What does the IKEA AI Assistant do?

Published on the OpenAI GPT Store in 2024, it recommends furniture based on the room, style, and budget, generates renderings, and identifies a product from a photo.

Modest, learning-oriented usage: about 1,500 users per month three months after launch, 20% referral to ikea.com, and 5% transactions on those visits.

← The index

Proof C Mixed signals

IKEA

brand assistant published on a third-party GPT store

IndustryRetail & e-commerceLeverAcquisitionFamilyConversationImplementationMartech platformStagediscovery

environ 1 500/mois

Monthly users, three months after launch

"clocked 1,500 users a month" S2

IKEA published its genAI assistant on the OpenAI GPT Store in February 2024: three months later, it had about 1,500 users per month, of which 20% led to a visit to ikea.com and 5% of those visits to a transaction.

Objective

Test the GPT Store as an acquisition and inspiration channel, by giving personalized furnishing recommendations, while learning what it takes to build a useful marketplace GPT.

The deployment

IKEA published its IKEA AI Assistant on the OpenAI GPT Store on February 5, 2024, for GPT Plus users in the United States. The assistant gives furniture and decor recommendations based on room dimensions, style, budget, and sustainability criteria, can generate rendering images, and can identify a piece of furniture from a photo. Ingka frames the initiative as a pragmatic, learning-oriented move. Three months after launch, Ingka's chief data and analytics officer told Digiday about 1,500 users a month. Twenty percent of those interactions led to a visit to ikea.com, mostly toward outdoor furniture and sofas, and 5% of those visits ended in a transaction. The exercise mainly served to test search behavior and to see whether the GPT Store could be an acquisition channel, before integrating the technology directly into IKEA's own surfaces.

Results Proof C

environ 1 500/mois

Monthly users, three months after launch

"clocked 1,500 users a month" S2

20%

Interactions leading to a visit to ikea.com

"Twenty percent of those interactions have resulted in visits" S2

Visits ending in a transaction

"Five percent of those visits resulted in a transaction" S2

Official Ingka press release on the launch and usage figures given by Ingka's chief data and analytics officer to Digiday, established press naming the brand. The volumes are modest and the claimed use is that of a test, which is sourced but does not go beyond C.

How it works

Documented architecture

The stack in detail

plateforme OpenAI GPT Store distribution store for the brand GPT, restricted to ChatGPT Plus subscribers at launch
llm LLM OpenAI sous-jacent aux GPTs OpenAI generative model powering the store's GPTs; exact version not specified in the sources
outil Generation et reconnaissance d'images (outillage OpenAI) furnishing renderings generated and identification of a piece of furniture from a photo, via the native capabilities of GPTs
infra Catalogue et site ikea.com product, style, and budget reference on the IKEA side, with tracked referral to the site for availability and purchase

How it runs, concretely

For ops teams

CadenceReal time at each conversation, in the environment of a third-party GPT store.

Operated byIngka's digital and data team, which configures the GPT and tracks usage and referral metrics.

1

Expressing the need customer
The user describes their room, style, budget, or sends a photo of a piece of furniture to identify.
2

Recommendation AI
The assistant proposes IKEA products, generated renderings, and selection criteria.
3

Referral to the site customer
The user clicks through to ikea.com to check availability and buy.
4

Measurement and learning data team
The team tracks users, referral rate, and transactions to judge the channel.

The signal that drives it

The referral to ikea.com and the transaction that follows. Because the conversation lives with a third party, IKEA sees only a partial signal of the journey and depends on the store for distribution.

How your customers perceive this type of use

Sourced studies

Les consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.

64%

Consommateurs qui prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (2024)

53%

Consommateurs qui envisageraient de passer a un concurrent s'ils apprenaient que l'entreprise prevoit d'utiliser l'IA pour le service client (2024)

pres de 75%

Consommateurs qui veulent savoir s'ils communiquent avec un agent IA (2024)

Acceptance conditions

Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)

Red lines

Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)

Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025

See full acceptance: by country, by use, by generation

How to replicate

Inference, not sourced

Data prerequisites

Product catalog and availability exposable to a GPT
Reference set of styles and criteria (budget, sustainability)
Tracking of referral to the site and of the transaction

Org prerequisites

Acceptance of distributing via a third-party platform
Governance framework for conversation data
Channel metrics (users, referral, transaction)

Possible stack

Brand GPT on a third-party store
Connection to the catalog and availability
Rendering image generation

Team to operate1 digital PM + 1 dev (catalog connection and tracking) + legal/data validation

The plan, step by step

Step 1

Frame the GPT's scope: brand tone, use cases covered, guardrails, and referral rules to the siteDeliverable: Approved GPT brief (instructions, examples, limits)
Step 2

Build the GPT and connect it to the catalog (products, styles, budget), with tracked outbound links (UTM)Deliverable: Working GPT tested internally, instrumented links to the site
Step 3

Validate legal and data (use of conversations by the third-party platform), then publish on the storeDeliverable: Published GPT + documented data governance framework
Step 4

Track users, referral rate to the site, and transactions from the referralDeliverable: Channel dashboard with the three metrics
Step 5

Judge the channel on volume and transaction before investing more (integration into own surfaces or shutdown)Deliverable: Documented go/no-go decision

First step: Publish a brand GPT connected to the catalog, instrument the referral to the site, and judge the channel on volume and transaction before investing.

Evidence

LevelProof C

LivenessMixed signals

The last solid public signal dates from May 2024 (Digiday), with modest volumes (about 1,500 users per month) and a use described as learning. Ingka then mentioned a future integration of the technology into IKEA's own surfaces, leaving it uncertain whether the assistant was still on the GPT Store in 2026.

Dates

Launch2024-02-05

Last signal2024-05

Verified on2026-07-11

Stack / platform

OpenAI GPT Store

Transposable to EU

oui avec adaptations

Publishing a brand assistant on a third-party GPT store is transposable anywhere the service is available. In the EU, the AI Act requires transparency about the use of an AI, and entrusting the assistant to a third-party platform raises the question of the transfer and use of conversation data, to be framed under GDPR.

Who drove it

inIKEA

Parag Parekh Global Chief Digital Officer, IKEA Retail (Ingka Group) S1
Francesco Marzoni Chief Data and Analytics Officer, Ingka Group S1

Agencies & partners

OpenAI Plateforme hote (GPT Store) et modele sous-jacent de l'assistant S1

Visual proof

Dated screenshot pending (PeekShot)

Sources

S1 IKEA launches new AI-powered assistant in OpenAI GPT Store Primary ingka.com · 2024-02-05 · accessed 2026-07-11 archive pending
S2 Ikea launched an AI assistant earlier this year. Has it actually driven sales? Established press digiday.com · 2024-05-21 · accessed 2026-07-11 archive pending
S3 Ikea launches generative AI assistant Established press digitalcommerce360.com · 2024-02-06 · accessed 2026-07-11 archive pending

An error, newer info, a source?

This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.

Suggest a correction