0:00

0:01

1 second

Sharon: A traffic jam is like privacy because

0:08

8 seconds

Mike: you have to stunt me on that one. Eh a traffic jam is,  Do you have an answer to this? You do, don't you? There's a It's a little

0:15

15 seconds

I don't know, Sharon. Why is a traffic jam

Sharon: I'm not telling you, You have to come up with it on your own. I'll let you think about it. How about that? 

Mike: Okay. Um 

Sharon: You think

0:23

23 seconds

about it while I drive us to the car wash. My next guest is Mike Branch from Geotab. Mike is VP of data and

0:32

32 seconds

analytics. Geotab is one of the greatest telematics companies ever. He also helped to launch an AI assistant for Geotab,

0:41

41 seconds

making fleet data a lot more accessible and transparent. I see him. Let me get him in the car. Hey, need a ride?I

0:48

48 seconds

Mike: I think I might. I think I might. 

Sharon: Come on in. Let's do this. 

Mike: Thank you. 

Sharon: What is Geotab? What do you guys do? 

Mike: We're a connected vehicle platform. So, uh, if

0:56

56 seconds

you want to know anything about your vehicle as a fleet, ever if you're harsh braking, if you're speeding, um, if there's a problem with the battery in

1:03

1 minute, 3 seconds

your car, all that kind of stuff, we connect up to the OBD port in your vehicle. That little plug that's usually the side of

I don't think you have a Geotab device in here. It doesn't look like it, but you should

Sharon: No I dont, . But, well,

1:13

1 minute, 13 seconds

well, after this episode, I may. Maybe I have now a connection. 

Mike: Yeah. Exa Exactly. And, and so we help fleets across the whole globe, giants, uh, Giants like 

1:20

1 minute, 20 seconds

UPS and PepsiCo all the way through to mom and pop shops. Um, you know, we have over 4.7 million connected vehicles

1:29

1 minute, 29 seconds

across the globe. So, we're managing all that data at scale, helping them uh drive down collisions, helping them reduce downtime, helping them reduce emissions. Transition to EV is a big thing

1:37

1 minute, 37 seconds

Sharon: Okay. So, Mike, I I know that you have a lot of really good information about your industry, about telematics,

1:48

1 minute, 48 seconds

about what you can do with this information. Like, give me the top secrets. Like, I know everyone just

1:55

1 minute, 55 seconds

wants to hear all the juicy juicy details. Tell us everything. 

Mike: Oh, you want to hear everything?

Sharon:  I want everything. 

2:02

2 minutes, 2 seconds

But but the good stuff. Like the juicy stuff. 

Mike: The juicy stuff. All right. Here we go. Okay. So, here.

Carwash Noise

2:18

2 minutes, 18 seconds

Sharon: That was amazing. And I'm really glad my viewers got to hear that directly from your mouth. Mike: Not too many people know this story, Sharon.

2:29

2 minutes, 29 seconds

Sharon: Wow. So, this segment is called Yay or Nay. Okay.

 So, is it smart safety or

2:37

2 minutes, 37 seconds

surveillance overkill? So, your telematic system alters your fitness app every time you visit a fast food drive-thru?

2:47

2 minutes, 47 seconds

Mike: hohoho Absolutely. Nay. Nay. 

Sharon : Really?

Mike: Yeah. 

Sharon: Why?

2:51

2 minutes, 51 seconds

Mike: Well, you know what? I wouldn't want my um telmatic system to know anything about my fitness. Those two worlds

2:57

2 minutes, 57 seconds

should not be intertwined. Uh unless 

Sharon: what if it helped your fitness?

Mike:  I mean,

3:03

3 minutes, 3 seconds

Sharon: it could it could be a good thing

Mike: It it could be a good thing, but you need a proper consent. You want that to happen. But I would say if 

Sharon: You're very responsible

3:10

3 minutes, 10 seconds

Mike: Absolutely. 

Sharon: Yes. I mean, you're you're in the business of being responsible.

3:14

3 minutes, 14 seconds

All right. Next one. Your parents get a notification every time you break hard.

3:18

3 minutes, 18 seconds

Even if you're 42 years old and paying off a mortgage, yay or nay? 

Mike: I would say yay as long as there’s consent. Like I you know

3:25

3 minutes, 25 seconds

what? Specifically, if it was for my kids uh and I had a device in the vehicle, I want to know that they're they're driving. Well, if and 

Sharon: What if

3:33

3 minutes, 33 seconds

they're 42 years old? 

Mike: They're 40 years old and and they consent, then fine. But I I can't imagine many 42 year olds

3:41

3 minutes, 41 seconds

consenting with that. 

Sharon: Exactly. 

All right. Your seat detects crinkling chip bags and asks if you prefer apple slices

3:49

3 minutes, 49 seconds

instead. Yay or nay? 

Mike: Uh, that's that's a nay. That's a huge invasion of privacy there. I think 

Sharon: really that you like

3:57

3 minutes, 57 seconds

Chips.

Mike:  But that it's detected the fact that I've got, you know, this these chips on the and and then I say, "Hey,

4:04

4 minutes, 4 seconds

you should you should have a fruit instead."

Sharon:  Yeah. That's a good thing. 

Mike: It is a good thing.

Sharon:  Maybe having fruit will will waken you up and you can drive

4:11

4 minutes, 11 seconds

better. 

Mike: Yeah. Right. I still think I still think they get a lot of nays here. I'm probably a little bit more responsible than you

Sharon:. I

4:19

4 minutes, 19 seconds

think you're too responsible. 

Your telematics logs every time you honk,

4:23

4 minutes, 23 seconds

rates it on justified or petty, and sends you a monthly summary. Yay or nay?

4:29

4 minutes, 29 seconds

Mike: I I kind of think yay to that. I think so. Um 

Sharon: I think so, too. Yeah. 

Mike: You could pick up some. There might be some aggressive behaviour there that is

4:38

4 minutes, 38 seconds

unwarranted, right? 

Sharon: That's right. And then you get a summary and you learn.

4:41

4 minutes, 41 seconds

Mike: Yeah. You get a summary. You learn from that as she comes into your app. Right.

4:44

4 minutes, 44 seconds

Sharon: Exactly. Last one. If you cut someone off or speed, your car sends an apology tweet on your behalf saying, "Sorry,

4:53

4 minutes, 53 seconds

that's on me. I'm working on being better." Yay or nay?

Mike:  Oh, yay. Yay. And uh it should uh maybe give them a little

5:02

5 minutes, 2 seconds

gift certificate to Tim Horton at the same time. Starbucks. Come on. 

Mike: Starbucks.

Sharon: Yeah. Uh well, again.

5:09

5 minutes, 9 seconds

Okay. So, first of all, congratulations.

5:12

5 minutes, 12 seconds

I know you are a, Geotab won the Picasso award about a year ago or so

Mike: We did we did, very excited

Sharon:. So,

5:18

5 minutes, 18 seconds

congratulations. Which means that you're obviously doing something really well with privacy. So, explain to me with the

5:25

5 minutes, 25 seconds

data that you're collecting, uh, what personal information are you collecting that you're even thinking about privacy?

5:31

5 minutes, 31 seconds

Mike: Uh, you know what, a lot of people don't think about that right away cuz they think, oh, you're, you know, your personal information is your credit card information, right? It's your healthcare

5:38

5 minutes, 38 seconds

information. Uh but uh your vehicle lays a bit of a track, right? So uh your

5:45

5 minutes, 45 seconds

vehicle driving habits if you're coming from home to work every single day,

5:48

5 minutes, 48 seconds

there's a pattern in that data and uh that pattern can divulge a little bit about you uh from a privacy perspective.

5:56

5 minutes, 56 seconds

So that is the geospatial element is the biggest concern for us when it comes to privacy. Um you know there's other pseudo identifiers like VIN as well too.

6:05

6 minutes, 5 seconds

Yeah. Where it's traveling. Um that's that's our biggest um risk area. 

Sharon: If you're a fleet company, isn't that the information that you want to collect?

6:14

6 minutes, 14 seconds

Mike: That's exactly it. And there in lies a conundrum, right? You you absolutely need that data to do your business. Um but you have to also give privacy

6:22

6 minutes, 22 seconds

measures to the fleet uh to allow them to turn off um uh GPS data whenever somebody's in say like a personal mode.

6:30

6 minutes, 30 seconds

For us, it's really important to uh to understand if we're dealing with uh data that might be personal or not. Um and

6:39

6 minutes, 39 seconds

you know, you have a driver that works for a company, they may take that vehicle home. You shouldn't be tracking the data that is in that kind of personal mode. Um and so as we're

6:47

6 minutes, 47 seconds

developing new data and insights for a lot of our customers, we can't be doing it based on a lot of this personal data.

6:53

6 minutes, 53 seconds

But to your point, absolutely our customers want to know where their vehicles are. It helps for routing,

6:59

6 minutes, 59 seconds

right? helps for uh a whole series of things. They couldn't run their business if they didn't have that GPS data. 

Sharon: Are you using that data for any other

7:07

7 minutes, 7 seconds

purpose or sharing that data for insights for other organizations or municipalities or anything like that?

7:14

7 minutes, 14 seconds

Mike: Yeah, we believe that you know there's certainly um a whole host of reasons that you can use this data for that can really benefit society. 

Um you know we

7:23

7 minutes, 23 seconds

recently did our platform Altitude which we take all this data privacy compliant and made it available um for municipalities to look at you know areas

7:32

7 minutes, 32 seconds

and cities where there's congestion and where you might look at better planning for uh for freight. Um we did a study with uh on the Gardener Expressway which

7:40

7 minutes, 40 seconds

as we're all familiar with you know there's three lanes in uh three lanes out and now construction has been done and you've got two lanes in two lanes

7:47

7 minutes, 47 seconds

That has a huge impact on uh on productivity in the whole city. And so as a result of some of the study we did

7:55

7 minutes, 55 seconds

to show that impact it was able to bring down the construction time. So I think another $73 million was put um into that

8:03

8 minutes, 3 seconds

project to bring down the time. But you can't do that without privacy compliant data. And that's why I always say like not all GPS data is is created equal.

8:11

8 minutes, 11 seconds

Sharon: Yeah. So what do you mean by that? 

Mike: So you could you could slam on your brakes at an intersection. A whole bunch of people do that. You want to be able to understand is that a dangerous

8:20

8 minutes, 20 seconds

intersection or not. Um and that is an okay use. You're not divulging private data at that point if it's happening from multiple vehicles in a common area.

8:30

8 minutes, 30 seconds

Dangerous driving. But you don't want to start divulging things like Mike drove from his home to the office every single day. 

There was that New York Times

8:39

8 minutes, 39 seconds

expose. It was, do you remember that? It was like one data set, zero trust. And so in that data set, what they exposed was individual vehicle driving patterns.

8:50

8 minutes, 50 seconds

Sharon: Okay. 

Mike: And he was able to very clearly see when somebody was um going to maybe change their job. They went from their home to Microsoft, home to Microsoft,

8:59

8 minutes, 59 seconds

then they went home to Amazon, then home to Microsoft. you could see that that pattern dulged information that it that

9:06

9 minutes, 6 seconds

it shouldn't. 

Sharon: So Mike, I understand you were instrumental in launching the um ACE platform, which is Geotab's AI

9:14

9 minutes, 14 seconds

assistant. 

Tell me about it.

Mike: Our our theory was if we launched ACE um that a lot of our fleet customers want to just be able to ask a question about their

9:23

9 minutes, 23 seconds

vehicles or their fleet cuz you're you have this data deluge, right? So you got these dashboards every which way. Um,

9:30

9 minutes, 30 seconds

you know, we're streaming 100 billion data points a day into our ecosystem.

9:34

9 minutes, 34 seconds

Sharon: God that’s crazy

Mike: a hundred  billion with a B and uh we have got 55,000 you know customers across so many different verticals. 

So to be able to

9:42

9 minutes, 42 seconds

create this oneizefits-all dashboard for everyone doesn't really make sense. So similar to chat GPT like can I ask a question about my fleet and have it give

9:50

9 minutes, 50 seconds

me the answer and that was the theory and and so when we ran some initial tests with customers they love this this idea right being able to ask you know

9:58

9 minutes, 58 seconds

who are my safest drivers you know um uh do I have a problem with you know any of my vehicles um batteries just anything

10:05

10 minutes, 5 seconds

you could think about for your fleet ask it 

Sharon: I imagine though with any generative AI tools there are risks never

10:14

10 minutes, 14 seconds

Sharon: Never. Wow. I think everyone needs to come to you and figure out what you 

Mike: Absolutely. We made, you know, AI that never hallucinates. 

Sharon: So,

10:23

10 minutes, 23 seconds

how do you make it responsible? Tell me about it. What was your journey? 

Mike: The whole concept of responsible is, I think, an interesting one cuz there's so many different kind of facets to it. So,

10:32

10 minutes, 32 seconds

you want to make sure that it doesn't go off on a tangent, right, and answer questions that it it it really shouldn't. So, we've done a lot of

10:41

10 minutes, 41 seconds

training there is it can't answer a question like, "Who should I fire?" it can or it can’t. 

Sharon: Okay. Okay. It cannot. 

Mike: Um and we have to make sure that that it

1

0:50

10 minutes, 50 seconds

doesn't, right? And so we implement a whole series of things like red teaming.

10:54

10 minutes, 54 seconds

So we've got um a small team of folks uh at the uh at the office who will go in and try to debunk it, right? And try to

11:01

11 minutes, 1 second

trick it into giving it. 

Sharon: This is like their full-time job. 

Mike: This is pretty much their full-time job

Sharon: That’s amazing

Mike; . Yeah. Um 

Sharon: how do you get a job doing that? 

Mike: It's pretty cool,

11:09

11 minutes, 9 seconds

Right?

Sharon: Yeah

Mike: It takes because it takes a little bit of understanding what's going on behind the hood and some creativity as well too. Um, so we're looking at ways to automate that a little bit more,

11:20

11 minutes, 20 seconds

which would be really interesting.

11:22

11 minutes, 22 seconds

Um, but yeah, it can't answer things like that. It can't answer things that are completely off base as well. We've had people ask it, you know, uh, who's going

11:29

11 minutes, 29 seconds

to win the World Series? Sorry, I'm a fleet data science uh, agent. I can't answer these kinds of things. Sharon: So, it's okay.

11:37

11 minutes, 37 seconds

So I know a lot of our viewers are thinking about AI. They are thinking about implementing AI and they are also

11:44

11 minutes, 44 seconds

hearing a lot of buzzwords like responsible AI. So what advice would you give them if they're just getting

11:52

11 minutes, 52 seconds

started um and they want to do the right thing? They may not know how to do the right thing.

Mike: I I think a lot of it is a people thing uh to begin with. You have to buy in throughout the organization.

12:02

12 minutes, 2 seconds

So you know we created a responsible AI policy, right? And that grounds how you make a whole series of decisions going

12:09

12 minutes, 9 seconds

forward. Uh so you have to come together as a leadership team because you can have a policy that's drafted but if you don't have full buy in throughout the organization it's not going to really go

12:18

12 minutes, 18 seconds

Anywhere.

Sharon:  Where is it? It's on your website.

Mike: It's on our website. You look up Geotab.com and look up responsible AI policy and you'll you'll find it in there. Ad we also have some tips and tricks of what we did for Geotab ACE.

12:28

12 minutes, 28 seconds

There's a whole document in there shows how we apply a responsible uh AI policy in the implementation uh of ACE. So I

12:35

12 minutes, 35 seconds

encourage all the viewers to go check it out. 

Sharon: Check it out. Um so someone told me 

Mike: Okay 

Sharon: That you um you like uh

12:45

12 minutes, 45 seconds

Chocolate-covered almonds. 

Mike: Oh yeah.

Sharon:  Uh when you go on a road trip.

Mike:  100% I do.

12:51

12 minutes, 51 seconds

Sharon: So here you go. Feel free to bust it open. We are on a road trip after all.

12:55

12 minutes, 55 seconds

Mike: All right. Like I can do this now. 

Sharon: You can totally do this now if you want.

12:59

12 minutes, 59 seconds

Mike: Sharon, I mean, you've given me something here that I'm absolutely going to, 

Sharon: but you know, you have to share.

13:04

13 minutes, 4 seconds

Mike: 100%. What do you think I am? Here you go. There you go. You get the first one, too. 

Sharon: Aw, thank you. All right. 

Mike: Awesome.

13:16

13 minutes, 16 seconds

That's a great question. Who's a better driver? My me or my wife? 

Sharon: Oh, do you both have Geotab devices in your vehicles? 

Mike: Uh, we don't, but I but I have

13:24

13 minutes, 24 seconds

to get one in on her vehicle. I have it on mine. I don't have it on hers yet. So we can So we can So I can So I can ask it in her time.

Sharon: Maybe she doesn’t want it in her car,  She doesn't want you to track her. 

Mike: So I can ask Ace that question.