AI Conversation Branding Study

How does the personality of AI assistants and chatbots shape the perception of brands? Our research shows that AI interactions can substantially strengthen brands.

Status quo

1 in 2 users

say that AI assistants and chatbots from brands feel impersonal and cumbersome

Personality as potential

48 %

are more willing to use them if they reflect the brand personality

AI interaction drives brand strength

78 %

higher brand strength (preference and differentiation from competitors) among users after AI interaction

Personality creates preference

2.5 x

higher willingness to use AI assistants after interacting with an AI with a strong brand personality

Download Study

Study design

For the study, we selected two brands with particularly strong personalities: Red Bull and Nespresso.

Both brands are internationally known and belong to comparable product segments as beverage brands.

The selection was made independently and without the involvement of the companies to ensure the objectivity of the study.

Three versions of a Large Language Model (LLM) based on OpenAI ChatGPT were created and adapted for each brand:

AI #1: no extra brand personality
AI #2: minimal brand personality
AI #3: strong brand personality

All models were instructed to behave like the respective brand – their difference lies exclusively in the type and preparation of the information on brand personality.

As a basis for the following studies, we defined four different dialog situations that are run through with all AIs and compared on the basis of these. The dialogs cover a spectrum of relevant AI chatbot interactions: from product information and questions about sustainability to inspiration.

The answers varied depending on the brand and personality traits.

A preliminary study with marketing experts (n=16) tested whether differences between the dialogues were recognizable – without them knowing which AI version was behind which dialog.

The AI-created dialogs were inserted into a neutral chat interface with the respective brand logo and shown as a screenshot.

The experts rated the dialogs according to brand fit, quality and preference

In an online study (n=1,153), the same AI dialogs were presented to female brand experts. They were then asked implicitly and explicitly about their perception of the dialogs, their evaluation of the brands and their usage habits of AI tools and chatbots.

Each dialog was evaluated by around 150 people and a control group was surveyed for each brand.

To take the interactivity and individuality of AI experiences into account, we conducted a final study in which participants (n=385) interacted live with one of the AI chatbots.

After 1-3 minutes of interaction, they rated how they perceived the brand and the chat experience.