and there's drop-off.

Ashwin:

So, definitely collect data around the entire journey and where the drop-off and all of that happens if it does. We found that it's remarkable that even today that seems to be an area of [inaudible 00:27:17].

Stephanie:

Gotcha. So, with all this data that you get access to when you're helping them build their data strategies and all that kind of stuff, is there any surprises that you've seen when going through some of the customers data and helping them organize it and build systems around it. Anything that you saw that you weren't expecting?

Ashwin:

Yeah, so actually we covered a couple of those challenges that we've seen, but the two main sort of surprises that we've seen are the two that we just covered. One is, just like we said, the checkout process. The page takes too long to load or it doesn't render on a particular device or particular browser. And then just the entire confusion around the data asset that the company has and how it's being measured, the metadata, and also there are opportunities for data sharing with partners and with vendors. These are really under leveraged, and if it's done in a thoughtful way it can yield real dividends.

Ashwin:

So, to give you an example, we have this major CPG customer, and we were helping them with their Ecommerce business and with their Ecommerce analytics. And then they said, "For our Ecommerce business we actually have a different supply chain because we have to compete with the needs of Ecommerce customers, which is very different. We need to have quicker deliver times and whatnot." So, they asked us to help them with their supply chain analytics. So, we started doing that, and then we realized these guys were buying their raw material product from farms, from various farms. And the farms actually have a wealth of data that can be combined by our customer across various farms to give them back valuable inputs to improve their efficiency, and also to improve product quality.

Ashwin:

Perhaps there wasn't enough advantage being taken of this opportunity, so I think there are opportunities that businesses just don't realize that they're sitting on, which they're able to leverage.

Stephanie:

Yeah, that's a really good example. We had a similar example when talking to Grubhub where they would share their data back with the restaurants to help them improve. Like hey, this person.... You'd get maybe a bad review whenever someone orders the nachos versus... Or they order at 5:00 PM, like what kind of chef do you have at 5:00 PM versus 9:00 PM when you get better reviews. Again, it's a really interesting point about how companies can partner with each other to share data to help each other.

Stephanie:

Do you think there's any hesitancy around that, because I could also see companies viewing even the farmer as maybe a potential competitor? If they were, I guess, worried about that, or worried about sharing data that could somehow come back and bite them later. Have you seen that hesitancy because I do see this as the way of the future, but I just don't know if I've seen enough people doing it yet?

Ashwin:

Sure. No, you're right. You're absolutely right, and that hesitancy is there, and it's fair concern that there could be competitive issues. So, for example, so many brands sell direct and sell to marketplaces. What information does the marketplace share with the brand, and what information does the brand share with the marketplace? There is a symbiotic benefit because when the brand has its own property that provides a certain richness of information about the product. And while they may still be doing a larger share of their revenue in the marketplace. But these kinds of concerns are there, competitive concerns are there.

Ashwin:

Then there are also concerns about data privacy because data privacy is a big issue and it can be done ensuring compliance, but one has to be careful of how one shares the data. What data is shared? Is it masked? Is it personally identifiable or not? And then the other issue is what we spoke about earlier is that they may be defining data in different ways. So, different entities that are defining their data in different ways, again, if it's shared, it may not lead to the right analysis because it may actually provide different perspective than what it's meant to provide if it's defined in different ways with [inaudible 00:34:10].

Stephanie:

That makes sense. Is there any way that you think it would be best to set up a data sharing program that would also make sure that the company doesn't lose focus? Because like you said, it could be a pretty big process to make sure that you're putting the right data points in there, masking it, actually giving your supplier or whoever it might be insights. Then I could also see that turning into 50% of your day job. So, how do you advise a company to think about that if they are thinking about sharing data with a partner so that they also don't lose focus on their own product?

Ashwin:

Sure, sure. So, and some retailers are doing it today already. Amazon does it, Amazon has Amazon Marketing Services where it shares a fair amount of data with its brand partners. It has certain definitions and certain ways in which it makes it available, which is pretty standard. So, then it's up to the branch to take advantage of it and use it in the way that it makes sense for them. And then they might have the other marketplaces that may be sharing this data in a different way.

Ashwin:

So, that's where we really come in is that we know how the different formats work and the different definitions work, and we bring it together in, let's say a dashboard that makes sense for a brand to consume and process different sources.

Stephanie:

Yeah, that makes sense. So, we keep talking about data definitions and how companies, oftentimes different teams will have different definitions for the data. I have personally experienced this in some of my old roles. And oftentimes, it's because maybe a team is very entrepreneurial where they're trying to start their own project and they're trying to create their own dashboards, and you just all of a sudden have funny different organizations using a different metric for, like you said, the conversions. Have you seen any best practices for large companies to be able to create a global spot for people to go and look into that dictionary to find what this data metric, if it already exists, what it means? Have you ever seen anything like that, that actually works well?

Ashwin:

Yeah, I think it's great question, and honestly there's no real silver bullet. Different companies are using different approaches and strategies, and the entire data and analytics journey is really evolving across companies. Different companies have different organization structures to [inaudible 00:36:45]. One thing that works, which has worked for some companies is having a chief data officer. Somebody who's really part of the CEO's office and who's empowered to drive that agenda throughout the entire business.

Ashwin:

But for certain other types of companies it doesn't work because they're so fairly diversified, they have different business units that have different needs, and they want that dynamism. So, in those cases there is a compromise where every business then goes ahead and sets up its own system and approach and uses that. So, then you typically have, on the one hand you have, let's say the core operation systems like your accounting and things, which work as a single source of truth. And then every business uses what we would call then multiple versions of the truth, which sit on top of a single source of truth and then they create their own logic and versions on top.

Ashwin:

So, we've seen both approaches, and both have their pros and cons. I don't think there's a definite answer.

Stephanie:

So, when it comes to having some versions of the truth in dashboards, I always get hesitant about dashboards because people can interpret them however they want. One person might be like, "Things look great." And the other one might be like... They might expand the time horizon and be like, "Things look horrible." It just depends on who's looking at it and what they want to see or what they think they see. So, how do you... You said that you are providing data a lot of times for these brands to make decisions, business decisions off of the data. How do you, or if you do this at all, guide them on maybe like, "Here's how you should think about this decision"? Or how do you make sure that dashboards are being read correctly?

Stephanie:

And this is not just for your company, but I'm thinking a lot of companies have dashboards that could potentially be advising people the wrong way. Well, not even advising. Providing data and people are reading it the wrong way.

Ashwin:

Sure, sure. No, absolutely. So, there, again, there's one very important is the people aspect and training aspect is very important. How to use that information, what it means, what it doesn't mean. As you said, you can look at something and interpret it in five different ways, and one person can say it's great and one person can say it's terrible. So, that training is very important, and along with that what we do is that we will set a baseline for expected performance for most key metrics. And then we have certain tools where we can actually append insights into those dashboards.

Ashwin:

So, we have this platform called Discovery, where along with dashboards, the system actually generates contextual insights. So, along with a number, it will explain what that number means, and why that number has moved from one point in time to another. So, then that helps people contextualize that information, and as they see that... And they can actually double click on that, so this allows people to interact with that information as well in a natural way. You can actually chat and receive information back on chat, or you can ask a question and the system will... A basic question, it doesn't do very, very deep questions, but basic or an analytical question and the system will understand your question and the calculation can come back and answer. So, things like this-

Stephanie:

That sounds helpful.

Ashwin:

Exactly, and then just on a very basic level, as we work with the business and we understand what they consider as good, not good, average, certain metrics, we can do things like color coding, highlighting into dashboards. Basic things like that can also help to just contextualize.

Stephanie:

Very cool. So, a couple times we've mentioned needing to train employees to have that data mindset and to actually know how to think about data and organize it. Is there any training tools that you recommend or courses or things that you've seen companies have success with by having employees go through them?

Ashwin:

Yeah, courses there are lots, and there are enough courses and there are lots of great trainers out there. But what is very important is you need to have a couple of internal evangelists within the company. There's this new term actually in the industry actually, which has become very popular, citizen data scientists. You have data scientists, which is like the kind of people they have. Technically people who can go deep in the data and who can drive the statistics modeling and all that. But citizen data scientists are essentially the translators. They're the guys who sit in between the executives who are making the decisions and the data scientists at our end.

Ashwin:

In some cases, the citizen data scientist may even sit in our organization, but mostly they sit in the client's organization. These people play a very important role of driving that awareness and culture within the company. It's highly recommended that every company either converts some of their existing resources into that or otherwise they hire some people [inaudible 00:42:20].

Stephanie:

I like that chief data scientists, and I have heard that quite a bit. I think it'd be interesting to have a course depending on if you're in Ecommerce, it'd be nice if there was a certain path that employees could go down to then be well versed and know how to operate within their industry. Because it does seem like there's just so many courses, I don't even know where to begin sometimes with them. And oftentimes you don't know what you need to learn either. Seems like there needs to start being some tracks that you can go down.

Ashwin:

That's true. That's true, and because of AI technology and because of all this transformation, there are lots of new opportunities also, new roles that people can take up. And I do believe that going forward we should all plan to spend some percent of our time learning.

Stephanie:

Yeah, so when it comes to skills, do you see people headed in a direction where everyone's becoming a polymath where they're a jack of all trades with many things, or do you see people really focusing in on a specific skill like, "I am an expert in AI for retail, that's my lane that I swim in"? How do you see the future shaping up for skills?

Ashwin:

Sure. So, I think that there is always an increasing need of course right now for specialists. But there will also always be a need for good generalists because you need specialists who can be deep in a certain function or technology, especially disciplines like AI, but at the same time we need generalists who can make sense of all these different pieces and put them together. So, I think that both career tracks... Therefore, you should just be clear which one of the two you want to be. Trying to be both then you're setting yourself up for failure.

Stephanie:

So, for everyone listening you're good either way. Whatever one you are, you just have to choose. So, to zoom out a little bit into the general Ecommerce industry, what trends or patterns are you most excited about for digital commerce?

Ashwin:

So, well, now we are in the midst of a human crisis, right? So, a humanitarian crisis with the COVID pandemic, and of course, we are all very mindful of the human tragedy, the hardship, the economic hardship it's put on people all over the world. It is exciting that this crisis has presented to really accelerate digital transformation and the use of digital channels. So, we have seen companies that have had digital transformation plans that have been one year, two years, three years long. And then, now they're talking, "Okay, we're going to accelerate and do this in two weeks, in three weeks." And it's actually becoming possible.

Ashwin:

So, what was thought to be impossible is actually becoming possible. We're seeing that if people really want to get these things done, they can. So, that adoption is exciting. It has potential to be... Digital has a potential to be much more personalized, more predictive than brick and mortar commerce. So, it offers a better experience for the customer, and it is good in other ways. It is good for social good as well because you can argue that it will reduce to some extent the impact of climate change. Less traffic, less congestion, less travel. And people get more family time for exercise or hobbies or what have you. So, digital commerce brings with it a lot of benefits as well, which I'm quite excited about.

Stephanie:

Yeah. Yeah, completely agree. It definitely seems like a lot of things have set up very quickly, and it's interesting watching the companies, there's a couple that I've been following, who just aren't moving to Ecommerce. How do you view companies like that who are taking a strong stance not to go online?

Ashwin:

It's interesting. While I do think that the industry at large is moving towards Ecommerce, and not just digital commerce, but digital everything, right? Digital entertainment, digital customer experience or digital communication. Most brands will need to do that to be successful. But sometimes there is always a market for those few contra players, right? Because there may be some consumers who may just want... Not want that new approach and new technology and who make the stand. They might have a spice boutique customer base who works with them. It won't last forever, but maybe it might help them for maybe the next seven, eight years, I don't know.

Stephanie:

Okay, so when it comes to everything that's been happening with the pandemic, how do you lead in times of change at your company or personally?

Ashwin:

So, a leader that has inspired me from his book was Satya Nadella at Microsoft. In his book, one quote that really stood out for me was he wrote that the C in CEO stands for culture. And he also talked about the importance of empathy in leadership. And so, I have taken it upon myself to foster the organization culture that we want to have. The culture that ties us together, and that is really helping us in these times. That culture that we fostered, which is bringing us all together.

Ashwin:

Along with that empathy, understanding of the challenges people are going through. It's not just being work from home, but anxiety about the disease, anxiety about the economic future, and along with that regular communication, transparency in communications. These are some of my key priorities, which are driving some of my actions in this crisis.

Stephanie:

That's great. Are there any challenges that you face when it comes to working remotely?

Ashwin:

So, it's actually been quite a surprise that when the pandemic hit and when we have to transition to work from home across all our different geographies, and 1,200 people moving to work from home, and I'm really surprised at how effective it's been. The work that we do is iterative, it requires collaboration and things like that, and it's working fine. And it seems like thankfully this happened at a time where all these technologies had evolved like Microsoft Teams and Zoom and others where it's still become very much possible. So, now I wonder why we used to be in offices all the time-

Stephanie:

Mm-hmm (affirmative

Ashwin:

Exactly. And so, it's not just us. It's the whole industry just thinking these things. We still need offices for bonding for having certain hard conversations, for inducting new employees, for fostering our organization culture, but we don't need them all the time. We can have 60, 70, 80% work from home. I don't know what the right balance is, we'll discover. But more than-

Stephanie:

Are you changing that at your company? When you can go back, are you only having a certain portion of your employees go back? Or how are you thinking about that?

Ashwin:

So, we've not taken a clear decision yet, but it will be definitely at least 50% work from home, maybe more. Could be 60, 70%. And we'll just have to experiment and find the right answer. We also want to see how things change when the pandemic isn't there and then how that changes people's orientation. But at least 50% work from home, and maybe much more is very much doable. And it will give hopefully people a good balance of engagement in office, and at the same time better quality of life because of [inaudible] and things like that.

Stephanie:

Yeah. Yeah, completely agree.

Ashwin:

But in terms of the pandemic, while work is not suffering, but the bigger issue is the emotional challenge. As I said earlier, people are not meeting their colleagues, not meeting their friends, so they have anxiety. Here we've tried to do a lot of things. We've tried to engage people through various activities like talent [inaudible] or talent contests and things like that. Yeah, so I'm amazed at the kind of talent we have in the company. People are singing and drawing and cooking.

Ashwin:

We've had different types of training. I rolled out a new skill learning challenge, where I challenged every employee to learn a new skill to enhance their career over the next 60 days. I said that I'll-

Stephanie:

How are you tracking that?

Ashwin:

So, it's not mandated. It's just a challenge, which is free for everyone to sign up for or not. But at the end of 60 days, if you have done it, you can report, and there'll be a prize for person or people who have enhanced their career prospects the most, and prizes for leaders who have helped their team members enhance their career prospects. Just to set an example, I said I'll learn how to program and write code in Python.

Stephanie:

Oh, man.

Ashwin:

It's been fun. I'm doing it with my daughter and it's been fun. And we also are providing some resources for things like mental health counseling and things like that if people are feeling anxiety and depression.

Stephanie:

That's great. Really good examples of things that other companies could take and implement on their own too. Especially the talent thing, that's really fun.

Ashwin:

Yeah.

Stephanie:

All right, you probably get this question a lot, but I have to end the interview. So, I'm sure a lot of people have asked you this or want to know this. Do you think that AI will replace jobs or will it just augment jobs and maybe some will not be around anymore, but it will also create new opportunities? What's your take on that?

Ashwin:

Sure, that's a great question, and that's a question that so many books have been written on it, and there's so much discussion in the industry. I'll just give you my point of view, and there are some people who think quite differently as well. But AI, what AI does is AI doesn't really automate jobs, it automates tasks. When you think of a job, any person's job is made up of a bunch of different tasks. Typically, what we've seen is the AI systems will automate some of those tasks, so that person is not becoming redundant, but some of their tasks are freed up.

Ashwin:

And so, then that gives the opportunity to use that individual to then do other things. To drive more personalized experiences, to take the businesses to the next level and things like that. But then that requires that right orientation in that individual, and then requires training. The company to provide that training to those people, or for the people to also take interest and train themselves through resources available. So, like I said, I think that now everybody in this new environment will have to consistently be training and upgrading their skills.

Ashwin:

But do I think that AI is going to come and replace other jobs? No, I don't think so. I think it will free us up from certain tasks and will enable us to widen the scope of [inaudible 00:57:19].

Stephanie:

Well, that is a good positive way to end the interview. Before we move into the lightning round, are you ready Ashwin?

Stephanie:

So, the lightning round is where I ask a question, and you have one minute or less to answer. And we will start with some easier ones. Once you can travel again, what's up next in your travel destinations?

Ashwin:

Oh, wow. Okay, well, I'd love to go to the Maldives.

Stephanie:

Great.

Ashwin:

Yeah, I'd like to go and do some scuba diving.

Stephanie:

Sounds amazing. What's up next on your Netflix or Hulu or wherever you watch TV shows? What are you watching?

Ashwin:

So, I'm currently watching Money Heist on Netflix. I don't know if you've watched Money Heist-

Stephanie:

I haven't.

Ashwin:

... know Money Heist on Netflix.

Stephanie:

No.

Ashwin:

That Spanish adapted English show, pretty cool. And I also like the historical genre, so I'm watching some shows like Last Kingdom and Vikings.

Stephanie:

Mm-hmm (affirmative), sounds good. I'll have to check those out. What's up next on your reading list, other than Python 101 manuals?

Ashwin:

So, that is taking up some of my reading time right now, and I'm trying to be disciplined about not wasting time consuming just unlimited amounts of coronavirus news and doing more productive things. So, right now I'm reading The Alliance by Reid Hoffman, and I'm also reading Why Should Anyone Be Led by You? Robert Goffee.

Stephanie:

Oh, I will have to check out that second one, I've heard of the first one. All right, the last harder question, what one thing will have the biggest impact on Ecommerce in the next year?

Ashwin:

Well, what's already had the biggest impact is the COVID crisis, but in the next one to two to three years, I think it's going to be AI.

Stephanie:

Well, that's a perfect way to sum up the interview. Ashwin, thank you so much for coming on the show. It's been a blast, and we'd love to talk again soon.

Ashwin:

Thank you so much Stephanie, it's been a pleasure. I look forward to talk to you soon.