AWS re:Invent 2022 - Optimize industrial operations with AWS IoT (IOT312)

Aug 16, 2023

AWS re:Invent 2022 - Optimize industrial operations with AWS IoT (IOT312)

Industrial customers use hundreds of pieces of equipment to drive complex business processes. In the past 3 years, 82 percent of companies have experienced sudden breakdowns, resulting in 5 to 20 percent reduction in productivity and in annual costs of $50 billion. In this session, learn about AWS purpose-built AWS IoT, AI/ML, and partner solutions for asset-intensive industry use cases. Discover how these solutions detect abnormal behavior in industrial machinery, making it possible for organizations to implement predictive maintenance and reduce unplanned downtime. Hear from Penske and Hitachi executives about how they use AWS IoT solutions to derive value and insights from industrial IoT data and achieve business outcomes at scale.

Learn more about AWS re:Invent at https://go.aws/3ikK4dD.

Subscribe:
More AWS videos http://bit.ly/2O3zS75
More AWS events videos http://bit.ly/316g9t4

ABOUT AWS
Amazon Web Services (AWS) hosts events, both online and in-person, bringing the cloud computing community together to connect, collaborate, and learn from AWS experts.

AWS is the world’s most comprehensive and broadly adopted cloud platform, offering over 200 fully featured services from data centers globally. Millions of customers—including the fastest-growing startups, largest enterprises, and leading government agencies—are using AWS to lower costs, become more agile, and innovate faster.

#reInvent2022 #AWSreInvent2022 #AWSEvents

Content

4.02 -> - How's everybody doing today?

6.54 -> - [Audience] Great.

7.95 -> - I need more energy.

8.82 -> How's everybody doing today?

10.653 -> (crowd cheering)

11.486 -> - Good, good.

12.319 -> At AWS IoT, we usually start our customer presentation

17.28 -> by asking ourself and our customers this simple question.

21.96 -> that is, if you knew the state of everything,

25.23 -> and you could reason on top of that data,

28.2 -> what problems would you solve?

30.93 -> The key thing to notice here

33.05 -> is knowing the state of everything in your company

35.7 -> and operations,

36.87 -> and to be able to reason on top of that data.

40.71 -> It's our mission at AWS IoT

44.25 -> to help our customers accomplish just that.

49.08 -> Hello everybody, my name is Praveen Rao.

52.08 -> I lead industrial IoT and analytics.

55.14 -> Welcome to our presentation today,

57.75 -> and I'm also joined by two of our esteem guest,

62.4 -> Rohit Talwar from Penske, Rajesh Devnani from Hitachi.

67.86 -> So we're gonna go through an introduction

70.5 -> when they come for presentation,

72.21 -> but we have certainly some very exciting stories.

75.72 -> So it is with this mindset of really co-innovating

79.23 -> with our customers,

80.28 -> and really helping them solve their operational problems

83.67 -> that we have been able to win the trust

85.47 -> of all these major customers that you can see.

89.82 -> And we also have a very rich set of partners

92.19 -> who kind of help us deliver on those things.

94.83 -> You would hear from one of such partners, Hitachi,

97.92 -> on how we are kind of collaborating to help our customers,

101.28 -> to win many of the IoT workloads.

104.04 -> And you'll also hear more stories about Penske

107.76 -> on how they were able to benefit from AWS IoT services,

112.53 -> and solutions that we have.

115.83 -> So when we say optimize industrial operations,

119.64 -> what we really mean by that

121.26 -> is basically customer comes to us,

123.6 -> because they want to lower their operational cost,

126.9 -> and increase their revenue,

128.82 -> and also improve their sustainability.

130.92 -> In other words, what we are really trying

132.84 -> to help our customers is to make their products

135.06 -> faster, cheaper, and better, right?

137.88 -> So we see three broad areas where customers come to us

142.86 -> for help with innovating and lowering the cost.

146.61 -> One is in the factory.

151.23 -> According to the latest study from Mckinsey manufacturing

155.73 -> in general produces about 18th hundred petabytes of data,

160.29 -> which is twice more than the the next nearest industry,

163.89 -> and the 90% of the data is locked in the On-prem.

167.94 -> So if you wanna save cost in your operation,

169.83 -> what's the first thing you need to do?

171.21 -> Move the data to the cloud, right?

172.8 -> Because you don't have to maintain all the infrastructure

174.96 -> and all.

175.793 -> Another thing is liberate the data, and democratize the data

180.75 -> so that different constituents can get meaningful insights

184.44 -> to help them with their decisions in the operations.

187.65 -> Whether it is understanding the machine availability,

191.04 -> OE, yield, quality, uptime, and many other things, right?

195.36 -> Because you want insights

196.92 -> to be able to make better decisions in your operations.

201.66 -> And then moving on,

202.8 -> other major area we see is in the products they make,

207.33 -> so there is more than 13 billion connected products

212.31 -> that are already there and equipments,

214.71 -> so more and more equipments,

217.23 -> and products are getting instrumented.

219.87 -> So that means products are getting smarter,

222.18 -> they're getting more connected.

223.83 -> And also other thing

224.82 -> that we are seeing from a trend perspective

228.026 -> is customers who come to us also want to understand,

231.42 -> how their customers are using the product,

233.85 -> that means they wanna get real insights

236.04 -> on how their products are used

238.32 -> so that way they can take those insights,

241.38 -> and then take it back into their product design,

244.38 -> that's the third area,

245.49 -> so make those products more customer friendly,

248.46 -> more manufacturing friendly, designed for manufacturing,

252.03 -> and then also make them more efficient.

254.19 -> So each in turn feeds into the factory as you can see,

257.85 -> the products are becoming more manufacturable,

261.35 -> and definitely more efficient, more sustainable,

264.27 -> so as you can see there is a flywheel effect,

266.4 -> so this keeps going, and then this is kinda

268.44 -> how you can continue to optimize,

271.77 -> and then extract more value from your industrial data.

277.68 -> So moving on,

279.45 -> so where do we see, customers see challenges,

283.02 -> and also opportunities for improvement.

285.39 -> Right?

286.223 -> So first and foremost thing where we see

287.97 -> is lack of visibility into operations.

290.73 -> What it means really, like we started,

293.76 -> lack of visibility means not understanding

296.85 -> the state of things within their company, right?

299.73 -> So what it leads

300.75 -> to a lot of reactionary measures in their operations,

304.68 -> and lot of manual work, lot of redundancies.

307.86 -> So that's something that certainly an area to come in,

312.21 -> and then improve, right?

313.74 -> The second area where we would see is unplanned breakdowns.

317.85 -> If you talk to almost all the CFOs in a company,

320.64 -> so their biggest thing that keeps them awake in the night

323.16 -> is they go through the budgeting cycle,

325.14 -> they budget, and then certainly something happens.

327.03 -> Now they gotta go find funds to address that problem,

330.99 -> not only that are also impacted would have on the revenue.

333.99 -> So that's kind of where,

336.15 -> so they need to have a more proactive approach,

339.87 -> more approach in which they understand

343.71 -> not only the state but also reason on top of it,

346.5 -> so that move break,

348.36 -> move from unplanned to more planned maintenance,

352.62 -> and pay as you go model,

353.94 -> so that way they don't run into disruptions,

355.92 -> not only in operation but also in the finances as well.

358.8 -> So third area is simply the waste that happens,

361.44 -> it could be waste because of the scrap products

364.05 -> that they make,

364.883 -> or the products that accumulate additional operations

368.37 -> once they're deemed scrap, right?

371.42 -> It could even be the recalls that they have to go through,

374.67 -> and associated brand images,

377.61 -> impact that would have.

379.726 -> And the fourth area

380.559 -> where we see a lot of industrial challenges

382.59 -> is the energy efficiency,

384.21 -> especially in certain geos like EMEA,

386.43 -> where you have geopolitical issues with the Ukraine war,

390.54 -> and certainly manufacturers are struggling

393.57 -> to keep up with the energy cost,

395.94 -> and continue to run their operations.

397.68 -> So they're looking for ways to,

399.6 -> how can I do more with less,

401.52 -> or more with the same in terms of the energy?

406.17 -> So given all these challenges and opportunities,

408.57 -> where do you get started, right, in terms of optimization?

411.93 -> I think the first and foremost where we recommend

414.21 -> is getting a good handle on your industrial data.

417.24 -> What you're seeing here is a classical pyramid,

420.54 -> or ISA 95 purdue model, however the way you wanna call it,

424.86 -> where organizationally, data is organized in stacks,

430.233 -> and they are measured in zero to four levels,

432.87 -> and also it's categorized into OT data, and IT data.

437.4 -> This is great, and it started with industry 4.0,

440.097 -> but the problem with that is

441.93 -> average age of the industrial machinery is about 25 years,

445.56 -> and there is 300 plus protocols,

447.81 -> they don't always talk to each other,

449.37 -> and all these layers are in their own silos,

452.22 -> so there is not much connectivity

453.72 -> or the data exchange happens among the layers.

456.21 -> So that leads to a lot of inefficiencies,

458.91 -> and even the innovation that happens happens in silo.

461.91 -> So what we are seeing with the ubiquity of cloud,

466.14 -> so that barrier is breaking, right,

468.063 -> over the last couple of years,

469.59 -> OT and ITs are coming together,

471.57 -> so the data is not stuck in one spot like it used to,

475.41 -> and there's also newer protocols like OPC UA,

478.62 -> that is becoming more popular

480.09 -> where you can translate from one to another.

482.88 -> So there is flexibility that is creating more efficiency

487.02 -> that would otherwise happen in a silo, right?

490.11 -> As we go into the future,

491.43 -> what we see is again with the the cloud presence,

494.73 -> and the ability to move data from,

497.67 -> not just from edge to cloud, but also cloud to edge,

501.3 -> we expect that barrier to be completely gone,

504.33 -> and so there won't be much difference

506.67 -> in the way things operate between the cloud and the edge.

509.79 -> So some of the newer things like A I / M L,

513.09 -> and even digital twin,

514.32 -> so you should be able to operate more at the edge,

517.44 -> and make decisions at the edge,

518.97 -> even when there is no internet connectivity,

521.46 -> as though you do it on the cloud,

522.96 -> so this is something that we are super excited about.

526.47 -> So this trend is gonna continue to evolve,

528.93 -> and it's gonna unlock whole bunch of efficiency

531.9 -> like it has never been seen before.

535.47 -> From the use case perspective and the outcomes,

539.07 -> there are five areas where our customers

542.07 -> are maximizing the benefits of IoT, and AI/ML, and cloud.

547.35 -> So the first area is

548.58 -> unlocking the higher operational efficiencies.

551.67 -> What we mean by that is essentially getting to know

555.69 -> not only the state, and also the reasons.

558.15 -> So that means remote health monitoring

560.43 -> like your OEE dashboards,

563.94 -> those are all a major outcomes that we are seeing.

566.61 -> In addition to that,

567.81 -> if you remember I talked about

569.04 -> like 90% of the data being On-prem.

571.47 -> So this is a great area for people

573.63 -> to move that data into the cloud,

575.7 -> and then not only get insights at a factory,

579.03 -> but also across the factories, right?

581.1 -> So that's kinda

582.14 -> where we almost call this historian and historian plus.

585.9 -> So there are many things that you can do

588.24 -> once you have the data in the cloud,

590.13 -> and contextualize it with other forms of data,

592.32 -> MES, your CMMI, your metrology data,

596.07 -> so it opens up a whole pandora of box

598.77 -> to uncover efficiencies like it has never been before.

602.97 -> And then we talked about the downtime,

604.65 -> so this is kinda where having a good handle

608.37 -> on the conditions of the assets,

610.95 -> and then going through the journey

612.75 -> of like getting descriptive, predictive,

615.36 -> and then prescriptive insights on what's the optimal level

618.72 -> to operate your factories and machines,

621.24 -> and the other equipment, so that's one area we are seeing,

624.03 -> and also with the advent of digital twin,

627.39 -> more and more operational twins are being produced

630.63 -> so that way you can remotely understand exactly

633.84 -> where the problem is,

635.37 -> and also what you can do about them, right?

638.28 -> So quality is another area

639.73 -> where we are seeing customers are deriving lot of benefits,

644.76 -> especially with the advent of computer vision,

647.31 -> and when IoT meets AI scenarios.

650.64 -> So there is a lot of things that we could do,

652.92 -> that could never been done before

654.6 -> in terms of identifying the anomalies,

657.15 -> clustering good versus bad,

659.07 -> and really understanding why things work one way,

662.13 -> and why it doesn't,

663.06 -> and comparing across different machines,

665.85 -> factories and enterprises, right?

668.43 -> To reduce all the scrap and other unwanted cost.

672.84 -> Fourth area is around the energy usage, sustainability,

677.64 -> getting a good handle on the energy,

679.95 -> and then metering it,

681.18 -> and then getting really efficient in all forms of operation,

686.25 -> in your equipment, in your buildings, everywhere, right?

689.61 -> Wherever you're using the energy,

692.07 -> so that's a key area, so where we are seeing,

694.95 -> and also an impact of environmental factors

697.92 -> on your energy usage,

699.3 -> that's another critical area

700.78 -> where we can input weather reports into your AI/ML models,

704.79 -> and unlock more benefits.

707.1 -> Last area manufacturing and supply chain

709.38 -> are like born twins, one cannot exist without the other.

712.71 -> Any disruption that happens in the supply chain

716.46 -> invariably affects the manufacturing.

718.92 -> So this is kinda where we are seeing a lot of our customers

721.47 -> at our track and trace solutions, control towers,

724.5 -> we announced a solution service yesterday, AWS supply chain.

729.51 -> So that's kinda where our focus is,

732.06 -> and our customers are benefiting from some of our services.

737.61 -> Now, just to kind of give you a preview on

741.09 -> what are some of the IoT and AI/ML services that we have,

744.36 -> that helps you with delivering all these outcomes.

747.42 -> So we classify our IoT services into three broad areas.

752.01 -> One is what you can do at the device level?

755.4 -> What is a cloud company doing at the device level,

758.19 -> you might ask, right?

759.18 -> But if you think about it,

760.98 -> getting a good handle on the device

763.11 -> helps you understand the state of that device,

765.27 -> and also helps you connect that device to the cloud

768.87 -> in a most efficient way,

770.28 -> and ensure that it is secure and safe at the device level.

774.33 -> Right?

775.595 -> And the second thing is around the connectivity and control.

778.8 -> This matters because this is kinda

780.63 -> where we can help our customer orchestrate a wide variety,

785.07 -> and wide number of devices, and also onboard them at scale.

789.18 -> At scale is a key word here, right?

792.93 -> Because that's kinda where you get the benefit of ensuring

796.41 -> that you keep an enterprise up and running

799.26 -> in a most efficient way.

800.76 -> Third area is around the analytics and streaming services.

804.09 -> This is where you take the data,

805.5 -> and you reason on top of that data,

807.24 -> and get insights

809.07 -> to help with all your operational decision support,

811.8 -> and also to orchestrate and automate where you can.

815.85 -> So what we are also seeing is,

817.68 -> these three are not layered.

819.75 -> As you can see,

820.583 -> I didn't put them on top of a pyramid or as a cake form,

823.35 -> I put them in a circle here,

825.03 -> so the reason being one feeds off the other.

827.67 -> So as you start leveraging, this is a scenario

830.928 -> where a whole is larger than the sum of parts,

834.3 -> so the benefit you get is,

835.62 -> far exceeds using them individually,

838.11 -> that's why we call this a virtuous cycle,

840.57 -> so it's a self-fulfilling cycle where more you use it,

843.51 -> more benefits you would get.

846.03 -> Just to quickly to double click on these things,

850.432 -> basically these are the services that you see,

853.74 -> we have four services in the device software areas

856.68 -> that helps you connect,

857.79 -> and secure your devices, and endpoints.

860.37 -> We have green grass

861.75 -> that can bring in some of the cloud capabilities to the edge

865.02 -> if you have stronger computers

868.02 -> at computation at the edge like PLCs.

871.86 -> And then on the connectivity and control services,

875.28 -> we have five services.

877.11 -> IoT core is by far our most popular service we have,

882.09 -> last I checked was 1.6 billion endpoint,

887.111 -> and we get about 2.2 under quarter connections

891.81 -> per minute coming into us.

893.88 -> So this is huge,

894.9 -> and this can clearly shown and proven in the industry

897.6 -> for several years now,

899.88 -> that it can scale and handle at the capacity

902.82 -> that is requested by our customers.

905.55 -> And then we also have fleetwise

907.5 -> to help with large, managing lot number of fleets,

910.5 -> particularly automotive,

912.15 -> and the RoboRunner in terms of orchestrating robots

915.81 -> in a distribution center or manufacturer,

918.39 -> any other place that you want,

919.77 -> kind of orchestrate your robots, right?

922.56 -> Moving on to analytics and streaming services,

926.07 -> so this is kinda where we have five services,

930.18 -> IoT events you can customize,

932.1 -> you can write rules and make actions on certain events,

936.03 -> whatever events happens that you gather

938.73 -> from your sensors or IoT devices, right?

942.21 -> And then we have SiteWise,

944.13 -> which is an industrial grade software service

947.46 -> that is built to collect, connect, contextualize data.

951.93 -> You can build asset hierarchies,

954.09 -> and you can also store time series data,

956.73 -> and make decisions on the remote health of your equipment,

960.51 -> and you can also build other dashboards

962.49 -> to get additional insights into your operations.

967.38 -> TwinMaker is one of our newest service

970.5 -> that essentially helps you build operational digital twin.

973.83 -> We also have partners such as Matterport,

976.11 -> who can give you a 3D visualization of your operations,

980.16 -> and you can deliver

982.47 -> some very interesting use cases and outcomes

986.01 -> by leveraging digital twin.

989.58 -> And we also to compliment IoT,

991.647 -> IoT and AI/ML go hand in hand,

993.42 -> almost one doesn't exist without the other.

995.67 -> We also have, I'm gonna touch on four services here.

999.81 -> One is look out for vision,

1002.12 -> so there is certainly a demos in the industrial tent,

1005.63 -> you guys can go check it out,

1007.25 -> where you can train AI/ML models

1009.47 -> to look for certain anomalies,

1011.36 -> either in your parts or in your operations.

1014.63 -> You can train them to essentially give you early warnings,

1018.71 -> whether it's a quality or any defect,

1022.04 -> you can automate them as well, right?

1023.81 -> That's another thing that you would notice.

1025.82 -> You also have look out for equipment,

1028.49 -> which can help you with condition based monitoring.

1031.61 -> Look for root causes of the disruptions in your operation.

1036.41 -> And we also have monitron, monitron is very easy to use,

1039.707 -> you can even buy it from amazon.com.

1042.5 -> It comes with both sensor and the insights,

1045.92 -> the dashboard that you can, and also gateway.

1048.68 -> It can detect your vibrations, your temperature,

1052.76 -> and then even if you don't have a PLC,

1054.917 -> you can literally stick it on any rotating equipment,

1057.59 -> and get insights on the inner workings,

1060.53 -> or the health of that equipment.

1062.39 -> Certainly you can use it in wide variety of use cases.

1065.42 -> And lastly, we also have AWS Panorama,

1068.84 -> which is basically a computer vision base.

1070.88 -> It can turn any cameras into more of an intelligent camera,

1075.44 -> and you can use this in wide variety of users,

1077.78 -> whether it's in a security, safety, worker safety situation,

1083.121 -> or essentially as you are seeing in the thing,

1085.638 -> managing the fleet of trucks,

1088.16 -> so there is wide variety of use cases

1090.08 -> wherever you want to turn your cameras

1092.27 -> into more intelligent cameras,

1093.89 -> and make decisions based on that.

1096.256 -> And then I won't drain you into the details here,

1100.49 -> I showed you the list of customers.

1103.01 -> Here is all the things that we hear back from our customers

1105.62 -> on how they are benefiting from our solutions,

1108.89 -> and our partner solutions as well, right?

1111.23 -> Many of these services gets embedded into our partners,

1115.82 -> and then essentially make them even better.

1118.04 -> As you can see, there is a wide variety of benefits

1120.77 -> in wide number of use cases.

1123.32 -> So certainly we have wide variety of partners

1127.22 -> who have taken these things,

1128.42 -> and then deliver value at even higher.

1130.73 -> In fact, we have one here, Hitachi Vantara,

1134.27 -> who have already taken their existing industry

1138.29 -> ready solution,

1139.123 -> which has been in existence for very many years,

1141.92 -> and embedded our services and made them even better.

1147.23 -> So with that, I would invite Rajesh Devnani from Hitachi

1151.16 -> to come in and tell us a story about

1153.086 -> what they're doing with our services,

1155.18 -> and how they're benefiting.

1156.17 -> - Thanks a lot, Praveen.

1157.595 -> (crowd applauds)

1158.915 -> Thank you.

1160.667 -> Hi, good afternoon all.

1161.84 -> This is Rajesh Devnani from Hitachi Vantara,

1164.21 -> and Hitachi Vantara is the global solutions services,

1168.23 -> and digital infrastructure arm of the Hitachi group

1170.72 -> from a global standpoint.

1172.55 -> So Praveen, the last slide Praveen presented

1175.1 -> made me recall and remember a blog post

1177.38 -> I wrote about five years back,

1178.76 -> which was caught putting the IoT cart before the horse.

1182.24 -> Things have moved on from there,

1183.83 -> I think it used to be all about technology at that,

1186.8 -> and all the bells and whistles of IoT.

1188.66 -> The focus is really pivoted towards the business benefits,

1191.207 -> and the value of that,

1192.68 -> and that's a sea change from what we saw before.

1195.32 -> So look at it past now in the current context, five years,

1200.03 -> a lot of churn has happened,

1201.53 -> a lot of IoT platforms came and went.

1204.17 -> And only the ones

1205.003 -> that are focused on the real business value

1206.57 -> have stood the test of time really effectively,

1208.82 -> and AWS obviously clearly lead charge of that.

1211.61 -> What I'll cover in the next couple of minutes

1213.29 -> is I'm gonna give you a brief walkthrough

1215.39 -> in terms of Hitachi's credentials,

1218.57 -> and vision strategy in the IoT space,

1221.81 -> and how we are aligning with the AW,

1223.76 -> who's again the market leader in this space,

1226.16 -> but we are saving the best for the last.

1228.08 -> So we'll have Rohit come in after me,

1229.61 -> and talk about how we are really realizing

1231.35 -> that in a customer context,

1232.85 -> and working in a collaborative journey with them

1235.52 -> to make it all happen.

1239.36 -> So just to get started, a bit of a brief on Hitachi.

1243.11 -> I'm not sure how many of you know about Hitachi,

1245.48 -> but we started life about 112 years back,

1249.02 -> in a small workshop in the prefixture of Hitachi,

1252.41 -> and that's where the name comes from basically.

1254.84 -> And from those humble beginnings,

1256.64 -> we have really come a long way,

1257.81 -> so it's about $85 billion in terms of size now.

1260.87 -> But what has really stood constant over all these years

1263.51 -> is the spirit of innovation

1264.8 -> that paraded us right from the start,

1266.227 -> and it was all about creating a positive societal impact,

1269.9 -> so that's really stood the test of time,

1271.61 -> as we have grown in Hitachi.

1273.05 -> Hitachi has grown to become industrial giants,

1275.27 -> so we do have real credentials in the industrial IoT space.

1278.72 -> We operate about 200 manufacturing plants globally

1281.33 -> across the board.

1282.74 -> We do work on all the critical infrastructure sectors,

1286.52 -> and we do provide very complex high value assets,

1290.09 -> not only do we design them,

1291.5 -> build them, manage them, monitor them.

1293.6 -> So IoT has been kind of in our roots,

1295.37 -> the whole industrial thing

1296.3 -> has been in our roots essentially.

1297.86 -> And we also have a digital practice,

1300.38 -> which has been in existence for about the past 60 years,

1303.32 -> the IT side of things.

1305.06 -> So what really accords us,

1306.68 -> is gives us a very unique position

1308.39 -> in terms of understanding both sides of the world,

1310.37 -> and bringing them together.

1311.78 -> So it used to be like ITs IT, and OTs OT,

1314.12 -> and now the twin shall meet.

1315.89 -> Hitachi as an organization brings that capability

1318.17 -> in terms of merging both these capabilities,

1319.91 -> and competencies together,

1321.08 -> and offering the best out of that.

1323.15 -> Now we do have a very intense focus on R&D,

1326 -> and one of the key streams for us in the R&D space

1328.52 -> has been around industrial AI,

1330.32 -> and that's very pertinent to the topic

1332.12 -> that we have today on hand.

1333.59 -> And it's about really

1334.423 -> how do you apply industrial AI at scale

1336.98 -> in terms of delivering transformative goals

1339.59 -> for our customers

1340.423 -> in helping them really derive the true value from IoT,

1343.19 -> whether it be in terms of the cost optimization,

1345.77 -> or in terms of their revenue augmentation,

1347.57 -> or even in getting into new business models

1349.52 -> entirely altogether.

1353.72 -> So Hitachi, we started our own foray into

1357.2 -> what came to be known as IoT a couple of years back,

1360.02 -> and out of that we conceptualized,

1361.91 -> and brought to live this concept called lumada,

1364.877 -> and lumada essentially stands for,

1366.8 -> it's an acronym which stands for illuminate data,

1369.2 -> which is essentially shining a spotlight on data,

1372.26 -> and producing rich insights of that data.

1374.39 -> So that came into existence as a basic core platform

1378.14 -> for IoT from Hitachi,

1379.542 -> but essentially it's a compendium of solutions,

1382.28 -> services, products all bundled together,

1384.44 -> so it's an overarching framework of sorts

1386.42 -> that we have to the marketplace,

1388.28 -> and does have components like data operations,

1390.8 -> and data governance, data management catalog,

1393.02 -> and there's a bunch of things in there.

1394.61 -> But what essentially differentiates lumada

1396.53 -> in the current context

1397.52 -> is all the industrial IoT solution course

1399.92 -> that we have really built on top of that.

1401.96 -> And when Praveen alluded to that,

1403.55 -> that's all what we are now kind of re-platforming,

1406.76 -> putting on the AWS stack, and bringing it together

1410.3 -> the power of Hitachi's industrial expertise

1412.61 -> in knowledge

1413.51 -> along with the cloud platform capabilities of AWS,

1416.48 -> to bring to our customers that innovation essentially.

1419.87 -> So industry clouds is the next step really

1422.57 -> in that evolution as we move forward,

1424.73 -> and they are gonna be defining,

1427.01 -> what's the future journey of the cloud platforms?

1429.95 -> So till date,

1430.783 -> the cloud platforms have been largely horizontal,

1433.04 -> and agnostic of the industry verticals,

1434.962 -> but there's a accentuated need for these cloud platforms

1438.11 -> to really start offering

1439.76 -> industry vertical specific functionality.

1441.92 -> And that's where people like us play in,

1444.17 -> that's where Hitachi really plays into the game.

1446.39 -> So we do bring a lot of rich competencies

1448.76 -> across multiple verticals, speed, energy,

1450.83 -> be it mobility, be it the core manufacturing sector itself.

1454.25 -> And we have 200 plus use cases,

1456.32 -> a solution course that we have developed

1458 -> to address very specific needs

1459.5 -> across a range of functionalities like production,

1463.58 -> maintenance, quality, safety,

1466.01 -> so the whole plethora of use cases that we have in place.

1469.13 -> And that's the journey we are embarking on with AWS,

1471.5 -> to move that together in that direction.

1479.582 -> So logical segue from there,

1481.22 -> AWS and Hitachi have been partners for long.

1483.5 -> It's a very strategic partnership

1484.85 -> that we have in place with AWS.

1487.16 -> It's a very multifaceted partnership,

1488.96 -> we are onto about 14 competencies,

1490.97 -> we have 300 plus certifications,

1492.83 -> we have 50 plus clients

1494.24 -> that we jointly engage with together on.

1496.49 -> I think the core of it

1497.38 -> is really around the whole industrial space

1499.34 -> because that's off very key importance board to AWS and us,

1502.49 -> in terms of creating the right business impact

1504.38 -> for our customers,

1505.73 -> so we are on this journey together like Praveen said,

1507.95 -> I'll talk about one of the solutions

1509.39 -> that gone onto the marketplace very recently,

1511.76 -> but even before that,

1512.72 -> we are sort of re-platforming

1514.4 -> a few of our very key strategic core companies,

1516.62 -> so there's asset performance management product

1518.6 -> that we have in our fall

1520.04 -> that's being re-platformed entirely in AWS

1521.9 -> to take care,

1523.367 -> take use of the core cloud native capabilities that exist,

1528.08 -> and ensure that the whole solution

1529.49 -> becomes a lot repeatable and scalable.

1531.83 -> And we can offer those to customers

1533.3 -> at a very logical price point as well.

1540.95 -> So this is one of the solutions,

1542.42 -> this is essentially a manufacturing insight solution,

1545.21 -> which is deployed at the AWS platform,

1547.82 -> and is also available on the AWS marketplace.

1551 -> And this is really addressing the core needs

1553.04 -> of the whole manufacturing segment at large.

1556.01 -> It covers the entire gamut of use cases,

1558.92 -> right from basic descriptive analytics around the forum,

1561.98 -> all the way up to the predictive prescriptive analytics,

1564.23 -> and the complete manufacturing lifecycle optimization

1566.936 -> value chain.

1568.4 -> And it does also offer a rich set of metrics out of the box,

1572.21 -> so Praveen spoke about OE as an example.

1574.501 -> We do OE, we also do things like on time,

1576.29 -> and full defect rates, first pass yields,

1580.64 -> the health and safety indicators.

1582.77 -> There's a whole bunch of things

1583.91 -> that we cover in the gamut of that,

1585.29 -> even production scheduling optimization,

1587 -> so it's a complete end-to-end platform,

1590.03 -> which addresses the need of

1591.53 -> core manufacturing industrial companies,

1593.6 -> and provides those capabilities out of the box.

1596.24 -> And this platform

1597.23 -> has been completely done natively on AWS now.

1600.59 -> So we are leveraging all the underlying components of AWS

1603.62 -> in this product.

1604.79 -> And this is sort of a reference architecture,

1606.89 -> or a blueprint that we have created

1608.42 -> so that we can achieve a lot of repeatability,

1610.61 -> and scalability in the entire process

1612.29 -> of doing this with customers.

1616.1 -> So before I wrap up and pass it to Praveen

1620.24 -> to invite Rohit upstage,

1622.25 -> I like to take a minute

1623.21 -> to talk about our collaboration with Penske.

1625.61 -> So Penske and Hitachi

1626.87 -> have been very close strategic partners

1628.88 -> for the past five play years.

1630.74 -> We kind of conceptualized and began this journey together,

1633.71 -> embarked on it with the clear intent

1635.78 -> of identifying real business values cases

1639.08 -> that can be impacted by leverage of IoT,

1641.36 -> and artificial intelligence, and machine learning.

1644.18 -> And out of that born was a particular use case,

1647 -> because Penske operates very large fleets,

1649.25 -> hundreds of thousands of fleets,

1651.08 -> so maintenance is really a very big cost factor for Penske,

1654.65 -> and this is what we set out to address.

1656.57 -> Rohit is gonna talk more about it,

1658.097 -> but it's been a very enriching journey for us

1659.663 -> over the past couple of years.

1661.64 -> It's not been easy by any means,

1663.65 -> it has been a journey, that's why I said,

1666.23 -> and there's no magic one per se,

1667.64 -> that you bring technology and things happen.

1670.1 -> It's been arduous, it's been long,

1671.51 -> but it's been very rewarding and enriching as well

1673.46 -> as we've gone along that.

1674.72 -> So with that,

1675.553 -> I'll call upon the Praveen back to invite Rohit upstage.

1678.68 -> Thank you.

1679.773 -> (crowd applauds)

1685.94 -> - I'm super excited to invite Rohit,

1689.54 -> and talk about what he has been able to do

1693.2 -> by leveraging some of our services,

1694.97 -> and our partner, as you know,

1696.74 -> computer vision and predictive maintenance,

1699.35 -> people have been talking about it for last 10 years, right?

1701.99 -> But what they have done is truly remarkable,

1705.32 -> and the complexity that they had to deal with,

1707.75 -> and to deliver at the scale

1710.39 -> that they have done is something,

1712.634 -> I'm super excited that we are so fortunate

1716.06 -> to have Rohit here to share the story,

1718.46 -> over to your right?

1720.308 -> (crowds claps)

1723.17 -> - All right.

1724.88 -> Thank you Praveen for the introduction,

1726.89 -> and Rajesh, thank you for that comment

1729.92 -> about best for the last,

1730.94 -> that does not put any pressure on me, obviously, right?

1734.985 -> So show of hands, how many of you know about Penske?

1738.26 -> Wow! More than I expected.

1739.97 -> How many of you know what we do?

1743.21 -> Okay, that's great.

1744.8 -> So let me introduce Penske Transportation Solutions to you,

1749.75 -> right?

1750.776 -> So that I represent Penske Transportation Solutions.

1754.43 -> We are part of the bigger Penske Corp,

1757.85 -> which is a 37 billion organization.

1761 -> Now, most of you may know about Penske

1763.55 -> because of race cars, because of indie car,

1766.64 -> but we are part of Penske Transportation,

1769.04 -> which is again, is a privately held organization

1771.89 -> part of Penske Corporations.

1775.19 -> We were founded in 1969, so right before Covid in 2019,

1780.56 -> we celebrated 50 years of being in business,

1785.21 -> 3000 locations globally,

1787.19 -> majority of these locations are in North America,

1791.78 -> and we have about 40,000 associates.

1794.96 -> This is again, Penske transportation only.

1797.99 -> What we do, we do full service leasing,

1800.6 -> we do truck rental,

1802.07 -> we do contract maintenance and logistics.

1804.35 -> So let's talk about that a little bit, right?

1805.73 -> So how many of you have seen a yellow Penske truck?

1809.93 -> Okay, so that's our rental fleet.

1812.15 -> So folks like you and me who wanna move, right?

1814.88 -> We would rent a Penske consumer truck,

1818.42 -> which are those yellow trucks.

1821.45 -> We also, a lot of large organizations,

1824.889 -> UPS, FedEx, Walmart, during the busy season,

1827.66 -> which we are in right now,

1829.19 -> where a lot of packages have to move

1831.35 -> because we all order on the internet,

1835.1 -> so we do commercial rental as well.

1837.65 -> So those companies actually rent trucks from us,

1841.49 -> so that's a B2C, B2B model to deliver those goods.

1847.76 -> So again, in our rental product line,

1849.86 -> we do both B2B, and we do B2C in our full service leasing,

1856.46 -> which is a whole different model,

1858.53 -> and it is our bread and butter.

1859.85 -> What we do is lease trucks to our customers.

1864.26 -> Again, an example of this could be where Walmart,

1867.35 -> Lowe's, Costco, some of these large organizations,

1871.01 -> their core business is not transportation, right?

1873.92 -> Their core business is retail.

1876.44 -> They want someone else to focus on transportation

1879.14 -> because that business is becoming very complicated.

1882.32 -> So they come to us, we work with them,

1884.66 -> we spec the entire unit,

1887.39 -> and the specking is based on a number of factors.

1890.21 -> It could be the type of payload that they wanna run,

1893.09 -> where they wanna run, how many miles they wanna run,

1895.7 -> and then we consult to them what type of tractor,

1901.22 -> or truck they should buy,

1902.9 -> and then we buy it for them, right?

1904.46 -> So we buy it, we place the order with the OEMs,

1907.28 -> and then we do full service leasing as it's called,

1910.25 -> which means that we lease it to them,

1911.87 -> but we take care of everything.

1913.49 -> We take care of licensing,

1914.93 -> we take care of titling, we take care of taxation.

1918.14 -> If you know anything about trucking,

1920.09 -> taxation goes down to the county level, right?

1922.79 -> It's very complicated.

1923.87 -> But we take care of all of that,

1925.46 -> and we take care of maintenance services.

1927.89 -> Maintenance is, again, very complicated business,

1929.9 -> I'll talk about that,

1931.04 -> but we take care of all of that.

1932.36 -> So the only thing that the customer

1934.135 -> that leases from us has to do,

1936.47 -> is put a driver in there, and fuel the truck.

1938.96 -> That's all they have to do.

1939.83 -> We take care of everything else.

1941.24 -> That's a large area for business,

1943.16 -> and we are leaders in this space,

1945.92 -> and then of course, logistics.

1947.06 -> So most of us are used to drinking Starbucks in the morning.

1951.23 -> Some of us in the northeast also go to Wawa,

1954.8 -> so the supply chain of all that,

1956.63 -> most of that is managed by Penske Logistics,

1959.3 -> which is our logistics organization.

1962.39 -> And then we maintain and manage over 400,000 trucks,

1967.25 -> tractors, and trailers.

1968.78 -> So here's an assignment for you, right?

1971.66 -> The next time you're on the road,

1973.61 -> count the number of Penske trucks,

1975.23 -> yellow trucks that you see.

1976.19 -> And trust me, you'll see, now start seeing a lot more

1978.17 -> now that I've mentioned it, right?

1980.043 -> Now, the only way you would know that it's a lease truck,

1983.36 -> I talked about the rental truck, easy to spot,

1986.03 -> lease truck is, if you look on the driver's side,

1989.06 -> there will be a small decal, which says Penske.

1992.03 -> Right?

1992.863 -> So you get extra credit

1993.92 -> if you're able to see a Penske lease truck.

1998.42 -> This was a game we used to play with our younger kids,

2001.6 -> much better game than, are we there yet type of a game?

2003.94 -> Right?

2004.996 -> So if you have younger kids,

2005.829 -> this is a good game to play.

2008.11 -> So, okay, like I mentioned,

2011.02 -> we are leaders in truck maintenance, truck leasing,

2015.13 -> and this is our maintenance vision,

2017.68 -> and it's pretty simple, right?

2019.42 -> It is to increase the customer uptime,

2022.81 -> and to reduce the operating cost.

2024.7 -> Some of the large organizations that I mentioned,

2027.76 -> they have their own SLAs with their customers,

2030.817 -> and they rely on us

2032.523 -> to make sure that that truck keeps on running.

2035.68 -> So the increasing customer uptime,

2038.74 -> and reducing operating costs are not mutually exclusive.

2042.34 -> Think about it, right?

2043.173 -> So if a truck breaks down,

2046.78 -> the driver is stranded on the side of the road, right?

2049.42 -> We still have to figure out how to get that truck,

2052.27 -> and that customer up and running.

2054.52 -> So we have to either give them a substitute unit,

2057.34 -> or we have to send someone

2058.72 -> to make sure that the truck is repaired,

2060.76 -> either case, it's a higher operating cost for us.

2063.43 -> And like I mentioned, as part of our full service lease,

2066.34 -> we do all the maintenance,

2067.96 -> so any higher cost hits our bottom line.

2070.24 -> So we have to make sure that the customer is up and running,

2073.87 -> and our operating cost is low.

2076.45 -> Now, there is a challenge to this vision,

2080.02 -> and I would say it's a challenge

2081.34 -> to the entire transportation industry.

2083.74 -> Actually, there are two challenges.

2085.57 -> One is unavailability of skilled labor,

2088.81 -> in this case, technicians, right?

2090.97 -> And then two are drivers.

2092.53 -> These are the two biggest challenges

2094.36 -> that our industry faces today.

2098.14 -> And then also what happens is

2100.24 -> a lot of folks don't go to school to be mechanics.

2103.96 -> We are seeing that educational side

2107.35 -> is not being entertained a whole lot

2110.17 -> by the newer generation,

2111.52 -> so that's again, a problem.

2113.23 -> The modern trucks are becoming very complicated.

2116.29 -> Think about modern trucks as data center on the road.

2119.35 -> It's truly a data center which runs on wheels down the road.

2122.62 -> It's becoming very complicated.

2124.6 -> It takes a very long time to diagnose a problem.

2129.01 -> And repairing a problem is a specialized skill,

2132.43 -> which is not readily available.

2134.71 -> So what I'm gonna talk today is two use cases

2139.75 -> where we leverage technology,

2142.51 -> and these are in production

2143.89 -> where we leverage technology, AI, ML, IoT,

2148.15 -> to help further our maintenance vision,

2152.56 -> and to kind of alleviate a bit this challenge

2156.88 -> around finding skill set technicians.

2163.304 -> So in 2017, 2018, we partnered with Hitachi,

2167.59 -> and Rajesh mentioned that,

2171.503 -> to kinda talk about this first use case,

2173.92 -> which is around reducing the diagnostic time,

2178.09 -> and also reducing the repair time, and repeat repairs,

2182.53 -> and I'll talk about repeat repairs, right?

2184.387 -> And this happened to me personally,

2185.95 -> and I don't know if it has happened to you.

2188.26 -> So my car had a check engine light,

2191.47 -> and I called the dealership, and so first of all,

2195.22 -> I get an appointment two weeks after I call.

2198.07 -> So I take the car there,

2199.03 -> they keep the car in there for two days,

2201.52 -> and then after that they give me a call,

2203.05 -> and they say that your car's good,

2204.85 -> take my car out, bring it back in three days again,

2207.58 -> because the check engine light was on.

2208.81 -> So what the technician did in this case,

2211.39 -> he or she was not able to figure out the problem,

2213.46 -> so they just did what they needed to do

2215.65 -> to flip the switch for the check engine light.

2217.87 -> So that's a repeat repair

2219.22 -> because I'm coming back right to the dealership

2222.67 -> to solve that problem because my problem is not solved.

2225.79 -> So our goal,

2227.59 -> so think about this in the context of trucks, right?

2229.78 -> So if a truck is broken down,

2232.03 -> and the technician only fixes the symptom,

2234.49 -> but not the actual problem,

2236.05 -> that truck is gonna come back,

2237.16 -> that's more downtime for the customer,

2239.89 -> that's higher operating costs for us,

2241.48 -> which goes completely against our vision.

2243.7 -> So our goal with this use case

2246.61 -> was to reduce the diagnostic and repair time,

2250.821 -> and also to reduce repeat repair.

2253.39 -> So the solution that we built, and we call it guided repair,

2256.45 -> is an AI solution.

2257.83 -> It's an AI/ML solution

2259.75 -> which guides the technician through the repair,

2261.85 -> and I'll double click on this in a bit,

2263.92 -> but it guides the technician through the repair,

2266.71 -> and it kinda tells the technician,

2268.51 -> Hey, here's what you need to do to solve this.

2272.71 -> The outcome,

2273.64 -> we were able to reduce the average repair time

2276.4 -> by 15 minutes per truck.

2279.13 -> Think about 400,000 trucks that we manage and maintain,

2282.55 -> extrapolate this to that, imagine the scale,

2285.76 -> right?

2286.775 -> Imagine the amount of capacity

2287.608 -> we've been able to create within our shops

2291.1 -> because of this technology that we deployed.

2293.92 -> And again, like Rajesh said, it was not easy,

2296.68 -> not an easy problem to solve.

2298.57 -> The other problem that we have to deal with

2301.54 -> is change management,

2302.59 -> because now you're telling the technicians

2304.3 -> that hey, listen to the technology, right?

2307.48 -> Do what the technology tells you to do,

2309.07 -> so that's a change management problem.

2310.57 -> So the whole cycle was not easy,

2313.57 -> but trust me, very rewarding.

2315.22 -> So let's double click on this a little bit.

2318.19 -> So a Penske truck pulls into one of our locations,

2322.24 -> and the technician checks in the truck using,

2325.78 -> so the truck pulls in for a repair,

2327.25 -> the technician checks in the truck

2329.05 -> using our maintenance applications and systems.

2333.52 -> And as soon as the truck is checked in,

2335.62 -> we do know the repair history, the diagnostic history,

2339.25 -> the fault code history, we know what type of vehicle it is,

2343 -> the make model year,

2344.32 -> and we also know what fault the customer complaint is

2348.79 -> because that's what kinda started the repair.

2351.37 -> So what we do is, with all that information,

2353.92 -> we transfer the fault code,

2356.41 -> and the complaint through API calls

2359.08 -> to this guided repair model, the AI/ML model,

2363.64 -> and the AI/ML model is already trained

2366.04 -> with 20 years of maintenance data.

2368.23 -> We have a lot of data, and this model,

2372.82 -> right when we built it,

2373.78 -> was trained for 20 years of maintenance data.

2376.12 -> It has identified all the patterns and it constantly learns,

2379.45 -> right?

2380.283 -> So as newer models come in, as new data comes in,

2384.64 -> it constantly learns,

2386.26 -> the guided repair model then outputs a prescription,

2391.39 -> so both Praveen and Rajesh talked about prescriptive, right?

2395.44 -> So this is a prescription like a doctor,

2397.78 -> it's giving out a prescription,

2399.79 -> and the prescription includes

2402.4 -> what the technician needs to do,

2404.14 -> and the probability of success,

2406.54 -> so it actually includes that, okay?

2409.93 -> Now the technician can take two routes.

2412.3 -> One is they can go with the prescription,

2415.57 -> or they can still continue with what they think

2417.91 -> is the right repair.

2419.35 -> In both those cases,

2420.34 -> we capture right what the technician is doing,

2423.37 -> and here's why we do that.

2424.42 -> So because if they take the prescription out of the model,

2428.41 -> and do exactly what the model says,

2430.69 -> but the truck comes back for a repeat repair

2432.97 -> in X number of days, we wanna know that, right?

2435.31 -> Because then we need to adjust the model

2437.05 -> because the model is not putting the right prescription out.

2441.19 -> The second scenario could be

2442.51 -> that the technician completely ignores the prescription,

2445.66 -> and does what they think is the right thing to do,

2449.29 -> and the truck still comes back for repeat repair.

2451.24 -> In both of those cases,

2452.41 -> we want to know what action did the technician take.

2455.65 -> So we run a lot of analytics in the background

2458.157 -> to figure all that out,

2459.717 -> to do training for our technicians,

2462.04 -> and also training for the model,

2463.54 -> so it's a whole feedback loop

2465.97 -> that we have with this use case.

2469.12 -> This did go into production by the way in 2019,

2471.97 -> and trust me, this 2020 was covid,

2475.3 -> a lot of workers in the US retired,

2479.05 -> and we were impacted because of that as well,

2481.03 -> early retirement.

2482.14 -> This solution really helped us right in that case.

2485.86 -> So this has been in place since 2019,

2487.6 -> and we continue to improve on this.

2490.36 -> So this is a reactive use case, it's prescriptive,

2493.266 -> it's a reactive use case

2494.71 -> because the truck is already at a Penske location,

2497.98 -> the event already happened, right?

2502.68 -> So what we started talking about in early 2020,

2506.98 -> we started talking about combining our IoT data,

2510.67 -> so our trucks actually talk to us, right?

2512.47 -> They're sending a lot of data.

2513.834 -> So we talked about combining our IoT data

2517.445 -> with our operational repair data.

2521.77 -> And the goal was to reduce breakdowns,

2525.52 -> actually Praveen had that in one of his slides,

2528.34 -> which is kind of eliminating or reducing breakdowns,

2531.58 -> so that was our use case

2534.22 -> because again, in the first case, the breakdown did happen,

2537.01 -> came in for a repair, right?

2538.45 -> We wanted to kinda reduce the number of breakdowns

2541.12 -> so the truck is not stranded on the side of the road.

2545.2 -> So again, with through our partnership with Hitachi,

2547.96 -> we built an AI based proactive diagnostics model,

2554.77 -> which again combines the IoT data from the truck.

2557.65 -> This is realtime live streaming data

2560.8 -> with our operational data.

2563.14 -> And this model did go live earlier this year.

2567.55 -> It is in production,

2568.78 -> and in fact, this is a quote, I won't read it to you,

2572.05 -> but I'll tell you exactly what happened here.

2574.33 -> So the night that we deployed this,

2577.63 -> within a few hours the model identified a truck,

2580.483 -> which would fail in the next X number of miles.

2585.19 -> So the prediction actually is

2586.99 -> that the truck may fail in 100 miles, 200 miles,

2589.87 -> same day prediction, right?

2591.25 -> So it'll give out same day prediction,

2594.22 -> so this happened early morning,

2596.26 -> I think it's like 4:36 AM,

2599.17 -> and the model informed the Penske locations,

2603.31 -> they called the driver,

2605.26 -> the truck was pulled into the nearest Penske location,

2608.38 -> the repair was taken care,

2610.15 -> and the driver was on his way in three hours.

2613.69 -> Right?

2614.523 -> Now think if we did not have this capability,

2617.26 -> so what would've happened

2618.19 -> is the driver would still keep going, right?

2620.59 -> The truck would break, stranded on the side of the road,

2623.5 -> make a call to Penske 24/7 call center,

2627.34 -> we would then try to diagnose the problem,

2629.71 -> send a technician to where the driver is,

2634.3 -> try to figure out the problem,

2635.71 -> hopefully the technician has all the right parts,

2638.17 -> if not, we tow the truck.

2639.73 -> You get the point, right?

2640.95 -> It just takes a very long time, not three hours.

2644.2 -> So this has definitely reduced

2646.57 -> not only the number of breakdowns,

2648.97 -> but also the amount of time it takes to repair a problem.

2655.6 -> So let's double click on that, right?

2657.01 -> So like I mentioned,

2658.66 -> our trucks are talking to us through IoT.

2661.45 -> So this IoT data is live streamed into our,

2664.72 -> what we call is the Penske's data platform.

2666.61 -> It's a very simplistic image here,

2670.6 -> but there's a lot going on there because think about it,

2672.82 -> it's all streaming data,

2673.78 -> we gotta make sense of it,

2675.487 -> and we gotta do something with it,

2677.71 -> we gotta get insights out of it.

2679.54 -> Now this data is then consumed by the proactive model.

2684.88 -> And again, the proactive model is trained

2687.04 -> with operational data like the guided repair model is.

2689.83 -> So it combines the IoT data,

2691.72 -> and it combines the operational data,

2693.97 -> and then we apply business rules to it, right?

2696.19 -> So for AI decision making, you need two things.

2699.28 -> One is the prediction, and the other one is judgment.

2702.7 -> Because if you just go with the prediction,

2705.34 -> which is completely dependent on the data,

2706.84 -> and don't apply any context, it may not mean a whole lot.

2710.26 -> So I'll give you an example of that,

2711.82 -> so for example, the AI model may predict

2716.86 -> that a truck may fail in about a thousand miles.

2719.411 -> Right?

2721.801 -> Now, we know that that truck is due for service

2726.79 -> in about two days, right?

2728.2 -> So it's already there, it's already planned,

2729.85 -> it's due for service,

2731.11 -> and on an average, that truck runs 300 miles a day,

2735.4 -> so 300 miles a day, two days is 600 miles, the truck,

2738.25 -> the prediction is, it'll fail in thousand miles.

2740.5 -> So do we really need to pull the truck off the road,

2743.65 -> or should we let the driver continue?

2745.3 -> And then when the truck comes in for service,

2747.61 -> we can take care of the problem,

2749.8 -> so that's the context, right?

2752.02 -> So those are the business rules,

2753.61 -> those are the smarts, that's our secret sauce

2755.5 -> that we kind of apply to this.

2757.39 -> And then outcomes, the days and miles to failure.

2762.07 -> And then, like I said, the judgment,

2764.86 -> the context that we apply and the miles to failure,

2767.08 -> all of that flows

2768.909 -> into the Penske's internal and external applications,

2771.85 -> gets communicated to internal stakeholders,

2774.73 -> gets communicated to customers through our digital tools,

2779.17 -> so all of that happens.

2781.03 -> Now, the brains of this runs on AWS.

2784.27 -> Right?

2785.103 -> So the Penske's data platform,

2787.411 -> the AI model, the processing, all happens on AWS.

2792.49 -> So again, look, I have the privilege,

2794.98 -> and honor of speaking about this innovative work,

2797.92 -> but the actual work was a lot of collaboration

2800.83 -> between Penske, Hitachi, and AWS,

2804.91 -> those are the real heroes of this.

2806.2 -> And like I said, this is not an easy problem to solve,

2808.84 -> and we continue to enhance this.

2810.76 -> Our goal is to continue building on this proactive model,

2817.18 -> because again, right,

2818.214 -> we have to keep the model up and running.

2821.65 -> It has to learn as new truck models come out,

2825.82 -> and like I said, trucks are getting very complicated.

2828.76 -> And then we are also working on some other use cases

2834.005 -> that in the next year or two,

2836.47 -> there are some other use cases,

2837.82 -> which are going to again help our maintenance vision,

2841.9 -> and some of the other roadmaps

2844.27 -> that we have within the organization.

2847.39 -> That was my last slide,

2849.91 -> so hopefully this was meaningful for you,

2852.46 -> hopefully that you take away something from this.

2855.1 -> So with that, I'll invite Praveen back.

2859.274 -> (crowd applauds)

2863.23 -> - Thank you.

2865.3 -> Thank you Rohit, that was an awesome use case.

2868.87 -> So just to kind of recap

2871.51 -> what you have heard so far

2872.95 -> from Rohit, Rajesh and I,

2875.83 -> and also what is in it for you,

2877.9 -> why you should consider AWS, right?

2880.54 -> So there are four things

2881.59 -> that I wanna kind of quickly recap and highlight.

2886.247 -> One is, AWS is born out of Amazon,

2889.69 -> and we've been doing this for well over 16 years.

2895.509 -> So many of you might know,

2896.59 -> Amazon is also one of the largest industrial customer too.

2899.44 -> We make products, and we have factories,

2901.96 -> and we use all of our services

2904.69 -> that we recommend to our customers as well.

2907.289 -> So it's something that we leverage ourself and benefit from,

2911.53 -> so it's not foreign to us.

2914.77 -> Second thing is

2915.85 -> we have well over 200 purpose built services,

2919.54 -> which are serverless and as you saw here,

2922.57 -> 14 IoT and also four AI/ML services,

2928.09 -> and also work seamlessly

2929.44 -> with other AI/ML services beyond industrial AI/ML

2933.22 -> like SageMaker and others, right?

2935.26 -> And they all are serverless,

2936.79 -> they work exactly the same way in terms of scalability,

2939.94 -> and you only pay for what you use,

2942.46 -> and you can scale things on demand.

2946.27 -> So that itself is a huge benefit.

2949.09 -> And the third thing is we have customers in every industry,

2954.31 -> every regulated environment

2956.17 -> including hospital, military, government,

2959.14 -> so certainly we take security very seriously of your data,

2963.34 -> so certainly we have industrial grade security,

2967.39 -> and compliance standard

2968.83 -> that meets all the regions,

2971.11 -> and all the regulatory bodies that are out there.

2975.48 -> And the fourth and important thing is,

2977.95 -> the scale and the speed at which we operate, right?

2981.43 -> AWS is present in close to 250 countries now,

2986.17 -> and we have more than 90 availability zone.

2989.29 -> What it means is once you prove out an idea,

2992.17 -> and now you wanna scale and scale it,

2994.42 -> and ensure that it also has the speed,

2996.73 -> and the bandwidth,

2998.53 -> so that's where AWS can really help you

3001.02 -> in terms of deploying things in seconds, if not minutes,

3006.6 -> really kind of take that benefit.

3008.58 -> And also if there is any change,

3010.02 -> change adoption, change management,

3011.91 -> you don't have to go through,

3013.86 -> like the way it takes once you prove it out,

3016.5 -> once you make a change, wait for another six months,

3018.36 -> so that's not the case anymore.

3019.77 -> It's almost near instantaneous.

3022.26 -> So it's a huge benefit, as you saw from Penske story,

3027.15 -> and many of our other customer service as well.

3029.76 -> And in ought to this,

3030.78 -> we also have a huge partner network

3032.61 -> in every geo that we operate,

3034.5 -> so that's something that you can take advantage of.

3038.58 -> Certainly I invite you

3039.72 -> to learn more about industrial solutions and IoT offerings

3044.19 -> by visiting the webpage here.

3046.35 -> And another thing that I would highly recommend

3048.24 -> is there is an AWS village, we have IoT kiosk,

3052.14 -> certainly go check it out.

3053.73 -> And then there is an industrial tent at caesar forum,

3057.84 -> so many of our services that you saw,

3059.88 -> we actually have a demo that you can kind of touch and feel,

3063.39 -> and kind of get a flavor for how it really works,

3066 -> and ask questions,

3066.99 -> and how it is applicable for your particular situation.

3069.96 -> So I certainly invite you to go check out the caesar forum,

3073.23 -> and I think today they close early,

3075.36 -> so they close at four o'clock.

3077.04 -> So the last couple of days they were open late.

3080.19 -> So certainly, there's still time.

3082.17 -> Go check it out,

3083.19 -> and always feel free to reach out to me, Rohit, and Rajesh,

3088.44 -> if you have any questions.

3089.91 -> We are here to help you, and also listen to your stories,

3093.21 -> and hopefully next year you can come in,

3095.79 -> and present with us,

3096.72 -> and tell your story like Penske did.

3098.82 -> Thank you all very much for attending the session today.

3101.502 -> (crowd applauds)

3102.335 -> Also wanna thank our speakers,

3104.1 -> Rohit and Rajesh for being so kind and coming,

3107.58 -> and tell their stories with us.

3110.49 -> You have a wonderful rest of the afternoon.

3113.16 -> Hope you all had a good re:Invent so far,

3116.22 -> and certainly reach out to us if you have any questions.

3119.31 -> Thank you.

Source: https://www.youtube.com/watch?v=gRnQnS9dGII