Anders Butzbach Christensen started his profession in product administration earlier than touchdown his dream job working for the Lego Group in Denmark. As we speak, as head of information engineering, he’s main Lego’s digital transformation with a particular concentrate on designing and constructing knowledge merchandise, together with self-service functions that know-how and enterprise groups can all use to higher serve their prospects.

On this episode of the Me, Myself, and AI podcast, Anders joins Sam Ransbotham and Shervin Khodabandeh to explain how the Lego Group is approaching digital transformation, and the way the toymaker is empowering its product groups by turning into a product-, architecture-, and engineering-led firm.

Learn extra about our present and comply with together with the sequence at https://sloanreview.mit.edu/aipodcast.

Subscribe to Me, Myself, and AI on Apple Podcasts, Spotify, or Google Podcasts.

Give your suggestions on this two-question survey.

Transcript

Sam Ransbotham: Our visitors usually use Lego as an analogy for a way organizations can construct up options with knowledge. However right now, learn the way Lego itself builds knowledge elements that join as simply as its bricks.

Anders Butzbach Christensen: I’m Anders Butzbach Christensen from the Lego Group, and also you’re listening to Me, Myself, and AI.

Sam Ransbotham: Welcome to Me, Myself, and AI, a podcast on synthetic intelligence in enterprise. Every episode, we introduce you to somebody innovating with AI. I’m Sam Ransbotham, professor of analytics at Boston Faculty. I’m additionally the AI and enterprise technique visitor editor at MIT Sloan Administration Evaluate.

Shervin Khodabandeh: And I’m Shervin Khodabandeh, senior accomplice with BCG and one of many leaders of our AI enterprise. Collectively, MIT SMR and BCG have been researching and publishing on AI since 2017, interviewing lots of of practitioners and surveying 1000’s of corporations on what it takes to construct and to deploy and scale AI capabilities and actually rework the best way organizations function.

Sam Ransbotham: As we speak, Shervin and I are excited to be joined by Anders Christensen. He’s the pinnacle of information engineering at Lego Group. Anders, thanks for taking the time to hitch us. Welcome.

Anders Butzbach Christensen: Thanks for having me, Sam.

Sam Ransbotham: First, inform us a little bit bit about what you do at Lego Group.

Anders Butzbach Christensen: I’m heading up the information engineering division inside the Lego Group. We at present encompass three massive international components, every inside my space. Two of the groups concentrate on self-service, enabling the group to make data-driven selections, and the final one is constructing a buyer 360[-degree] view that permits us to construct customized experiences.

Sam Ransbotham: Let’s begin with the primary one. What does that imply, to be self-service?

Anders Butzbach Christensen: So rather less than two years in the past, we began out our exploration of the digital transformation inside the Lego Group, and for us that mainly meant that we wanted to do numerous upskilling and we wanted to concentrate on having the fitting competencies and groups and methods of working inside the group — so mainly constructing the fitting digital basis. And to ensure that us to allow the 4 buyer teams that we now have — the customers, the shareholders, the companions, and our colleagues — we wanted to guarantee that that they had all the fitting tooling to take action. And an enormous a part of doing that’s self-service, enabling them to make data-driven selections. So, what we did was that we took this centralized knowledge platform, [like] virtually all massive corporations have right now, after which we made that accessible for everybody to make use of, mainly. And that’s self-service.

Sam Ransbotham: So, what does that appear to be? If I sit down tomorrow with the Lego Group, and so they received’t let me play with the bricks, how do I play with the information?

Anders Butzbach Christensen: What it mainly means is that it’s tremendous simple for the product group all through the group to return with their knowledge, carry it into the platform, after which to mess around with the information — rework it in no matter approach they need to — after which expose it for no matter use they’ve. That is perhaps for analytical functions, but it surely may be for knowledge science functions, and so forth. Making that journey as simple as doable and accessible to all varieties of talent units inside the group is what it seems like.

Proper now, it’s used for mainly the whole lot: That’s all varieties of knowledge coming from our web sites flowing into the platform. After which we have a look at how prospects behave on the web site after which present the very best advice expertise to them.

That’s one factor, however we additionally use it for forecasting, for instance. We now have numerous totally different knowledge units coming in from our demand planners throughout the globe that every one get constructed into a fantastic knowledge product that’s used for creating this forecasting mannequin.

Shervin Khodabandeh: Anders, what I’m listening to is knowledge platforms and knowledge engineering, however I’m additionally listening to knowledge science in there — suggestions and demand planning. Does your group do each?

Anders Butzbach Christensen: The best way that we’re organized inside the group known as the information workplace. We do have a knowledge science [area]; they concentrate on numerous the information science work. We additionally do use knowledge science inside my space, however the best way that we make the most of it’s for enablement. So this could possibly be, how will we construct knowledge that permits individuals to innovate sooner? In our use case, that’s enabling artificial knowledge on the platform. So at any time when somebody comes alongside and needs to make the most of a knowledge set that doubtlessly incorporates personally identifiable info [PII], they want a authorized approval, proper? And that’s as a result of we have to deal with our prospects’ knowledge.

So, what we do is that we give them an artificial knowledge set that’s generated primarily based on the schema, after which they will get began immediately. It has zero similarities [with] the unique knowledge set, but it surely offers them the identical output. That’s a approach for us to make that knowledge science group innovate approach sooner than they initially would, as a result of it may well take months earlier than you get that authorized approval.

Shervin Khodabandeh: Sure, sure. Artificial knowledge’s actually fairly good. Sam, when you keep in mind, after we had the dialog with …

Sam Ransbotham: Humana.

Shervin Khodabandeh: Humana, sure. In addition to, I believe, with Moderna.

Sam Ransbotham: Oh, mm-hmm.

Shervin Khodabandeh: Artificial knowledge for the aim of experimentation. It’s fairly fascinating. Anders, you talked about “knowledge product” a couple of occasions. It’s been a reasonably large buzzword for a few decade, at the least that I’ve heard. A variety of organizations we converse to are constructing knowledge merchandise. How would you outline “knowledge product”? What would represent a knowledge product? What are elements of a knowledge product?

Anders Butzbach Christensen: If we to begin with have a look at the definition of the information product, then for us, a knowledge product is a mind-set. It’s a mindset, proper? So we used to consider knowledge as a byproduct — one thing that is part of a product out within the enterprise, or one thing like that. However to us, it’s actually essential to concentrate on knowledge as a product, and that’s why we name it a knowledge product.

In case you are very technical, it could possibly be a database that consists of numerous knowledge units with numerous totally different knowledge attributes inside it. In concept, it’s numerous totally different inputs that go into one product, proper? However for us, it’s actually the mindset. It’s the truth that that knowledge is now not simply one thing we use as part of our day by day work. It’s how we make selections. It’s how we create worth.

Shervin Khodabandeh: Once I take into consideration a product, a product would have a product highway map, and it could evolve over time, and you’ll innovate on a product. Is that additionally what’s happening along with your knowledge merchandise?

Anders Butzbach Christensen: Positively. We view it as another product. It’s no totally different than one other software program product we construct. instance is that if we have a look at the 360[-degree] view that we’re constructing inside my space, there are numerous totally different knowledge units that go into that knowledge product. I believe we now have 4 or 5 totally different IDs throughout the Lego Group that we have to sew collectively. In an effort to construct this 360 view, we have to herald these IDs, and to ensure that us to supply worth to … let’s say that we need to construct a personalised e mail marketing campaign or one thing like that. The best way that we advanced this product is by taking a look at what’s the minimal viable product that we have to construct. What knowledge units do we have to herald, attributes, and so forth.? After which we will mainly evolve from there on. You don’t want all 5 IDs to be able to ship that worth. You possibly can truly herald perhaps one or two after which begin constructing these customized campaigns.

Shervin Khodabandeh: The fantastic thing about what you’re saying is, along with constructing these intelligence merchandise on prime of your knowledge product, you’re not simply constructing it one time, as a result of the remainder of your group may use numerous these belongings sooner or later. And that is fairly essential, as a result of numerous the time you see in a corporation, perhaps they construct a personalization engine particularly for what they need, and all of that knowledge pipeline and knowledge engineering then go to waste. And any person within the subsequent room would do the identical factor, and so they’d go all the best way again to scratch from the uncooked knowledge once more.

And I believe what you’re saying is, you’ve created the commonality, or the frequent layers of information, after which of us simply use what they want once they want it, but it surely’s not duplicative.

Anders Butzbach Christensen: You’re spot-on, and that really [speaks] quite a bit to the reusability of information merchandise throughout the group. After which, when you may have onboarded all this knowledge into the platform and constructed all these knowledge merchandise, it’s additionally actually essential that you just make it discoverable to the group in order that others can put it to use for different functions and create different varieties of worth.

What we do is that we now have a discoverability software, the place you possibly can go in and have a look at the totally different knowledge units. After all, there are non-public knowledge units, as a result of if it incorporates, let’s say, personally identifiable knowledge, then it must be non-public and it’s essential to request entry. However that may be knowledge merchandise that don’t should be actually non-public and may be publicly accessible for individuals to faucet into.

And it’s actually essential for us to make it discoverable on the platform so that you could begin stitching the information collectively and create new varieties of worth, and also you don’t need to reuse or re-ingest or issues like that, as we now have seen throughout the totally different organizations for years.

As a substitute of being very restrictive about the way you construct these kind of platforms, it’s essential to consider it as empowering the product groups to get as a lot freedom as humanly doable.

Shervin Khodabandeh: It’s kind of just like the precise bodily Lego blocks, proper? Fast digression: We most likely have 300 Lego units that my youngsters have constructed. They usually serve their authentic objective, following the directions, after which they find yourself creating different monstrosities or cool creations, so I might completely see that analogy.

Sam Ransbotham: I used to be simply enthusiastic about that. We now have so many individuals who come on our podcast and point out, “Oh, it’s like a constructing block,” and so they all the time make an analogy to Lego, after all. And Lego itself is utilizing these constructing blocks and their very own analogy.

Shervin Khodabandeh: Sure.

Anders Butzbach Christensen: Couldn’t agree extra, however I believe it’s additionally actually essential to say that what we do is that Lego is a bodily toy model, proper? We construct bodily toys, and what we do within the digital house is that we allow higher experiences for these bodily merchandise. It’s simply to reinforce that have for our customers, and to make the expertise of shopping for merchandise and interacting with merchandise far more enjoyable for our customers. And I additionally suppose that that [speaks] very a lot to how you need to use issues like machine studying or AI to then guarantee that we do it in a accountable approach.

So, what we’re additionally taking a look at is, how can we make the most of machine studying to audit the platform and guarantee that we don’t have knowledge mendacity round ceaselessly, but it surely truly will get cleaned up, and on the identical time, additionally have a look at PII detection? How will we guarantee that our builds and product groups truly know if they’ve PII knowledge inside the platform or not?

Sam Ransbotham: Folks can’t see this, however within the background, Anders’s workplace is full of these toys. I’m fairly jealous, as a result of my background is fairly plain right here. What sorts of issues are individuals constructing along with your knowledge bricks?

Anders Butzbach Christensen: We truly do have numerous cool experiences that received printed not too way back. We did the launch of Lego Tremendous Mario, I believe final 12 months, and that was an enormous success. That may be a toy that isn’t solely bodily, but it surely’s truly additionally interactable by these digital experiences, and that’s one thing that was an enormous success to us.

Shervin Khodabandeh: Once I hear [about] the information product and varied usages of it … I believe inside the enterprise, you talked about planning and optimization, and I might think about, as a buyer, interacting with Lego web sites or directions or buying advice engines and issues like that. However you additionally talked about one thing across the precise expertise of constructing, like throughout play. Are there additionally merchandise that allow higher play or totally different play experiences?

Anders Butzbach Christensen: Round three or 4 years in the past, we launched a cellular utility the place you might use AI to do totally different play experiences. However … you wanted to have a telephone or an iPad to be able to work together with the expertise. I believe for us, it’s combining it, like we did in Tremendous Mario, the place it was inside the toy, that made the distinction.

Shervin Khodabandeh: Anders, that is fairly fascinating, what you’ve shared with us, and I’m certain a lot of our listeners and folk who put on your hat in different organizations are questioning if there’s a secret sauce or if there’s a recipe. Do you need to share with them and us, what does it take to truly construct what you guys are constructing?

Anders Butzbach Christensen: I believe one of many actually essential issues for us has been to empower our product groups to truly construct these merchandise and take the possession of them, and the best way that we did that was by establishing three totally different pillars inside the firm. We needed to be product-led, architecture-led, after which engineering-led. So, what it means is that being product-led … we truly acknowledge that the totally different groups are merchandise themselves. We now not have these tasks which have a begin and finish date, but it surely’s merchandise which might be evolving.

Then the second developing is definitely wanting on the structure itself and ensuring that the merchandise we construct, like knowledge platforms — it’s not solely one thing that solves a short-term want but in addition a long-term want, and that we guarantee that we don’t must redo issues over time. I believe that mind-set has been actually essential for us, and it additionally units the course for our groups.

Then the engineering-led for us is quite a bit round the best way that we ship our applied sciences and make them accessible throughout the corporate but in addition outdoors the corporate. We truly strongly consider that the engineers are the specialists. So when a product supervisor or administration comes alongside, they, after all, set the scene by speaking in regards to the “why” and “what” we do issues, however the group is solely liable for the “how,” and so they want to determine, you already know, after we speak about artificial knowledge, “OK, if we have to work with artificial knowledge, then how will we do this in the easiest way? What does a POC [proof of concept] appear to be? What does a minimal viable product appear to be, and the way will we evolve that over time?”

They should determine that out and inform us, after which we, after all, have a look at issues like, how can we ship worth as quickly as doable? And that’s getting these POCs on the market, examined, and ensuring that they ship worth to the group. For us, the engineering-led is de facto essential and one of many key the reason why we now have had the success we now have right now.

I believe we generally additionally must keep in mind that we rent these expert individuals as a result of they’re the specialists and they’re the most effective of the most effective, proper? So empowering the groups is only a key factor to be able to obtain nice success.

Sam Ransbotham: How did you find yourself on this function? I imply, our present is Me, Myself, and AI. I’m curious the way you truly ended up concerned in all these items.

Anders Butzbach Christensen: I believe that’s an extended story. I’ve an academic background inside laptop science and in internet growth, and I truly began out as a marketing consultant seven years in the past. I did cellular functions and web sites and moved into challenge administration of the purchasers that we’d constructed these merchandise for. After which, I believe as so many different individuals in Denmark perhaps dream about working for Lego … we’ve all performed with the bricks, and we dream about working for them.

Sam Ransbotham: It’s not simply Denmark.

Anders Butzbach Christensen: You’re completely proper, and we additionally see that. However completely by coincidence I noticed this job advert that mentioned one thing about “senior product proprietor, large knowledge,” and I most likely didn’t know what large knowledge was, however the “senior product proprietor” I knew, as a result of we have been additionally working with agile inside the consultancy company. So I utilized for the place with out understanding an excessive amount of about what the area was. After which, after 4 or 5 rounds of interviews I, by some coincidence, managed to get the function.

I began out within the Lego Group. I had our suggestions group for Lego.com and Lego Life, which is our social app. Then I did imagery moderation and tagging, to guarantee that we now have protected content material inside our platforms. Then, after half a 12 months, I received the job to digitalize our provide chain, constructing demand forecasts, and so forth., which was an enormous challenge, and I didn’t know something about provide chain, forecasting, or something. After which we determined to speed up our digital transformation, after which I turned head of information engineering.

Once we began out, there weren’t numerous corporations that had tried that earlier than, so we wanted to do numerous discovery and analysis to see how different corporations do this after which attempt to sew the whole lot collectively. And that mainly leads us to the place we’re right now — to the place I’m. And I believe the curiosity round knowledge is what’s saved me right here.

Sam Ransbotham: That’s nice. OK, so that you’ve set me up for this: Lots of the belongings you’re speaking about along with your digital transformation look like you’re pretty mature in them at this level. You talked about, for instance, the artificial knowledge that you just’re engaged on, and the PII identification. What’s subsequent? What sorts of issues are you and the Lego Group headed towards?

Anders Butzbach Christensen: There’s a big journey forward of us. There’s, after all, numerous product groups that use the platform right now. I believe we now have 600-plus knowledge merchandise on our platform. However there are nonetheless numerous digital product groups that aren’t utilizing it, particularly by self-service. There’s a lengthy journey forward of us to construct up the ambassadors across the group and make it as simple as humanly doable to construct these knowledge merchandise that create numerous worth.

Shervin Khodabandeh: You’ve been on this journey for fairly a while and have numerous classes discovered and experiences. In the event you have been to perhaps fast-forward, what could be preferrred? In the event you sit down and say, “Wouldn’t it’s cool if, 10 years from now, when organizations are speaking about knowledge merchandise, they might have X, Y, Z?” What’s “X, Y, Z” 10 years from now?

Anders Butzbach Christensen: What I dream about is that every one knowledge producers expose their knowledge into this platform in order that it’s accessible within the discovery software, and I dream a few one-stop store, after which it’s very easy, it doesn’t matter what talent set you may have, to sew that collectively into a knowledge product.

And for me, that goes for technical groups constructing transformation scripts in Python or one other programming language, to businesspeople who go right into a drag-and-drop software after which immediately it’s accessible in a reporting software. So, what I dream about is that this one-stop store for everybody throughout the group, enabling them to turn into data-driven. And if I look forward, that’s the place we’re in hopefully earlier than 10 years.

Sam Ransbotham: You already know, Shervin, as we step again, we’ve heard from lots of people who’re expressing some related varieties of issues however in several domains. Clearly, not everyone seems to be making kids’s toys … or toys for grown-ups; as Shervin simply talked about, he nonetheless performs with Lego. However we’ve seen individuals in well being care, we’ve seen … ChatGPT developing, and there’s simply numerous applied sciences that we couldn’t even think about 5 years in the past. They’re now publicly accessible within the arms of regular individuals.

Shervin Khodabandeh: Initially, that is fairly fascinating, and thanks for sharing. We’re going to maneuver to a different phase, the place we simply ask you 5 rapid-fire questions and also you give us the very first thing that involves your thoughts.

Anders Butzbach Christensen: Go forward.

Shervin Khodabandeh: All proper; let me truly get the questions. What’s your proudest AI second? Or, let me generalize — knowledge second?

Anders Butzbach Christensen: One among my proudest moments with knowledge was after we truly moved from having numerous several types of knowledge coming in with several types of knowledge qualities that we couldn’t sew collectively. It didn’t create the worth we wanted, as a result of they simply didn’t do issues in the fitting approach. So, what we did was that we constructed a software that educated the information producers on the best way to truly create good, high-quality knowledge merchandise. And that was an enormous success and, I believe, one of many key causes behind that knowledge product turning into successful.

Shervin Khodabandeh: What worries you about AI?

Anders Butzbach Christensen: The concern is just not that it’s going to take over our jobs. I believe the concern is that we sadly by no means can cease studying, proper?

Shervin Khodabandeh: That’s superb. That’s very true. It’s additionally the identical factor you mentioned about, you’re by no means completed with the information. What’s your favourite exercise that includes no know-how?

Anders Butzbach Christensen: For me, it’s bodily exercise. Once I get off work after many hours, I must do one thing that takes my thoughts off knowledge and tech.

Shervin Khodabandeh: What was the primary profession that you just needed? What did you need to be if you grew up?

Anders Butzbach Christensen: First off, I truly needed to be a lawyer, which I discovered wasn’t for me. I’ve an aunt that may be a lawyer, and I borrowed considered one of her schoolbooks, and I checked out it and I discovered that there’s approach too many legal guidelines. I gave up on that pretty rapidly. After which I needed to be a dealer, after which I discovered that that requires taking a look at shares on a display for a very long time, and that wasn’t for me both.

Then I moved into tech and constructing several types of functions, and I believe what I actually discovered cool about IT is you could construct merchandise that create numerous worth, numerous income, with out having to purchase numerous issues. You don’t want a bodily product, and I believe that was one thing that basically received me into IT.

Shervin Khodabandeh: What’s your best want for AI sooner or later?

Anders Butzbach Christensen: My best want is that it’ll make the world a greater place, and I’ll go away it there as a result of meaning … that may be completed in several methods.

Sam Ransbotham: You’re not going to inform us how. You’re going to carry that for the subsequent time we discuss to you, I assume.

Anders Butzbach Christensen: Precisely.

Sam Ransbotham: I believe what’s notably attention-grabbing about right now’s dialogue is, numerous the folks that Shervin and I discuss to speak about Lego as constructing blocks, and so they make an analogy of the issues that they’re doing of their group: “Oh, we’re constructing these Legos so that folks can construct knowledge.” And what they don’t notice is … they suppose they’re speaking about Lego bricks, however they’re truly speaking about the best way that Lego approaches knowledge. I believe that’s fairly fascinating. I believe that’s the form of factor that lots of people can be taught from. Thanks for taking the time to speak with us. [We] admire your sharing this.

Shervin Khodabandeh: Yeah, thanks, Anders. This has been nice.

Sam Ransbotham: Thanks for listening. Subsequent time, we’ll discuss with Rathi Murthy, CTO and president of Expedia product and know-how. Please be part of us.

Allison Ryder: Thanks for listening to Me, Myself, and AI. We consider, such as you, that the dialog about AI implementation doesn’t begin and cease with this podcast. That’s why we’ve created a gaggle on LinkedIn particularly for listeners such as you. It’s referred to as AI for Leaders, and when you be part of us, you possibly can chat with present creators and hosts, ask your personal questions, share your insights, and acquire entry to helpful sources about AI implementation from MIT SMR and BCG. You may entry it by visiting mitsmr.com/AIforLeaders. We’ll put that hyperlink within the present notes, and we hope to see you there.