Chief Data & Analytics Officer at largest IKEA retailer says AI can design your home, but humans make it feel like one

Across the conversation, she explores the tension between speed and responsibility in AI adoption, the shift from reporting data to truly influencing decisions, and why building data literacy across the business, from leaders to frontline co-workers is becoming just as important as the technology itself.

The episode also touches on the realities behind scaling AI in a global organisation, the importance of human judgement alongside algorithms, and what it really takes to turn data into something meaningful for customers, not just measurable for the business.

Screw It is a podcast from Ingka Group. New episodes are available on Spotify. 

About the Podcast

Screw It is a new podcast from Ingka Group, the largest IKEA retailer, exploring the “art of assembly” in business, sustainability, and life at home. The series invites global experts and leaders to discuss how we piece together better homes and societies—even when life looks nothing like the manual. From a company that wants people to sit comfortably, but recognises that progress is often uncomfortable, Screw It ditches the corporate script to embrace the “wonderful mess” of building a better future.

About Ingka Group

With IKEA retail operations in 32 markets, Ingka Group is the largest IKEA retailer and represents 87% of IKEA retail sales. It is a strategic partner to develop and innovate the IKEA business and help define common IKEA strategies. Ingka Group owns and operates IKEA sales channels under franchise agreements with Inter IKEA Systems B.V. It has three business areas: IKEA Retail, Ingka Investments and Ingka Centres. Read more on Ingka.com.

Screw It episode 4 – Conversation transcript

Dani: Welcome to Screw It, a podcast by the world’s largest IKEA retailer. Today I’m joined by Burce Gületkin, the Chief Data and Analytics Officer for IKEA Retail, to talk about AI, what’s working, what isn’t, where the hype is getting ahead of reality, and what it actually takes to make AI responsible in a large business organisation.

Dani: Hi, Burce. So nice to have you here. Thanks for joining us.

Burce: Great to be here. Thank you, Dani.

Dani: Before we dive into the world of AI, I want to know a little bit about what drew you to this type of role. Tell me a little bit about you.

Burce: I was born and raised in Izmir. It’s a small town in Turkey – actually the third biggest city, so not so small compared to Europe. My parents are teachers, and I have a little brother five years younger than me. At some point, when I went to university, there was a debate about whether I could go out and study in Istanbul or stay in Izmir. There was a newly established university there, and my parents, being teachers, said they couldn’t afford for me to go to Istanbul, so I had to study there.

Burce: I started studying management, with a lot of resentment actually, because I found it very boring. But because the university was new, they were able to offer me different courses. One professor said, “You can actually try different things here. That’s an opportunity.” I’m a naturally very curious person, so I went into computer science, mathematics, and started taking lessons from different programmes. I had a very wide starting point.

Burce: That curiosity led me through many hoops. I started my working life in finance, just because I was good at numbers. From finance, I moved into commerce and customer experience. I led data-driven marketing and campaign management, and from there moved towards data science. That combination of education and work experience let me move from one interesting space to another.

Burce: When I found myself within the world of data analytics, I felt like I had found my calling. It was really like, wow; this is a space that continuously evolves. You never stop learning. You are kind of behind as soon as you stop following through and updating yourself. That made me very happy in this space. Then I did data and AI in financial services, afterwards in CPG, and now in retail at IKEA.

Dani: I think you’re used to change, right? You’ve tested different things. Coming to IKEA is also the next step. Like you said, you’re new to retail, but you have everything you need to do it.

Dani: What is something people don’t know about you that they wouldn’t expect from a data analytics officer?

Burce: I love karaoke. I love singing. I’m not too good at singing, but I sing.

Burce: It’s funny because although I’m super concrete in my decision-making, and very much based on data, I also trust my gut very strongly. Maybe that’s also unexpected from a leader like me. I look at all the proof points and everything, but in the end, I also go back to my gut. It needs to feel good.

Dani: So, you mix both. You have the information, you look at it analytically, and then you have your gut, and both show you the way.

Burce: Yes.

Dani: That’s so nice to hear. And now we have come to AI – the hype of AI in the retail world. You’re leading AI in one of the largest retail brands in the world. What’s something people assume works brilliantly when using AI, and something that doesn’t work?

Burce: I think the hype is really built around perfect worlds. Very controlled experiments. They say a certain percentage of jobs will be replaced based on controlled experiment design. What people overlook is that enterprises and markets are super messy.

Burce: The data is messy. The processes are messy. It’s not specific to IKEA, but in general, that’s not how we designed the way we work. We designed it for humans. We didn’t design the way we work for AI.

Burce: A job is not done task by task. A job is a collection of tasks. I think people overlook that when they create hype and doom scenarios, all the jobs will disappear. The reality is messier.

Dani: I’ve heard a lot about the spaghetti at IKEA – that it’s a complex spaghetti that needs to be dissolved. That’s how difficult it is when you come into a large enterprise. You have to unravel all of this.

Burce: Yes. I heard about the IKEA washing machine – that you kind of go from one part to another.

Dani: When it comes to insights, you’re gathering so much data about customers and life at home. Where do you think data and insights can make a difference?

Burce: We can know a lot about customers, but from my point of view, what matters most is understanding what drives your life and where you are in your life stage. The more we know that the more we can be personalised towards our customers and increase customer lifetime value.

Burce: There are interesting moments in life when you make big decisions, and we want to be close to customers in those big decisions.

Dani: That’s so interesting because having that much data about people holds so much power. You know so much about them at every stage – from having a young child to when they go to university. With that great power, there’s obviously a lot of responsibility. What do you think IKEA can do to preserve that responsibility and integrity of the data people share?

Burce: There are a couple of things, but I think the most important thing is making sure we don’t use this data to manipulate customer behaviour. That’s the most important thing.

Burce: Being there for you to make your life better and have a better everyday life is one thing. But saying, “Here’s a teenage room, and we will overprice it because we know you’re at that decision point,” is not what we want to do.

Burce: What we want is to use that data to be close to the customer and partner with the customer in that decision moment. That requires different ways of working with data, always protecting the data we have in the most secure way – from a cyber point of view and from an access point of view. But what you do with that data becomes the most critical thing when you have a trusted brand like IKEA.

Dani: That makes total sense. And for our viewers, where do we use AI at IKEA? Where is AI embedded? I know that’s a large question, but maybe from the consumer’s point of view – when you’re buying a product, where is AI embedded in the journey?

Burce: At many stages, actually. Especially if you are interacting with IKEA on the web, you see inspiration from different rooms. That is brought to you based on what we know about you at that moment. We can say, “For this customer here, let’s show this content,” and it becomes more personalized to your preferences.

Burce: If you have a question for IKEA, you will see Billie popping up in the corner, and you can interact. That is based on generative AI and the latest models.

Burce: Throughout the ordering process as well, when you place an order, we have both rule-based systems and goal-based systems to make sure you receive your order based on how you need it. Is it fast? Is it cheaper? We look at your preference and send your order in that way. It’s called goal-based order allocation. From the beginning of the experience to the end, AI is at work.

Dani: Super interesting. And I know there are also design tools where people can design rooms, put in their own furniture, and use machine learning for design. That’s even before the purchasing journey.

Burce: Yes, even before the purchasing journey. I enjoyed that even before I became an employee here. As a customer, it was exciting for me to try IKEA Kreativ on the website and design my own PAX constellation. Recently, I used the kitchen planner to get myself a kitchen, together with co-workers.

Burce: You can also design your home. What excites me most is getting to a point where you can take a picture of the space and explain your mood — the mood you want the space to have – and then we can give product recommendations. I think that’s the next step from what we have today.

Dani: In the retail environment, where does AI struggle more than people think? Where is AI still not there, even though people think AI is giving them all these recommendations?

Burce: I think AI struggles in two ways. One is when the data is not ready – when the underlying data and processes are not there yet or not connected. Then AI struggles a lot.

Burce: The second is something technical called agent drift. Work is not usually finished with one task. You have multiple tasks following each other. If it gets complex, AI can drift and forget what the previous step was or how it should communicate with other agents. Task complexity makes agents drift right now.

Burce: If I compare two years ago to now, it has changed a lot. I think change will accelerate in the coming period. At the beginning, language models couldn’t even do mathematics – they couldn’t do two plus three. Now they can solve really complex problems. Every problem we had a couple of months or years ago is being solved. So, there are struggles, but solutions are coming.

Dani: We’re at a point now where there is so much hype around AI. It has reached a boiling point, and we’re overstimulated with conversations around what AI is providing us. As consumers, we’re thinking, “Now we have generative AI, now companies are giving me agents, now this, now that.” Do you find some of these phrases are just hype? Are people making things very shiny and calling it the next best thing, when actually it might not be convenient or feasible for a company to use generative AI?

Burce: I am all for solving problems in the simplest way possible. Sometimes we have a problem and we ask, “Do we really want to solve it with an agent approach, or is it simply rule-based?” Some problems need to be solved in a very deterministic way, and AI by definition is probabilistic.

Burce: We have to find the problems and apply the right solutions to them. The simplest solution is usually better. AI can get complex in the hype, and I find that unfortunate. It always feels like it is ahead of us, and we need to catch up.

Burce: There are doom scenarios saying AI will make all of us jobless, and the world will change drastically. Then there is the other scenario that AI is our salvation and will solve everything, including climate change. I think I am quite in between.

Burce: Life is messy. Life is not controlled. That is my number one thought. Number two is that we have to make AI work for humans. Humans should not work for AI. We need to make AI work with humans and staying in control of that is really important to me.

Dani: I love this idea that you can be optimistic without being fully optimistic or pessimistic. You’re looking between both and seeing what can work. That’s interesting from a leadership perspective because we see a lot of tech leaders and CEOs talking about what will happen with AI – jobs disappearing, people having to reskill, and it creates a lot of stress for people.

Dani: That leads me to the next question around people. We know that people at IKEA are at the heart of everything we do. But now that AI is becoming more capable, how do we ensure that people stay at the centre?

Burce: I believe AI becoming better doesn’t change the fact that we can put people at the centre of our business. Our most important asset is people and how they connect with customers.

Burce: If you take away the manual work and make the role of the person connecting with the customer, understanding their needs, and guiding them towards better decisions for their life at home, then you can create positive change together with AI.

Burce: Designing with AI, checking out with AI, making recommendations with AI – all of this can happen together with AI. But the genuine connection between the co-worker and the customer becomes more possible and real when manual tasks are taken away.

Burce: As long as we think about working with AI to make jobs better, even if some tasks or jobs are replaced, we should leave the most meaningful work to co-workers and to us.

Dani: Is an example of remote planning for kitchen services? I can go online, design it, and then make an appointment with a design expert to go through my kitchen because I still want that human element.

Burce: Yes. Almost everyone wants that human connection before making a big decision. A kitchen is a big decision. I recently made one myself. You definitely need a human for that decision. Even if you don’t need it at every stage, when you are installing it, there is a human connection. We can’t underestimate that.

Dani: What do you think are great skills people can reskill themselves in to adapt to how AI is changing how we work?

Burce: I have two answers. One is around soft skills – though I don’t know if they are soft skills, but asking the right questions and defining the outcome you want becomes the most critical skill.

Burce: You need to be able to say, “This is where I want to get to, and here are five agents – get me there.” Defining the outcome and asking the right question is very important. Where am I? What do I want to achieve? Being articulate about that cannot be communicated enough.

Burce: The second part is learning to work with AI. We have zero excuses not to learn. Every company in the AI space – Google, OpenAI and others – has learning courses for many people. You even have AI for elderly people, AI for students, and AI for different settings.

Burce: Start working with AI and learning. Start experimenting in your own personal life as well. I always give the example of how I use my agents, almost like minions, to search my Gmail inbox for my kids’ upcoming events and automatically put them on my calendar, including tasks like ordering a birthday gift for a nine-year-old boy for a Sunday birthday party.

Burce: That’s how you learn to build an ecosystem around you with AI and agents that can do tasks on your behalf to make your personal life better as well.

Burce: Thirdly, as an employer, we have the responsibility and accountability to upskill and reskill all co-workers working for IKEA so that they are ready for the next steps in their jobs, or even outside IKEA if needed.

Dani: You touched on something important there. We talked about evaluating data. One challenge today is that we are getting a lot of information through AI, so critical thinking around what we receive is important. How do we make sure that if I’m using five agents at home, what I’m being fed is accurate? What is my responsibility as a person and as someone in a company?

Burce: It’s always fact-checking the outcome. It’s funny because if you fact-check AI and say, “This was wrong,” it will say, “I’m sorry, I apologize,” and then it gives you the right answer if you push a bit.

Burce: There was a meme asking AI how many Rs are in strawberries. It would say two, and then if you say no, it will say, “Oh, I’m so sorry, you’re right, it’s three.” So always fact-check.

Burce: The second thing I do is almost like a trick I use for myself. If I have a very strong opinion about something, I ask, “What is the opposing argument against this opinion of mine?” Then I start to argue with myself to make my decision. That’s how my mind works.

Burce: I do the same with agents. If I have a big question to solve, I create almost like a council of agents representing different personalities. I say, “You are a McKinsey consultant. You are an experienced Chief Data and AI Officer. You are an IKEA customer. You are IKEA’s Chief Commercial Officer. Here is my question – debate it for me and come back with a recommendation and different perspectives.”

Burce: Getting different perspectives from AI helps improve the decision-making process.

Dani: That is such an insightful thing. What I take from you is: one, don’t outsource your critical thinking; and two, look at other points of view and make sure you have variety to make the right decision.

Burce: Definitely. Outsourcing your critical thinking is so risky. Never start your thinking process with AI. Do your thinking process first. That is one of the pieces of advice I would recommend because it creates cognitive depth.

Burce: The brain is a muscle, and like any muscle, if you don’t train it, you lose it. If you start your writing with AI, for example, you stop writing genuinely and authentically. You stop thinking, and your critical thinking process is impacted. It creates cognitive debt.

Dani: That is so fascinating and such a good takeaway. Going back to the organisation, we talked about reskilling and upskilling. How important is change management in getting people updated and ready for what comes next? The moment they get skilled, there are new skills they need to learn. How do you manage that change management process in a large organisation?

Burce: I’ll make a strong statement. I really believe we should go and work together with people. How you can do this work together with AI – I just did a pilot in three countries, and we saw the overlap there.

Burce: If you co-create together with people, and upskill them in the process of co-creation, I don’t think you need a huge change management programme afterwards. You don’t come after and say, “Here is a tool, here is the change management, here is our PowerPoint, this is why we need to change.”

Burce: The big part of the work is show and tell. Be there with people in the redesign of the processes they need to work with. It’s very difficult to do that at scale in an enterprise like this, but we need to start somewhere.

Dani: I think you make such a good point. Change management works when it works for people and helps them through the curve. But at scale, it can become something where you bring something, give it to people, and say, “Learn,” while missing the whole part of working with them and tackling it together.

Burce: Exactly. Doing it with them becomes the most important and critical thing. People don’t know what is possible before they experience it together with you.

Dani: Let’s take you to the future. What will feel completely normal in five years that might feel uncomfortable today when it comes to AI? I’ve heard a lot about robotics in the household helping with chores — something that today feels a little uncomfortable.

Burce: I think interacting with robots in our daily life will become more and more normal, just like interacting with chatbots became normal. People used to feel extremely uncomfortable with chatbots and immediately wanted to connect to a human. Now the tone of voice and connection has become normal.

Burce: I would not really like a humanoid robot in my house today, but probably in five years we will feel comfortable with it, especially if it can do the dishes and laundry.

Dani: The laundry would be amazing. But do you think regulation will be there? Part of the uncomfortable feeling about robots is that we don’t feel legislation is there to protect us.

Burce: At least in Europe, I believe we are catching up on regulation faster than innovation, which I don’t necessarily think is the best thing to do. A lot of people make comments about innovation not going as fast as regulation in Europe. I think it’s a good thing that we are regulating and trying to understand the change this brings to us. At least in Europe, I believe we will take the steps.

Dani: Has working with AI made you more optimistic, or are you still cautious?

Burce: I am cautiously optimistic. I think we will figure out the way. Big changes have happened in history. One example is horse carriages and industrialisation – how horses were carrying people around, and slowly cars came, and then horses left that role.

Burce: That is the scary part. I think horses would probably have thought, “We’ve been here for 250 years doing this job. Cars can’t replace horses.” That is a bit like my inner voice right now.

Burce: But then I also think about the world ahead of us. Things we believe are not possible will become possible. My true hope is that we will continue to find meaning in life in different ways and with different jobs, even if it looks quite different from today.

Dani: Change is the only constant. Change is here, and we need to embrace it. Two last things before I leave you. What should always stay human at IKEA, no matter how good AI gets?

Burce: The connection with the customer should stay human. When we meet customers in the store, I don’t want to meet humanoid robots. I want to meet a real human who is connecting with me and helping me make my house feel like a home. That connection needs to stay human.

Dani: I totally agree. Last question: what is something you hope AI never gets good at?

Burce: Nothing. I hope it gets good at everything, actually. That gives such big possibilities and progress for all of us. I hope it gets good at everything, so we can outsource many things to it – as long as we want it.

Burce: Maybe what I don’t want is for AI to get very good at controlling people.

Dani: So, you want people to have the handle. You want people to use technology for their benefit. AI is at the service of people.

Burce: Perfect.

Dani: Well, I want to thank you again for joining us on this episode of Screw It.

Burce: Thank you.