This episode features an interview with Jack Berkowitz, Chief Data Officer at ADP. Jack has spent the last 20 years as a senior leader at a variety of companies, including being the Vice President of Products and Data Science at Oracle. In this episode, Jack talks about the importance of keeping your product simple, data sharing, applying ethics to algorithms, and much more.
This episode features an interview with Jack Berkowitz, Chief Data Officer at ADP. Jack has spent the last 20 years as a senior leader at a variety of companies, including being the Vice President of Products and Data Science at Oracle.
In this episode, Jack talks about the importance of keeping your product simple, data sharing, applying ethics to algorithms, and much more.
--------
How you approach data will define what’s possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.
Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.
Learn more and register at www.snowflake.com/summit
Steve Hamm: , [00:00:00] Jack. It's great to have you on the podcast.
Jack Berkowitz: Thanks, Steve. It's really great to be here.
Steve Hamm: Okay. Yeah. Now most people have heard of ADP. They know the company prepares their paychecks, but we'd be good to start off. If you could describe the business knowledge dimensions.
Jack Berkowitz: Yeah. So ADP does that. A lot of people know us. The company has been around for about 72 years, maybe 73 now. And. We process payroll, but we do a whole lot more. So we pay about one in six people in the U S but we also provide HR human resource systems for a large number of people as well. We also then process the taxes , uh, for additional people.
So a lot of you may have your W2's coming from ADP, for example , um, we deal with insurance. We deal with retirement. Um, we deal with. Um, money movement. In other words, so a company may give us the money on deposit and then we'll pay taxes [00:01:00] or to all the different jurisdictions where their employees work.
And so we move about two and a half trillion dollars a year. Uh, we issue somewhere around 70 million, maybe slightly more than that. W2's , um, we even get involved in , um, uh, gig workers. Um, and we do this also around the world. So, um, you know, we, we, we have a map footprint here in the U S but also in 140 countries around the world.
Steve Hamm: Yeah. You know, you've, um, you've described, I mean, what is an, almost like a fountain of, of data , uh, it's a tremendous amount of data and, uh, it can be used by companies internally. Or for, to manage their own kind of finances and business and operations, or it could be used kind of combined in different ways and use to analyze the health of the economy or the business community or economic trends and things like that.
So what are some of the more unusual uses or are these kind of bigger, broader, more powerful [00:02:00] uses of ADP's data?
Jack Berkowitz: Or at least the U S economy , which, which then amplifies out to the world economy. Uh, one of the things that we produce , uh, actually as, as a, as , uh, in the public interest, we have a organization here called the ADP research Institute. And every month we generate the ADP jobs report, which tries to give a view of.
Um, the U S economy in terms of hiring and , um, unemployment , uh, growth of wages. And. You know, during COVID , um, this was really interesting and leading indicator about what was going on. So, um, you know, the financial markets look at that data , um, you know, um, once a month, the federal government takes a look at things like that.
Um, university researchers are using it , um, And so that's one of the great, great uses of the data. And it's been so successful here in the U S that we have it in Canada as well. And you'll see us continue to look at it over [00:03:00] time. Now, off of that, we have other groups using the information for things like demand planning or , um, location planning, where companies.
Because we have such a fine grain in terms of what people are making, where are they located? Um, what type of sub-skills are available? For example, in, in South Texas, I was just reading about, you know, Tesla and the rocket launches and things like that. So what type of skills are available in South Texas versus what type of skills are available available here in outside of New York city?
And so we have all of that type of information that people use for lots of different reasons.
Steve Hamm: it's interesting. The Bureau of labor statistics and other, other federal agencies and the state agencies to collect and publish a lot of employment data, a lot of economic, you know, jobs, data and stuff like that. So how is your data different and what makes it so useful?
Jack Berkowitz: Yeah, that's a great question. So a lot of the data that people see in the [00:04:00] labor markets is based on surveys. Um, and so what do I mean by that? So, um, you know, your HR department will receive a survey either from the federal government or from a compensation benchmarking company. Um, and they'll fill out that survey and they'll say, Hey, this is what's going on.
This is how many people we have. This is the average salary, or the survey may be collected directly from you. So you may go onto a site where you're trying to get it. You know, a job posting you say, yeah, I make 180,000 a year. Something like that. ADP's data is actually derived directly from the payroll and HR transactions that are going on through the, through the transactional systems or through the financial networks.
And so our data is very precise and it's also essentially in real time or within a month or so. Um, so the surveys take a long time to, to be mailed in mailed back there's delays to them, there's processing. And you know, if you look at the BLS, essentially about 13 months later, [00:05:00] they true it up to the actual tax data , uh, that the government , um, as you know, when you pay your taxes,
Steve Hamm: Yeah, but you, you have , uh, a bit big slice, but it's still a slice of the market. Do you extrapolate the, the patterns and trends you see in, in, in, from your clients to the larger, you know, economy and, and, and, and jobs picture, or do you do it a different way?
Jack Berkowitz: No, that's exactly right. We, we, we do extrapolate , we, we use external information just the way anybody else would and combine it together. Um, and then we, we were actually using a set of neural networks to do this and we can project , um, particularly for the U S and for other areas , uh, with high precision.
Um, information that covers almost all us workers
Steve Hamm: Oh, so it's more, it's kind of like you do modeling and simulations based on your subset and projected to the whole
Jack Berkowitz: That's right. That's right. Between 18 and 20% of payroll data, another [00:06:00] 20% of tax data. So that combined gives us enough of a foundation, such that when we do the modeling and the extrapolation, it's highly precise.
Steve Hamm: Yeah. Yeah. It's interesting. Now you, I guess you have a whole array of different size and types of clients so that you can kind of replicate the total picture from them. That's that's interesting.
Jack Berkowitz: Yeah, that's right. You know, in the low end of the market, you know, we're by far the market leader in the small business arena. So zero one to 50 employees, and then we have a large chunk. Of the mid market as well. But as we even move into the upper market companies, you know, in the five and 10,000 area, we have a really nice representative sampling there as well.
So we're able to go from the very biggest down to the very smallest and represent, you know, what they need. Um, you know, our biggest customers for this data , uh, is actually those, those clients, right. They're using it themselves for people, analytics, for benchmarking and [00:07:00] everything else. And so we're getting direct feedback on how to improve it over time as well.
Steve Hamm: No leading product development for the ADP data cloud. So we want to kind of establish some basics here. What, what do you mean by that? What is the ADP data cloud? You know, what's the strategy for taking advantage of the data cloud? And , uh, I know, you know, I want to hear about the kinds of applications you're developing.
Jack Berkowitz: Yeah. Yeah. So the data cloud is really a collection of capabilities. On the one hand , we, we build a people analytics product, and that's what we sell to HR departments on the other extreme , uh, we build machine learning. That is part of our applications that we provide to our clients. So if somebody is having an HR application or payroll application, that application will become intelligent through the machine learning apps that we build. Uh, and then the data cloud also is that collection of data that then on the [00:08:00] other side, we monetize and we may monetize it through licensing of data. Like we talked about. So people may , uh, purchase the data or license the data for doing demand planning, for example, or through employment verification. So we have a number of partners , um, that that will do employment verification.
So if you're taking a loan , uh, for a mortgage or a car , Um, and you opt in and you say, Hey, yeah, I w I want you to verify my income. Um, there's partners that that will do that verification. And for that, they will source the information from us.
Steve Hamm: Right. No, I understand that you're building a whole new kind of data distribution business in the cloud. How does that change the way ADP engages with its customers? Kind of what's the quality of the engagement and, and how do you build , uh, and market data products for that?
Jack Berkowitz: So. We are going to the cloud. Um, you know, snowflake is a big portion of what we're doing. Um, we also work in, you know, in AWS to make that happen. [00:09:00] The reason to do this is really about pace and flexibility , um, pace for us , uh, to be able to have that information out, you know, as fast as possible. So, you know, monthly data with only a few days over the lag , um, Variability or flexibility for our clients.
So our clients may be saying, Hey, I want to get information about, for example, in demand planning, I need a different cut. I needed a different slice. I need more industries, or I need a different geo location. So it gives us flexibility , um, and it gives us , um, you know, the ability to deliver it in different ways.
So, you know, one thing about our information it's great. The other thing is you kind of need to have a knowledge about payroll. You know, uh, I get paid once a month. Other people get paid twice a month. Some people get paid weekly, some people get paid biweekly. And what we want to do through the distribution is simplified after people.
And then, so they can consume it in the ways that they want. So, you know, why should they have to come to us and get us to FTP a file to them having the [00:10:00] cut? Why can't they just consume it right inside of the data system that they're using. And so that's what we're looking at in terms of distribution, to make it easier for people to take it and easier for them consume it in a way that they understand it.
Steve Hamm: Yeah. Yeah. Hey, do you think over time, like this will be your main channel of distribution, your main means, or is it just going to be an element or a.
Jack Berkowitz: No, it's a great question as to, you know, how distribution is going to happen. And it really gets down into use cases for the consumers. And what do I mean by that? Um, we just launched a , uh, a SAS product called real income and a SAS product has visualizers that are great. If you're an executive or, you know, somebody like me who has an operational requirement or an operational duty to see migration patterns of people or commuting patterns of people.
So that's one of the capabilities, real income hats has a visualizer inside of it [00:11:00] that shows, you know, who's migrating to Austin and who's leaving Austin. Uh, not very many people by the way are leaving Austin. Um, but, but other people want the raw data, right? Data scientists want the raw data. And then some data scientists don't want the raw data.
They actually wanted in a table format. That's joined to all of their other data. So. It really depends on the use case and, and the profile of the end user , uh, to, to, to take the information.
Steve Hamm: okay. So it'll be kind of a mixed distribution channel for awhile then it sounds like,
now I noticed that you worked for Oracle for a number of years before this. So you've been on the software vendors side. Now you're on the customer side, but you're. Your, your job is to create applications, create software, kind of how you are kind of a bridge between two worlds. How has this affected your perspective now in your new role?
Jack Berkowitz: That's a great question. You know, I came up as a software developer. Um, and I came out building product and I'm building product here, [00:12:00] but I'm also building it on top of platform pieces that I don't need to build myself. Right. So, you know, it's really one of my building. Um, you know, our teams are building content and visualizations of content.
Um, and so those pieces of infrastructure are those pieces of analytic capability. Um, we're more, more likely to go and, and. For lack of a better phrase, OEM it , um, incorporated into our, into our products so that we can provide the content. And, and that's really it. We see ourselves as a content team. Um, we will build technology if we need to.
Um, but mostly what we want to do is , um, use the best of breed technology, bring it together, get it out the door fast so that our users get the greatest content available. Um, and that's, that's, that's what we're focused on here.
Steve Hamm: Yeah. So it's a, it's a stack and a lot of the bottom layers come from somebody else. And I imagine you throw in a lot of open source code and stuff like that. Into the, into the mix where you build. Right. Um, I, I'm curious how important is [00:13:00] visualization these days, visualization of data. Is that something that's really taking off that your, that your consumers within the company or your, your clients really want to see or is that, you know, is it a big deal?
Jack Berkowitz: Again, it depends on who and it , um, for our HR users or our payroll users, the people who, who, who, who are using the bulk of our people, analytics. It's vital. It's vital. These people. I think the, the misnomer is, is that , um, people in companies have infinite time to explore data using Gillies. And for most of us, that's not the case, right.
They, you know, the HR person or the payroll person, or, you know, the person doing operations, planning, demand, planning. That they may want to do some deep investigation. Quite frankly, if you give them a picture or better yet augmented analytics that processes the data on there, perhaps on their behalf. And I lights an exception on a visualization.
That's what they prefer, then they [00:14:00] can to go dive in. Um, and so we build and spend a lot of time talking about visualization. I have a very big. User design team as part of my, my group. Um, but really what we're trying to do is make visualizations incredibly simple for people to absorb really complex information.
Um, if you could see me right now in my office behind me are, you know, all four volumes of Tufts, these , uh, books, and , uh, you know, these are, these are, these are things that , that, that we look at frequently. And so, uh, visualization is important to us.
Steve Hamm: Yeah. Tufty you know, I think it takes a long time for people to understand that a big complex thing. Isn't a good visualization, you know, something that does graphically tells the story that's as simple as can be, is going to be the powerful way to go.
So.
Jack Berkowitz: I started my career as a flight deck designer. I have a under, I have a graduate degree in human factors engineering. There's not a lot of us out there. Um, and I started working on , um, uh, airplanes when I first got out of graduate school 30 years [00:15:00] ago. And I remember somebody telling me early in my career at the FAA, Hey, look, let's not overcomplicate things.
If a pilot is flying into a mountain, let's just have the alarm, tell them to pull up. And it's a really good point for designers and people like us that are into data to keep in mind. Um, you know, sometimes the, the user just needs a pull up and then later they can go in and explore what I went on. And so, uh, it's just something that I always, we always try to keep in the back of our minds , uh, as we're building.
Steve Hamm: that's interesting. It's kind of like, keep it really simple on the surface, but then give people like infinite capabilities of drilling down and exploring, you know, when it's. When they have the time and inclination,
Jack Berkowitz: Exactly. Exactly because they do need to get to the core reason. Um, and, and, and we need to provide them all that ability to get there.
Steve Hamm: no. ADP is kind of in the middle of migrating a ton of data to the cloud and then also distributing data products from the cloud. [00:16:00] But I imagine have been challenges. Have you, as you've done this, this is, this is a new thing. So I want to explore a little bit, what are the challenges that you've encountered and how have you,
Jack Berkowitz: So data gravity remains the largest issue for us. And , um,
Steve Hamm: I don't know what that is. I'm
Jack Berkowitz: yeah, let me explain. So where's the, where's the weight of the data. Where's all the data actually at, so it's great. You want to migrate it to the cloud, but boy, you know, we have 70 years of, of systems running in existing data centers. So one of the big challenges is for us to actually gather and align, move and clean the data as it goes up to the cloud.
Right. The second major challenge is, is an age old challenge that people in the BI world have had, or the business intelligence world or the data world have had for years. And it's the second problem isn't solved. And that's, that's the shared semantic. So what do I mean by [00:17:00] that? You know, the metric head count in the company?
Well, you would think that's an easy thing to do. Well, there's a lot of variations, you know, so am I counting the people. That are here today? Yes. Am I counting the people that have resigned but not left yet? The people in the UK might be on garden leave? No, maybe am I counting the people that have accepted the job, but haven't started yet.
Am I counting the people that are out on maternity leave or paternity leave? So the definition of metrics is the second major challenge that we've had as we moved to the cloud. Not because of anything other than that, I'm. Uh, that our team is part of a 9,000 person technology group. And across 9,000 people, there could be lots and lots and lots of different interpretations of, of business roles and software and best, best in breed, best practices.
And so our ability to bring the data together as a data platform, and [00:18:00] then to have, you know, those, those, the variability or the flexibility, but still standard metrics. These have been our two biggest challenges as we moved all of this into the cloud.
Steve Hamm: now I'm guess I'm just guessing here, but it seems like probably a lot of your data and a lot of your processing traditionally has happened in IBM mainframes. Uh, Does that create any challenges for, for, you know, migration to the cloud?
Jack Berkowitz: Yeah. So a lot of its been an IBM and a lot of it it's been an Oracle and a lot of it's data and other things , um, No, I think, I think it could generate problems for us , but, but one of the greatest things about , uh, that the company overall has, has been United United on this. And so, you know, we've been able to work to get API APIs so that the information can be streamed out , uh, from, you know, the mainframe systems.
So we, we have a streaming API on top of, on top of our mainframe systems. We have streaming API APIs. Uh, and streaming systems from legacy Oracle [00:19:00] database is another systems databases. Um, that only happens though, because all the development teams are bought in, on our notions of, of data match architectures and things like that.
And that only happens because our team has done a really nice job explaining our strategy about moving to the cloud. I'm really proud of. The amount of architecture and the engineers working together with some of our key partners , uh, to be able to describe what's needed. And , um, you know, I, I've been here in the role about two and a half years.
I haven't had, you know, sort of what I read about online, about, you know, people wanting to control their data. And I haven't had that experience here at ADP and , uh, that's, that's been really refreshing and it's helped us hit our goals.
Steve Hamm: Oh, the idea of, of business leaders wanting to hoard their data and not share it with anybody, that kind of thing.
Jack Berkowitz: Yeah, maybe that exists, but , um, we haven't seen it here. And the more I talked to my peers, either in other technology companies or other financial services [00:20:00] companies, I think that's an aging practice. I think, I think a lot of companies now realize. Hey, no, it is all about the data, you know, maybe five or 10 years.
It wasn't, but today I think everybody gets it. It's all about the data and we've got to share it to be able to move forward. And so, you know, moving stuff of a mainframe, you know, if you want to talk about the classic hard problem to solve , uh, but, but the team solved it very efficiently.
Steve Hamm: Yeah, it's interesting. I've heard people comment that you know about that hoarding data thing that yes, there's value in keeping the data close to a specific business unit, but for an enterprise or, you know, internally or with its partners, there's much more value in sharing it. And I think that is a lesson that is now accepted by most everybody.
Jack Berkowitz: I would hope so. I think there, there is the, the concern. Obviously around security and, you know, we're dealing with data. It's not [00:21:00] necessarily health data, but it's, it's pretty sensitive to individuals. What do you make? You know, uh, that's something that's very sensitive. And so, um, you know, we are, if there's one thing about ADP we're, hyper-focused on the security of everything , um, you know, the.
The encryption, the role-based access controls, the the masking, all of these types of things they're present throughout all our data pipelines throughout our data analytics throughout our machine learning and throughout our data distribution as well. Um, anonymization is utmost in any of the data that we release.
And , uh, these are important principles , um, that, that, that we live in.
Steve Hamm: Right, right. That makes a lot of sense. Now, early in our conversation, you mentioned snowflake, you're a customer. Um, so let's go back in history a little bit. When and why did you begin working with snowflake?
Jack Berkowitz: Well, we've been working together , um, you know, to prove this out [00:22:00] over the past year. Um, does it March now, right? So it was really at the beginning of the pandemic is when we started to really get engaged , um, shortly after or no, it's April now. Excuse me. Um, so it was , uh, about that time shortly after that time is when we really started to dive in and it was really for three reasons.
Um, and it's everything that we've talked about so far, one of them was around pace. Uh, You know, our ability to just iterate and have new , uh, data products, new capabilities available. Second thing is, is about access and that shared control of the data and the metrics and everything else. Uh, and the third thing is the distribution , um, snowflake , uh, it, the ability for us to deal with distribution in interesting ways and.
Right. You know, we're hearing from clients that they want to use it and we're interested in using it. So it made a really nice move for us to adopt the platform.
Steve Hamm: Do you find that a lot of your clients are already using snowflakes? So that, that all that data sharing is just so much easier. [00:23:00]
Jack Berkowitz: Um, increasingly , uh, we still have clients that are, that, you know, are, are, you know, Looking at the use of different types of systems. Um, but the flexibility is, is there and we are getting the demand. Hey, do you guys have a snowflake account? Can you just distributed to us that way? Um, just one of the biggest consulting companies in the world, just ask me exactly that question two weeks ago.
And it was nice to be able to say, yeah, we can handle that. Um, and so we're, we're, we're seeing it take off that way.
Steve Hamm: Mean, it was kind of a habit, your away thing, right? Yeah. Hey, so let's get into a little bit of detail here with the snowflake data cloud. If you could mention a couple of applications, a couple of ways that you're using it kind of most intensely and with the most, you know, the clearest results, that would be helpful.
Jack Berkowitz: Yeah. So, so the first thing that we're doing with it is providing. Our benchmark data, whether it's compensation, data or benefits data, or any of those types of [00:24:00] aggregations, sort of that national data that we were talking about , um, we're using it to calculate in store and distribute that information , um, downstream and yeah.
And the other thing that we're using it for is taking exactly the same information and backing one of our existing SAS applications. With snowflake, as opposed to our previous approach. And so snowflake is essentially acting as a, uh, a multi-tenant SaaS enabler for us, uh, for, for a visualization capability, um, for our HR and our payroll customers.
And so those are two of the, the first two places we're using it, um, is around this anonymized aggregated benchmark compensation, benchmark information.
Steve Hamm: Yeah, no, that sounds great. Now I understand that you have some big plans for using Snowflake's data marketplace as a sales and distribution channel. Why is the data marketplace model so [00:25:00] powerful for you and how do you plan on using it?
Jack Berkowitz: Well, it's powerful for us. Um, in terms of distribution, you know, snowflake has, you know, a growing number of clients. And it helps us with our sales acceleration, our ability to get traction with the data scientists, that those customers at those clients, at those customers, those joint customers. Um, and as we can put it into the context that they're going to be doing the balance of their data analysis, um, our data just shows up as, as, as additional things that they can use.
It just really streamlines. It takes the friction out. Uh, I think that's really, it, it takes the friction out. Now, one of the other things we're doing is we're combining our information ahead of time with, um, some other data partners to give even richer data sets. And so one of the ones that we've, that we've been working with is Intercontinental exchange.
These are, um, people that work in the financial markets, um, Intercontinental exchange also, um, the New York stock exchange as part of [00:26:00] Intercontinental exchange, we're working with the data group there. Uh, I, ice data services, uh, build a value, add combination of our information and then in turn, distribute that, uh, out via snowflake.
And so that combination is for, you know, Munich, well bond ratings, just something that ADP would never have gotten into. And I think that's the key. These are use cases that ADP would never have gotten into. We don't know much about. But by using the data exchange, we're able to combine the two data, you know, between ice and ourselves and come out with a new data product that nobody's ever even thought about before.
And I think that's the exciting part of it. Um, who knows where it's going to go. You're just going to see combinations of information for all sorts of different use cases over time.
Steve Hamm: that's really interesting. So if, uh, if, uh, Customer goes onto the data marketplace and kind of shopping around with a cart, you know, looking for all kinds of data, basically, you're saying, Oh, you don't have to like pick every individual, you know, source we'll combine it for [00:27:00] you in kind of use cases so that you just buy the package.
It's all there. It's pre-integrated ready to go. That kind of thing. Right.
Jack Berkowitz: You see these sort of ecosystems of ecosystems. Right. I could source shame of a consortium, uh, emerge, right? So here's the data you need for, or, you know, understanding the economy. And instead of me having to go pick a thousand sources, what do I know about, you know, Irish shipping, one of the best use cases I ever heard or data monetization was Michelin tires being shipped in and what that meant to the economy in France.
Right.
Steve Hamm: Oh, as an
Jack Berkowitz: do I know? Yeah, it's an indicator. What do I know about tires, but the idea that, that somebody is going to combine that together, and then you can just say, Hey, I've got an economic indicator. That includes 25 sources. Fantastic. Now it's early days. Not all of those exist, but you're going to see these over the next few years.
And we're excited to be part of [00:28:00] that, uh, ability.
Steve Hamm: I sense.
Now your businesses, we've been really. Abs really vital for the economy and for the business world the last year during COVID, because some, you know, it's such an unusual situations. Things are so hard to predict, you know, strengths that were not expected, popped up. I mean, look at the stock market weaknesses.
We still don't know where all the weaknesses are. Uh, and I would imagine that in the, in the coming months, as. You know, governments and businesses, try to kind of put things back together and put them together in new ways even, but that your data and also the technology used to deliver it will be really important.
So I want to kind of ask you, look ahead over the next year. What are the most important trends that you see in cloud computing, data analytics and, you know, specifically in your marketplace?
Jack Berkowitz: Yeah, it's interesting. Right? Because this time, last year I was sitting right here at this desk, looking at the screen, [00:29:00] watching the economy smelt, and there was no other way to put it. Um, but where do I see it going next? Um, well, certainly we're seeing the strength pickup through these early signals and we see things like job postings at way higher than pre COVID levels. We see things like background checks way higher than a little bit levels. And so, you know, we certainly see strength in the economy. Overall. I think that technologies that are important. Are related to pace it's related to security. And as we discussed, um, I think there's a world of ethics as well. That's going to start to play into whether it's people, data, or any data as to whether or not you have the information, but should you be using it?
How should you be applying it? Um, we all remember that [00:30:00] there were hedge funds that shorted the market in a year ago. and I suppose in some people's minds, well, they made money. So that was okay.
Steve Hamm: they shorted it. And as a result, it went deeper and worse
Jack Berkowitz: it went deeper.
Steve Hamm: might have. Yeah. Right.
Jack Berkowitz: Yeah. I think all of us in the data world, um, should really be thinking about. The ethics of when to apply and how to use, maybe it's not the data, but maybe the algorithms or maybe just the decisions of how to apply. Um, I think it's, you know, if you look at the news today about what's happening, maybe not in the U S with the vaccination program and how well things are going here.
But if we think about everybody else around the world, um, you know, I think one area that is not necessarily a technology. What about technologists you know, what do we do for the information so that we [00:31:00] don't find ourselves in the same situation, five years down the road. And so I, I have a lot of trends in terms of, you know, new technologies.
I'm talking to the heads of various know technology partners all the time, demanding, Hey, I need this feature in that capability. Um, we're pushing the outer edges in terms of security, but, um, one area that I, I personally, uh, am spending a lot of time on is data and AI ethics. And you're going to see that as more and more people pick up ML, operations, ML, ops tools, you know, the ethics, the ability to monitor, you know, it, is there a bias?
Are you, are you not giving. Uh, equal opportunity to women owned businesses. Are you not giving equal opportunity to people of color through your algorithms? Now this needs to be transparent and you as a technologist need to take a decision as to how you want to behave and how, what you want to do.
Steve Hamm: Yeah, a,
Jack Berkowitz: it's a little bit [00:32:00] different of a discussion then maybe. We were all having two years ago or even a year ago. That's it.
Steve Hamm: Well, I actually did some research into this recently looking at what organizations have kind of ethical guides or practical guides for using. Um, artificial intelligence and, and usually data is a piece of this, but you know, the us department of defense actually had one of the highest quality and most thoughtful, uh, guidelines that I had seen around.
And it was really kind of impressive, but, you know, you understand. I mean, they have an immense amount of very sensitive data and they have, you know, 1.5 million active, you know, uh, military employees and other stuff. So they really aren't the cutting edge of a lot of these questions. So, uh, I was kind of surprised, but pleased to see that they had done that kind of
Jack Berkowitz: Yeah, we at ADP, we did exactly the same thing about just before this time, last year. We published our guidelines there on our [00:33:00] website of AI data and AI ethics. Um, we actually have a sitting board of people from ADP as well as outside experts where we get together. And we talk about data and AI ethics.
We review programs where people want to use data. People want to build AI capabilities and. We, you know, we check our programs to ensure that they're aligned with ADP's values that right before this podcast, I was in an AI data and AI ethics board meeting. And I think, you know, you're going to see this as a trend, hopefully over the next few years.
And, um, You know, that's the thing, right? You know, this pandemic has been something, uh, significant. We'll all remember it in our lifetimes. And I think it's something that we as technologists and I am, you know, think of myself as technologist should end, can contribute back to the world.
Steve Hamm: right. Right. I like that idea of having the advisory [00:34:00] board. And, and kind of, you know, it's so important to think about these things ahead of time because technology and business are moving so fast and it's hard to put the genie back in the bottle. So you've got to kind of anticipate, you know, What might happen and try to head it off.
I mean, I even think of this as like nuclear non-proliferation, you know, I think it's like AI and data, uh, you know, making sure that this doesn't become damaging. And I think it looks like you guys are really on a path to, to assuring that or trying to make sure of that.
Jack Berkowitz: Yeah. Yeah, it's interesting. Right. Um, I reflect back on my career and these weren't concerns that I had in the internet rush. Right. I'm an older person now I've been through it. Right. I remember when somebody said, have you heard of J J T E yeah, I remember those days. I remember I remember the internet.
Boom. I remember the bus did these weren't concerns, you know? Um, but they can, but we have so much power at our fingertips today. [00:35:00] That these are concerns that we all must have.
Steve Hamm: Right. Yeah. So we've talked about kind of the next year or so of the kinds of things you guys are watching, but I want to ask you to put on your visionary cap for a minute and look out like five years or more. How do you see data, the data cloud, and just big data in general impacting business and even society.
Jack Berkowitz: You know, the, the level of automation is I think going to be stunning to us. you know, we've always had this notion since probably Dick Tracy cartoons or things like that back in the forties about this mixed initiative, um, interaction between people and machines and, you know, today it's still a pulled from person Siri, do this, you know, I think.
When we start to really have the mixed initiative, the push from the system, this [00:36:00] is a real Bush, you know, not, Hey, do you want me to remind you of the alarm? But, you know, just whether it's alerts, whether it's, I've taken care of this for you, that level of automation I think is gonna really transform.
I don't know if it's five years or 10 years. But it's in that time range, right. We're seeing it now, the, just the explosion of, of, of electric cars. I was just looking at the new review of the Mercedes that just was released last week. And you know, this level of automation is going to be kind of amazing.
Now people will shape it to where they want to use it, but, you know, it's kind of crazy. It's 2021 and I need a new iPhone. 2007 is when the iPhone came out. He came out 15 years ago. I can't believe it working years ago. I can't [00:37:00] believe it at this moment. Still feels like a new device to me at a time right now, but think about what it's going to be 10 years from now.
And, um, I'm just amazed to see. All of that's going to be enabled by data, all that it's going to be enabled by these technologies that we're all working on. Um, but it would be really interesting to see, you know, um, I, I just saw these photographs where I can take an old photograph from the forties and animate animated so that the person seems to come up.
Steve Hamm: Oh my God.
Jack Berkowitz: What, what. Well, it's just mind boggling. And then, so, you know, when I sit down and I talk with somebody, not just from my, you know, my parents' generation, but even the generation before, and they're still with us, right. People from world war two. Uh, think about what they saw when they were, when they were in their twenties and thirties or yeah.
When they were in their teens and twenties. Uh, we think about the kids today that are eight and 10 years old. What they're going to see
Steve Hamm: Acceleration is
Jack Berkowitz: when they're coming out of college and. Well yours. Yeah, exactly.
Steve Hamm: You know, [00:38:00] it's funny. My father died two years ago at age 98. When he was a teenager, he was plowing fields with a mule. And in the 1960s, he was designing moonwalk vehicles. And so it's almost like from the bronze age to the, you know, the electronic age of the digital age and, and just a few decades, but things have accelerated since then.
And I think that's what you're talking about. I think it's really amazing. And I really think we have to think about a partnership between. Artificial intelligence and human intelligence, because it's really through an interaction and an engagement that you really get the big benefits. It's not, it's not, it's not just about automation anymore.
It's really about, um, collaborating and getting things done that neither humans nor machines could do as well on their own.
Jack Berkowitz: That's right. That's right. And, and, and how that comes together is going to be fascinating. You know, I just saw that you can buy one of the Boston dynamics [00:39:00] dogs for $74,000, but. I'm not sure I'm ready to do that yet. I still like my, my, my, my guy, but, um, but yeah, uh, you know, that's $74,000. It's going to be $740 before you know it.
And, and then what do you do? How are you going to interact? Is it going to play fetch? I just don't know
Steve Hamm: I don't think anybody's going to be carrying backpacks on their, on their hiking trips anymore. They'll just have a little mechanical dog behind them carrying
Jack Berkowitz: exactly. Exactly.
Steve Hamm: Yeah. All right. Hey, um, you know, we always like to finish off our podcasts with a question that's kind of, you know, more personal, lighter, more fun and stuff like that.
And I understand that you're one of those people who's who typically in the, in the past, before COVID would fly tens or hundreds of thousands of miles a year for business. COVID obviously has changed all that, you know? So I want to understand how has COVID changed your work life and how has it changed the way you think
Jack Berkowitz: Great question and yeah, I, you know,
Steve Hamm: that's not [00:40:00] really a light question I realize, but yeah,
Jack Berkowitz: playful,
Steve Hamm: but it's personal,
Jack Berkowitz: I'll give you a couple of decent examples. Um, yeah, I did. I, I, I flew one in two back-to-back years. I flew over 300,000 each year, actual flight miles, not yet the United premiere accelerator miles. And. Um, it's a lot of lonely time when you're sitting in a plane and better yet.
A lot of lonely times when you're sitting in the airport, waiting for that plane to take off. certainly I sleep eight hours a night now and I don't feel guilty about it. probably lost 15 pounds. I'm not on the blood pressure medicine that I was on because I was jet lagged all the time. Um, But I, I have time to think deeper I don't think I'm the only one that has that time.
Now, that being said, I think most have a little bit more [00:41:00] sensitive. You know, I work as part of we're a worldwide organization. Like I said, we have people in 140 countries. We have people that work in my team in Brazil, in India, in Barcelona, Spain, You know, right now they're under, uh, an awful on slot.
And I think about them not as units to production, not as members of my technical team, but I think about them as people not to say, I didn't think about them as people before, but it's really a depth now on a depth of pondering about who I work with. Where my family is how they're doing, how my kids are in they're in different cities, around the world, around the country. a lot of it is, is remember I was talking about my dog Baxter and I, we, we get outside a little bit more now than we used to. We wrestle in the grass a little bit more than we used to, [00:42:00] and I think that's, uh, some good, good, good, good Daff.
Steve Hamm: Yeah, that's really interesting. It's interesting what you said about kind of slowing down and thinking deeper. Cause I hear that from a lot of people and I, you know, and this is kind of a corollary, but the other thing I hear is that people have changed the way they listen. That may be in the past, the big rush.
It was always kind of listening to respond. And now it's kind of listening to comprehend and understand in it in a deeper way. And I just hear a lot of people talking about that and I think that's going to cause improvement in business, but also just in, you know, in life, the way we relate to our loved ones around us and stuff like that.
So it seems like a good thing.
Jack Berkowitz: Yeah, I think, I think that notion of active listening. I had a college graduate professor who, you know, tried a beach that to us all those years ago, active listening, ask the questions, listen to the answer.
Just don't [00:43:00] hear it. And you know, I haven't talked to Jeff in 30 years, but he was spot on. And I think, I think this is the notion of actively listening to people is something that I hope all of us remember, and then pass down. Uh, to our children and, uh, the people, who've all of us in our, in our careers as we go forward because you know, business can be tough.
The world can be tough. Um, but we have time to listen to each other.
Steve Hamm: well, that's a, that's a good way to, to leave it here. I, I want to thank you for today's conversation. It's been kind of packed with information and interesting thoughts, but also I think with some of these bigger ideas, I mean, you know, it, it's very encouraging to me and I, you know, I've been up. A business reporter for many, many years and a tech reporter and been in, in, in various roles.
And. I defend business a lot to my friends, you know? And, but it's really great to hear you talk because I feel like this is an example. This is the way a responsible, uh, [00:44:00] business people operate, you know, and you know, we, you talked about AI, you talked about the data and it re it's really encouraging me to, to me to hear this.
So thanks very much, Jack.
Jack Berkowitz: Well, thank you, Steve. It's been, uh, been my pleasure and I'm look forward to catching up again sometime soon.