Session Information:

Chief Procurement Officer’s 2022 Outlook

Session Transcript:

And this is another great leader of intelligent automation in the supply chain, and I'm talking about definitely Liu who is here with us. She has, over 10 years, an enterprise-level, big data, open-source platform, with extensive hands-on solution design delivery in big data, analytical AI, and machine learning.

Daphne is an experienced team leader with exceptional strategy planning, innovation, creativity, and project rollout skills. She's a frequent seminar speaker and workshop trainer in enterprise-level business intelligence and analytics machine learning solutions and artificial intelligence Daphne on behalf of our global audience, thank you so much for being with us and sharing your expertise with all of us.

Good morning.

Thank you for joining me this morning.

Today, I'm going to talk about the C Vasa supply chain green journey. First, let me introduce myself. I am the Head of AI and ML in Steve as an admittance Matrix Supply Chain Management Application.

We serve a business, a supply chain, globally, who's solve today's at global CO 2, 3, requirement, or compliance growth, seba, with Evil ourself, to improve our supply chain CO two consumption.

Starting, 2021, I have been involved heavily with our CO two supply chain project.

Today, in this presentation, I will share my approach using AI and ML tools, or auguries them to reduce the CO two consumption in our supply chain management.

So, on the screen?

OK, here.

So you guys see on the screen right now, this is Ally, my, OK, and the sly.

You can see, I am sorry, my screen. Go to the.

Sleeping Mo.

OK, so this is our three green components, so we bring in, OK, so this is make sure that it's an application we use to manage our supply chain.

So With my screen, OK. So that's a three components.

one, is a CO two, CO two aware, one is SEO too intelligent, one is the CO two discovery.

So now let's go to the next slide, that's our, right now you are seeing our three components that I'm going to talk about.

So before we introduce our component to reduce the CO two consumption, we need to understand the CO two formula.

So as you probably already know, some, some of the CO two calculation is based on your weight, based on your distance, and based on the CO two factor.

So this is three factor, made up, too. We multiply them together. We get our total CO two consumption.

So, in this is re factor to us, like, in a week, when we management management, our supply chain, what do we need to consider, OK?

So in way, we are looking for package, how do we package them? Our pallet.

How do we make them to a container?

In this sense, what do we need to think, OK, we need to plan our C D pair, this CD pair? What is our optimize distance?

Our poor pair and our stops for the CO two factor?

Wow, our truck size, the airplane type we use, what kind of trip involved in our plane in our role, Like it's the median, the stars, or it's a long distance trade.

So, in all a bar, I'm going to speak our solution in, supply chain, in other, in other war, is to reduce in a package way, as much as possible.

And we want to use an optimize trip distance rau, we need to plan in advance.

Also, we want to consult no, two pallet or container. And we want to use a large trucks sites carriers.

And we want for the air, if we need to use a wrap, we want to use a cargo plane, rather than barely free.

And if that timeframe permits, we want a ship with a rail ocean, our explain why?

Go to Next Fly.

Here we go. So. Our approach.

OK, so, for our CO two redaction, we are going to optimize our FTE out.

I know you have a laptop LTL. So in our goal is we going to consolidate to make more FPL?

That's our goal and we We make a decision We want to replace as much as possible.

The ocean the air route, or the ground route with rail route an ocean Rob And all this.

It needs to be done No, in the order time or scheduling time, so you have a time, you know to plan all of this OK, so in the air and grom that consolidation This part we got to give them consolidation to alert.

So they get it, so we will suggest to our customer we think this is yours?

Opportunity to consolidate Package, you know into a pallet or even go into OK, and the air rau we research in advance.

What is our mini must stop, OK, because your air like your plan?

How do you conduct them, that matter, reduced to minimum distance?

And we also want to choose the CO two friendly airlines.

So now, let's talk about wrong part.

OK, so in the grandpa, you can see on this chart, this is a CO two factor.

You look at here, so you can see if you ship a package in airplane, if you calculate this factor, is 40, say, things more than you go. ocean ****.

So the ocean ****, of course, is the most CO two a friendlier carry that the route. But however, not all your package, you can go with a ocean route. Sometime, you know, you just need to go with the air.

Or sometimes, you really need to use truck.

So you just have to keep in mind, if we cannot go to toggle, we cannot go to trend, OK, so our solution come in.

So, wrong. What do we need to do?

So we know the factor between different trucks.

So now, you can see, for the 3.5, 7.5, you know that the different sites are tracked.

You want to use a larger sized truck, your solution first.

OK, so we go into beta and weight, but that also has a cost to consider. So you have to find, you know, your perfect solution between that, pick up a good truck size. Also, it's an economy for you to do, ship your package, or your palate.

And also we are looking our trip consolidation algorithm. We base it on the origin. You know where your ship it, we base it on, or poorer could be your air. Paul, what happened to that pool the what happen to the ocean port there? You know, we give them visualization and we will do our adjustment.

Of course, our destination, how many good, and we also do forecast, how many good goal, but there, or should we keep them in the hob waiting for a big trial coming to carry them? You know, So, we do have a consolidation algorithm behind this is a little bit behind our solution.

And we also has an algorithm, we call it shipping unit optimization.

You know, we're looking for opportunity to tailor our operation, That's an opportunity.

You can consolidate the package become a bulk, OK, either to most of the time to a pallet or even to a container, then we gave them the rail alert.

If we think this shipment, we have that window, that timeframe is, it's for me, OK, then we will send the rail railroad alert to our customer where you can see the let's go rail.

So now, let's talk about air.

So on the airside, no, we are choosing the Bailiff re read. We want the Kaggle plant because of the Bayley Frey CO two factor is higher than the cargo plane. So you want as much as possible cargo plane.

And you also want to prioritize. So the air carrier, you know, some airlines, they have a cargo plane. Some, they don't.

And we also prioritize our air carrier. We have an algorithm. So from this CP pair, how many stopped they got?

So for example, if I want to ship something from Bangkok to Chicago, I want to choose maybe only one stop between them. I don't want an airline. Go to two more stop because that it's created a distance which you can save for CO two consumption.

We also use the CO two airline Index. This index, you can find it.

You know, it basically prioritize the ranking by the green favorites. So usually a newer airline, they are ranking higher than the old way.

Oh, airline.

And if they do more green effortless that's on top of our ranking, we will use them more than those airlines, they didn't do any effort.

And of course, for the air we are packaged them to palette, even container. But pellets are easier.

Sometimes you just cannot of goods to go to a campaigner.

And then of course, we will give you ocean route alert. In this case if we find out your deliver day, six we had, we will give you this kind of say, OK, that's what you should do. Give that alert to our client.

So you are going to see in the next I'm going to present the intelligence dashboard. In the dashboard you will see CO two emission index. You will see in this dashboard you will see with our forecast for their CO two in their trend and the seasoning we use our linear linear regression model. You will see the result of that. You will also see how do we provide the trip consolidation opportunity.

And you will see we have a CO two emissions saving calculator, and we can go to our dashboard now. Sorry, I need to.

In fact, the screen.

Here's the dashboard.

When I share a screen, I lost my resolution.

We can see your dashboard just fine in the previous shot, in the previous law, OK.

So we will do this slot here, OK. Yeah. Because my background, as well, when you switch to the other application, was fine on this side. It looks fine, OK.

So, OK, You can see it OK. So, OK, fine. Because on my side, it looks.

Can you see normally, we can see normally Yeah, OK, in my site, I share, it just looks weird, but I'm going to just, can see it. So the first one, we show the CO two. This is for the dashboard we show our clients. So, they can see their Guang, how today, a CO two consumption, by growing and by air.

And we, we have a, we break up that consumption, bimonthly also has a running total and we also can see America. You can see the map is for the consumption for the CO two for America and Europe because they use a different standard, different formula, OK. And you also see the SSI, CO two consolidation opportunity by origin.

So you can see different origin: You know, who is my number one, origin produce, the most of the CO two, and you can work on that part and this dashboard. Well, you know, a trigger, so they can see that calculation. That way too, you can see that.

And on the right hand side, you also see the index we call a CO two index: The forecast for you for your air, for your brand, for your America or for Europe.

CO two consumption.

So you can plan like this year, you know how much and we will compare your previous year to try to save the CO two for you, OK. And then this button this a troop consolidation opportunity is a pollutant shot, OK. So this balloon child will show you there's a big balloon show that's an opportunity. You need to look at, you know, in that trip in that place, in the origin or maybe destination.

Because you are running step seven several different trip, you should be able to consolidate together to do that, can't see duration. So, you may want to refresh the screen, because now the dashboard is not displaying correctly, but I don't It's not a feed issue. It's on your computer, though. There is something that's not updating the graphs. We only see kind of fuzzy images on the graphs.

OK, so I know that's OK. So let's use this one.

Instead, we will use this one.

This is a screenshot.

I know my, my screen resolution is Jessup change into F two. I share this.

Orca?

So, yeah, so you can see the index. So you can see this balloon here. This balloon char we'll tell you your opportunity, how many trees you should consolidate together and here we give the guideline. and we'll tell you on this chart, you can see if you consolidate how much your CO two can see in this chart.

OK, and if you follow the execution plan How much more you can save? you know What is your airline consumption? If? You'll go through Airline if you will switch you know like our execution plans Subjection of device you will save.

So it will be from 4,000,004.63 mainland you can save to 3.72 and same thing on the ground and your Rail so we calculate that for you so I'm going to the next slide.

OK, when I Sorry, I'm going to do it this way.

Foster. So, sorry.

OK, so we can be more green, I guess. You know.

We do have our first step two to help our customer, to reduce their CO two consumption. So what have we are working on now?

So that's what we are working on, OK.

The CO two, C T, we want to have a discovery panel. We got this feedback from our customer. So they want to freely to check from this C, D, to the CD, what will be my CO two consumption, and what is the best route for me. So we are working on that.

We also has a forecast, you know, based on air and ocean port. So in the future, we're going to show them your airport. What is your airport CO two consumption looked like. What is your ocean port CO two consumption looks like.

And we will give them like a calculator, so they can put their test on their way and we'll let them visualize the CO two.

No, in the dashboard, like the one I, and, of course, our consolidation opportunity of device, it can be better right? Now, we send alerts, but right now, we want to make it even smarter, it can alert, like, you know, if we find certain days, you can count solids and together. We'll immediately send you the alert before even the shipment.

No.

Before the shipment started, and we have a route weser to subjects before you plan your route, and we're going to have a raw weser to help you to plan your route.

So, that's what we're going to work on in the future.

And, the final, what do we want, is we want a calculator for CO two and which will show our customer, their CO two shipment, the whole shipment, and their execution plan. How do they execute this trip? You know how to run it. Run this raw, how to get their package, you know, so we will ship it.

A CO two formula in there to help CO two consumption.

I'm sorry for my screen today, you know, and thank you for joining me today. Let me share Matrix, a C bus metro, green journey with you.

So if you have a question, I am very happy to answer. Thank you so much Daphne for sharing that journey there with us.

With civil logistics on the, on what's normal, on how you're really implementing AI and machine learning solutions within the business, which is a very unique application use case. You know, ESG has become so important in the industry, on the investment community as well. As we, as we make better decisions aligned with ESG Matrix. And you're really showing an example of the practice of, of what that looks like when you're implementing the strategies. So I know we definitely appreciate that in any questions that the audience has. Please make sure to ask to ask, ask those questions. I am I'm going to ask you Daphne to stop showing your screen to the audience. On your goto Webinar interface here we can hit the button that says, stop showing screen to the audience.

And then I'll bring my camera back on for the Q&A. Thank you.

And that definitely my first question has to be with, uh, how do you, even get started was, I bought on, where you are. I mean, you are now in a pre, the advanced level of development here, where you have metrics, identify where you have a, you know, a machine learning that took place, and you some, you know, AI. There has been working on developing what is the best model for what's going on in your system and making decisions for your operations, which is really fantastic.

But the reality is that there is a tremendous amount of work that you and your team has put into this to get to this stage.

So I want you to help us.

If you can rewind the tape and go back to the beginning, when you started your journey, where did you started? How did you, Steven start, you know, building this the systems?

Thank you. That's a great question. Jose. We, we are purchasing AI and ML tool. Let's say they said, Look, at two years ago, at a company, I want to spend a lot of money. You know, every company is doing that. So they are looking for a POC idea, you know, everybody, every team, you know? So then I brought this idea because on the other hand, the company is working on CO two because we have a lot of the Europe of client.

And, you know, CO two in Europe is very, very important to them, you know? So then I brought this idea to the execute him.

They are, know, very excited and they support it. So that's a two years ago. So we have it, in other, what, we already have, our AI and ML, a tool. We use, We purchase a, you know, and we set it up. So this is our first POC. And, like I say, the first one, you already see the result is that dashboard, unfortunately, my screen today? The resolution is now that good, but I can share, you, know, like if any, one interesting to see that I can make, I need to mask some data. Do you know what I'm saying? I can share that, hopefully, will help. But others, that, you know, their journey.

You know, so that if you ask me how I started, that's how I started. And we just take it. And because my background is from the Analytic it so I already done all of this ... dashboard in our company. So when we bring this AI and ML, actually, I just added this AI and ML recipe to our current.

So it's not like we start from beginning. We already have that.

So, and like you say, I'm still exciting. Stop the future. That's what me and my team, we are working up.

So, we will bring more exciting to our solution, you know, if there's an opportunity in the future, I will share.

Absolutely, Now, this is fantastic. Because, again, you're showing your presentation, is not about showing the pretty stuff, and what it looks like for a high level.

You are, really, delve into the details here, and we all know that the devil is in the details and that, and how hard it is to actually implement something and make it operational. And you, and you, and your team, have done just that, it's operational now on your logistic decision making.

Yes, so, let me go back to, seems to talk. This is a more of a detail implementation, which I think is phenomenal, And, we appreciate you sharing that with us.

Let's go back into the you said it very well that you started you had an analytical foundation.

Like, you had relationships with your business, and understood the operations, and understood what matters, and understood the metrics that you need to track.

And, by the way, for the rest of the audience, that is not as experienced. Don't take that for granted, because that's a really hard thing to do, because I didn't know. I just, I just, I just get into your organization to align on what is that we're going to measure, how we're going to measure it. And then you establish the foundation. Well, good. That's very good.

Now, tell me, specifically, now that you have the foundation established, you're collecting some measurements.

How did you build your machine learning and AI engine on that to take you to the next level? What are the steps that you take to build? Now, this? They are the AI and machine learning engine on top of the analytics that you already had?

OK, great, thank you for that question, So, you know, we are a global company, so sometime we have to collect that data, it's not like an in house. You know what I mean? Like, it's different, you are dealing with a different time zone, different people different grew, the business nature, you know. You have a different reason this. We are serving their prison this rule at different. So, you kinda need to align, but because of the CO two calculation, like I say, what are we looking for? weight, distance, C, L two, factor.

So, based on that, co-ordinate with all your team different beads and as a ground air ocean, you know, collect it. So then you POC was one and to show the rest of the group because everybody's a little bit different, you know. So every good, especially operation, then, yes.

At first, it's hard because you are making operation to change their habits. Say, hey, from now on when you do this, you need to look at this. You need to look at the track. You need to say palate, tell me your truck size, tell me your palate. We want a palette.

No like that. It's hard. I will tell that is the hardest part. Even though today I don't have all them, but we have some of it so I can bring it in for some customer. So that will make customer very happy, because I got a visualization. Now. You are showing me, once I have this year, next year, I'll tell you, what is your goal this year you want to say. And we will monitor that go for that.

That's why I call a CO two keiko later.

So once I implement that, and I do have, like I said before, when would you choose to the algorithm? You know AI ML is very early. So now, a lot of how do you say user friendly tool out there? But this year is getting more and more better. We got more, and more like a tool we can choose from.

So we use: I, I don't, can I say that tool, I shouldn't say that tool.

It's OK to say, OK.

They didn't pay me, OK, I just wanted to share the tool we use, OK. We ended up landing this tool called Data IQ, OK, So it does save our developer time with AI and ML.

I'm very happy with the product, and the execute is very supportive.

No, because it's very expensive.

Also, we have to spin up a spot cluster to help the .... So, once we have that, it is like, you don't have to do the AI mail from the scratch, now, you know? It really speed up, you know? And so, we can reduce our developer, maybe one months.

And we got, you know, we have a bit.

The dashboard you see is, is several algorithm together, is not just one algorithm, is a combination of the algorithm and rizal.

We bring it into our analytic engine, and then we produce that.

Know, so, I don't know if that answer your question. No, it does. It does. I mean, there's so much in terms of complexity to doing something like this, that we probably can talk for hours. But, yeah. This is a good view of what you have done now. Let me ask you this now that you are developing the, the algorithm is getting improved. If you, if you will. What are some of the ways that you can check the Algorithm to make sure that the decisions are the right decisions, Do you have a system for verification or validation of the algorithm.

Yes, that's why we gave Dashboard.

Now, actually, right now, we have two dashboard, but because of today's time and my monitor or resolution, I losses, so I didn't show you the second dashboard, So that's a two dashboard to show historical data. That's the one you guys just saw today. We have another one is the future is like the other day just Place.

So, we might have up some, we have a two days window, because that's an a priority.

Some, we have a 2 to 3 weeks, so that we can do something.

So immediately our solution was our client put it in the order, OK, and we run our engine probably 24 hour. We spit out this dashboard to them immediately.

So, so they can look, of course, this is enough. In other words, they are checking the client checking because we go into subjects, the opportunity, and then the operational people are checking, that's for them to follow before they don't have anything.

So now, based on that second dash, what we call the forecast, the future, and then they all have a CO two discovery, if I am I consolidate this together, what will be my CEO.

So we will have that kind of, so they know, OK, is that possible.

Validate, maybe to trip because it's a stop, You know what I mean? So, you cannot, you cannot do off. So sometime, we merge, you know, even to tomorrow, three-d. on jelly.

Helping them doing this, kind of CO two ..., then they will place in the final order. Right, now is still the operational people.

They prove it and then they send you an answer. It's not like the whole thing automatically. Now, we're not that. Well, yeah, but 95%, our algorithm is pretty accurate.

And sometimes our algorithms even smart and surprise us, like they'll do an even smart decision then those people think, oh, I can do that. I never know.

Know, so things like that algorithm will use. The older that. I like it very much Surprise me at law, it's a neural network.

That one is, like when we do the Rao, no optimized route is, surprise us all the time. Because think about, your operation cannot know, all the Raleigh seba.

So once you implement this knowledge to your algorithm.

Very, very powerful. That is it.

Solid and Honduras Rao you know, for different business, different me and it just start smart to pick up the ... we call it optimize brought for you and every Rao after our research, we tie into a CO two factor. The calculation that Rob, like I say, everything you need pre, planned it, don't just do it randomly.

Know? So we are. Once you have all that place, you can do, it's, it's, you know, I know, like, you say, I can talk about it all day long, I mean, you can see my excitement, because the Some of the stuff I see, Wow, this thing is so smart.

I really enjoy it, and we're going to do more, and right now, I'm in the Showcase to be, isn't it because you need to communicate with the business operations, So right now, you know, like, this week, or next, next month, I'm in the Showcase them all to show.

The Hosur teaches them how to use it.

Well, that is just fantastic, Daphne, and I hope that you can continue to develop this, and the, and then share with us, you know, what the, what the future developments look like as well.

Yeah, I have a final question here from you from our audience, and the question is about the what advice do you have for someone right now who is a leader in supply chain logistics? And is implementing, this, is thinking about implementing these more sophisticated approaches?

Should their systems, to their processes, to their data, you know, what would be an advice for someone at that, at that stage, two, to move forward with?

artificial intelligence and machine learning, What would you suggest that they keep an eye on?

Of course, Yeah, I would love to, because that took us almost six months to, like I said, we interviewed vendors, you know, until, like I say, I pick that IQ, You know, and also we have, you probably already see the dashboard, the Analytic A playful you accompany need to set up that setup that, too. and the reason we picked, Like, Now. I pick this tool like my analytical tool, and you call this date IQ is everybody will think, or an emotional is so hard, Like, It's so difficult. How do I started? That's why you need this tool. I pick this, too.

I find out, hey, it's an easy, not, because I've been training people all the time for it. So, once I learned this little tool, I feel this is easy enough.

UI is easy, I can teach you this, because our goal is not the whole corporate, like the whole enterprise, depending on one team, like my team, it's like that.

five people, do you all? I'm saying, so, you, I want a tool. I'd be able to teach AI and machine learning tool.

So, once you set that one up, you make not a high anymore. They got AI. Machine learning is not like there. Any more is coming here to my team, and to my business, to my operation and people, I can teach them.

So, I can tell you, today, using data is not only me. now.

We have a business Baffa dashboard, or some, you know, do on Power BI, capital, do. You don't want that kind of skill set.

I can teach them to use this tool. In other words, it's me, it's, it's influenced the whole enterprise, want to do that, because I'm able to teach an easier tool for them to use. So that will be my suggestion.

You will have to start with that, know, and this is just They didn't pay me money. OK, I want to say that this is just my journey, and I feel this tool helped me a lot.

Daphne, your excitement is palpable. It's wonderful to hear and see the progress you have made. This is a real best in class application of AI and machine learning shoe supply chain. A lot of people talk about it, but they talk in general terms.

You are implementing ESG related metrics to the effect your day to day decision making that is leading application. So we want to thank you for sharing that wisdom and expertise with us, who our audience, What we had several questions about, how do I get in touch with Karen with, definitely, directly. I have put on the chat to take a look there, the link, and that link to our poster that the conference.

And then on that link, you can see Daphne's name on LinkedIn, and you can click on it and then send her an invitation to connect, and the talk directly with her.

We want to thank you, again, Daphne, for sharing your wisdom and expertise with a global audience today.

Thank you. Thank you for inviting me and give me the opportunity, and let's keep I'll keep you updated. How about that? Yeah.

Dying to see how this journey continues to progress. It's very exciting. Congratulations. Again, to you and to the team.

Thank you. Thank you. Thank you. Bye, bye. Bye, bye, Bye.

Ladies and gentlemen, that was Daphne Liu with Seva Logistics showing really great applications of artificial intelligence and machine learning related to ESG Matrix and how that's impacting their day to day work in decision making.

You know, when we get into the details like that, it's, it's, it's the, the It's wonderful to see What it takes to get to two great applications that impact decision making. Like, the one that definitely is showing she had a little bit of trouble with the resolution of her monitor, but you've seen from the PowerPoint presentation that, which is a screen capture of the dashboards that they use. And she has volunteer to share with you what those some of those screens look like. So, some of you are doing some benchmarking and the, and you can connect directly with her, again, use the link that I put on the chat. And then on that link, you know, find her name on LinkedIn and send her a connection request on LinkedIn. We're going to wrap up the session.

And I am going to step out, because I have a major commitment that I need to fulfill for the next session, but we're gonna have Chris Hodges, reviser's going to be introducing our next speaker and wrapping up. They chew for us and our next speaker is coming from variable and that I'm talking about go ****, the Vow. Who is the director for the Lead Center of Excellence, every variable.

And, and he's going to talk about rethinking just in time for supply chains. So is going to lots of experience only, for sure. But he's think, and he's going to talk about just in time and the evolution of just in time, especially the face of all, the variable demand and supply disturbance that we have had in the in the in the marketplace.

So great, practical session for someone who is a true leader in practitioner of improvements, any innovations and transformations in supply chain. So close the session for now. We'll be back with Chris Raj Shah: Just at the top of the hour. Thank you and bye bye from me for today. Have a great rest of the day, everyone.