AI-Driven Supply Chain Transformation | Heineken at AIPCon 6

all right guys good morning from blue to green from Telecom to beer right so let's talk beer it's a little bit early but um when I my name is Lawrence vro chief operations officer so um when I enter the room this morning I thought wait a minute it's only been six months since actually I was in the audience and in those six months actually it feels much longer but

in the six months I didn't know before coming here what p he was 6 months ago somebody told me a week before gares invited me to join this AIP Con in in September 2024 afterwards we sign made an agreement we did two proofs of concept and guess what we're presenting here now so it could happen to you as well because Palante is about speed um moving on to what our proof of

concept first of all we did a commercial one so we helped our team develop an application what is the next best item to sell in 350,000 accounts right so but we're not going to talk about that that today maybe another time we're going to focus on supply chain because that's what I do uh I have a team managing an international supply chain so we produce hinin those Eis Takata beautiful

Brands uh in Europe Mexico and some of actually in the in the US we import 33,000 containers um to nine demand points or nine warehouses across the country uh we're shipping 66,000 trucks uh uh on an annual base to 450 uh uh customer customers and those Distributors ship it to the Walmarts

the uh convenience stores and the hotel bars and restaurant and our overall aim is to make sure that whenever you want to do this that that beautiful green bottle is there right that's aame now before we engage with Palante here we were already on a digital supply chain Journey because guess what during covid many of you probably experienced that it was a mess we got breweries

having to close uh ports congested trucks not showing up but what we've been doing over the past three years it's like anything in supply chain any disruption has a tendency to repeat itself so why not solve the problem when we're in but then find a way to connect data in such a way

that you can detect that similar problem earlier because if you detect the problem early you don't go that deep in terms of mess and secondly the correction mechanisms which are typically Manual of speeding up or slowing down in terms of over Supply how can we automate that so we be a whole

Suite in powerbi uh connecting the brewery the warehouses all of our customer data so we know exactly every day how much they have on inventory how much they're selling we started to track via Ocean uh uh satellite tracking uh the vessels to connect what the containers are coming in the Big

Challenge there is to make sure that whatever is in our data like a handen 12 pack at a customer is something else right so how do you connect those because then the combination of data is not 1 plus one is one is two but 1 plus one actually is five right then we buildt some algorithms on top of it

to be able to speed up or slow down however that came a little bit to an end because um connecting all this data we wanted to add retail stores it became too much so and also the algorithms had to be started by humans could not run all the time uh and you get like hundreds of follow-up

actions also my previous uh peer presage said that as well you have to have the execution capacity to implement all those changes so on a personal note I love formula one so I always try to make a link with that by the way the races are starting again this weekend but um we had a great

chassis of the car but our engine was underpowered so we went to Palance here because we wanted to have the best engine in out there in order to make this nice combination of the chassis with the engine now bear with me for uh the next slide because I think first of all we we thought we are getting a great engine which we did but we also got a great team to work with so the P Engineers

were super smart great people to work with uh uh we had also some good fun so basically beer and Tech it's a great combo right um obviously we were only drinking H and zero while building right so it was not it was like that so but but the team really jelled well well together so and that tiger

team actually in three months build what took us three years to build before right so that gives a little bit of the magnitude of of speed um so bear with me on this slide because actually this is the main message of my eight minutes fast forward your supply chain using AI to solve tomorrow's problems

before they happen it sounds odd but actually that's what we do now let me give you one example and then we go through some real life examples in in The Foundry software um if you have a warehouse and you know for 3 weeks what orders are going coming out and you connect that with ocean Freight data about what ship is coming in you can and of course you know how much you have on inventory you

can basically just like in a movie or on Netflix you can fast forward your supply chain three weeks because have three weeks of outbound data three weeks are inbound so that means that you can see where the problems are going to happen and in the supply chain there's always imbalance there's always you know a warehouse that has too much or too little of a product and but you can also do

that at customer level at Port level at Brewery level and we connected all of that I'm going to show you three algorithms that we use we have many more but just to give you a flavor of it it's going to be a little quick so um what you see here is let's say the background of people

love Maps so these are the maps all the stores are in there the warehouses in there you see the flows uh we focus basically on the west coast because we are here now um there's a potential future out of stock so we have a new algorithm called Abu I'll get to the name later but Abu actually looks at if a customer has ordered the product and for example because we have a lead time of three

weeks which is normal in the beer business it has a risk of running out of stock while we are still having to execute the order Abu is saying you got to pick up that product in this case I hand 00 uh bottle uh sixpack and automatically move the delivery date earlier so whereas normal Supply

chains run fiveo so who places the order first and we execute in that particular order actually this algorithm helps us to pull something ahead of the pack to bypass a few others and make sure that actually the product is available at a the C because in the end remember what I said when you want to do this we we need to have the we need to have the beer there right so it's about deviation

from a previous process let's move to our next one um next example is about drage drage for those not involved in the maritime industry R is actually getting a container from a port to a warehouse now if a vessel arrives the containers are always stored in this stack the first few days are for

free but after a few days you start paying a lot of penalties so Dem merage and detention is a big cost so that's we want to prevent so typically the General General process is the container that has been hanging out in a port for a long time will be picked first we have a new algorithm called Dr

Dre and Dr Dre will pull it from I mean in this example was the H the M you know if you were looking at it the M was just arriving and we got to pick it from the stack immediately so again an example of acceleration putting something ahead of the pack now those two algorithms we we were using before however with Palante we can use them all the time not dependent on human intervention

also we can scale executing all those deviations now the last example because I think I'm running a bit out of time um is smart shipping so making beer will cost like three weeks it take you three weeks to brew the beer maybe one second to package it and then we need three to eight weeks to you know to get it here depending on where you are New York is shorter Seattle takes a bit longer

to get there now what what we do is we brew and produce on forecast however by the time that all the resources are arranged materials people Etc and you're about to fill the bottle or the can things will have changed so all of these uh uh imbalances can have moved so we together with

pante here we developed an algorithm that can take two days before the container is being loaded take whatever is produced take an existing booking and swap what goes where so an example that was in the background we needed a hen 12 pack was planned to go to Houston that cargo we're going to put in a container going to Seattle because again we see in the future

where the problems are going to happen now uh you had me talk about Abu Dr Dre Neptune those are creative names and it's very important because the users came up with those because I want to make sure that whoever uses AI or algorithms owns the algorithm and doesn't see it as a threat

so now in our DCS our daily control system we talk about oh Dr Dre said this and is doing that so you bet you're adding a person to your team a sidekick to the people and Empower them secondly I think it's also has change has to be leader l so for me as head of operations I'm the probably one of the biggest users of our Command Center algorithms Etc because if

I adopt the change I know my team will and and secondly reward and recognition super important we have inside the Hing group we one of the best Supply chains on many dimensions also our customers are sharing with us that our availability for the last two years is better than

most even domestic producers which is something I would have never have imagined so what's next we're going to we show the peak sneak peek under the hood that everything is in Foundry and all the algorithms and are moving and and doing things in order to actually speed up or slow down we're going to a second stint with with palent here where we're going to move all of this into

production so so far we're proof of concept we're going to move in production actually drive the benefit in the year where we actually implement it so that's all next and um let me end by saying I think Frank and I are head up digital we're here we'd love to talk to you we would like you to challenge us because we want to be the best and then we should be open to sharing what

we have and make it better together with you and last but not least my final note is fast forward your supply chain using II to solve tomorrow's problems before they happen thank you very much