From Smart Designs to Delivered Homes: My AI Strategy Lab Reflection

No ratings yet. The AI Strategy Lab team assignment led my team (Team 28) and me to come up with an extra GenAI layer on top of the already existing IKEA Kreativ software. This extra layer turns choice overload into a shoppable room design and is called IKEA RoomMate.

My contribution included the value side of this project for both IKEA and its customers and the future-market positioning. At the start of the project, I treated IKEA RoomMate’s AI as the source of value. However, I ended up seeing it more as an amplifier of IKEA’s existing assets.
When it comes to added customer value, RoomMate reconnects the evaluation and choice stages with the actual purchase, whereas with the predecessor, Kreativ, customers would drop out at the evaluation and choosing stages. Also, RoomMate builds on Kreativ and can be seen as a form of incremental bridging. RoomMate also has a competitive advantage against alternative GenAI models, as it encompasses the entire customer journey from designing a room to creating an actual shoppable list of real IKEA items and providing delivery options.

One of our key decisions in this project was to start from the customer’s perspective and not that of the technology. RoomMate changes IKEA’s existing workflow in terms of the customer’s journey from idea to purchase and IKEA’s internal processes. Looking back, I realise that we mainly focused on the changes related to customer journey and not really the internal changes for IKEA, such as new roles, data reviews and ownership. In addition, when I started working on the future-market slides, I based RoomMate’s differentiation around personalisation. However, my argument was changed after some research on the topic to IKEA’s defensible advantage being the integration with the products, inventory and delivery.
What surprised me was that a tension had occurred in my own slides. I presented cross-selling and bigger baskets as value for IKEA, while RoomMate both designs the room and sells the products. Because of this I realised that customers will not trust a recommendation tool that mainly helps the seller. That is why I focused on trust, transparency and customer control. In hindsight, I realise that a weakness in my part was that the value claims mainly relied on sources that focused on AR, rather than proving that customers’ purchases would increase due to an AI generated design.

Feedback also caused some changes in our work. The general feedback in lecture 4 mentioned that many teams explained value creation well, but value capture poorly. This applied to us as well. RoomMate is free, so I had to make clear how IKEA actually earns money, which is through extra furniture sales and not the tool itself. Furthermore, my value claims were based on research that showed customers who feel more confident are more likely to make a purchase. Lecture 6 taught me that this type of evidence shows a link, but not a cause. That is why our recommendation is an A/B pilot with a clear rule, which is that IKEA only scales RoomMate if conversion rises by at least 1.2 percentage points. Also, the AI canvas helped me see that RoomMate only predicts a good design. The customer is still the one deciding whether to actually purchase it. Therefore, RoomMate only creates value if customers trust the design enough to act on it.

The collaboration within our team was very pleasant. When doing the assignment, I realised that each of our sections depended on earlier sessions. This meant that we had to have continuous and clear communication about our personal and communal progress. Different slides had to align with other slides, such as the slide about trust and the slide about governance. For example, both slides mentioned the risk that customers could doubt whether RoomMate picks products for them or for IKEA. We agreed that the slide about governance would explain the risk and safeguards, while my slide would explain why trust is necessary to compete with other AI agents. We made sure to keep the assignment coherent by discussing it in meetings and checking each other’s work. We all kept track of what each of us did and provided constructive feedback, so that everyone stayed on the same page. From this, I learned that a strong strategy consists of different parts that tell one consistent story.

The strategy lab assignment also changed my understanding of AI and business strategy. Before this project, I thought that AI would be the main source of value. An article about strategy in an era of abundant expertise showed me that AI can make knowledge and design cheap and available to many. What stays scarce is things that AI cannot copy, such as logistics, customer data or trusted relationships. The advantage of IKEA lies in turning a design into a delivered home, and not in the design itself. Additionally, I learned that personalised tools can be biased when the seller controls them. RoomMate is such a tool. However, because IKEA runs this tool, customers might wonder whose interest it actually serves. In the future, it is possible that customers will end up using their own AI agents to shop. In that case, IKEA will have to earn trust instead of solely offering good technology.

In conclusion, the AI Strategy Lab changed the way I see AI in business. At the start of this project, I thought of RoomMate’s AI as the main source of value. Now I understand that it is an amplifier of what IKEA already has, which is their products, delivery network and relationships with customers. Furthermore, I learned that value claims need real evidence and that customers only act on AI recommendations they actually trust. In future projects, I will first ask myself which advantages competitors will not be able to copy, before asking what the AI can do.