Every AI Brand Makes a Promise. Can the Product Keep It?

All brands make promises to their customers. But what happens when the products they’re selling can’t always keep them?

That’s the main challenge AI brands face. Every day, AI chatbots and assistants interact with millions of people worldwide, answering their questions, giving them advice and influencing their decisions. Unlike traditional products, the behavior of AI products isn’t always predictable. Every conversation with an AI agent is different, and even the people who built these tools cannot anticipate all their answers.

For brands, these are considerable challenges. An AI assistant portrayed as reliable can still “hallucinate” and make mistakes. One designed to be warm and understanding can respond in a cold or indifferent way. And when this happens, consumers can feel confused and start losing trust in the brand. Because the invisible promise the brand made (“our AI technology is transparent/reliable/approachable…”) hasn’t been delivered.

So, how can brand strategists and managers build AI brands that are trustworthy and coherent when the product itself cannot be fully controlled? How can they ensure the brand’s promise and hierarchy of values are always reflected in the experience people have with the AI product?

Every brand faces situations where it has to choose between different values. A values hierarchy helps make these decisions by defining which values matter most when they conflict.

So how does this work exactly? A values hierarchy is built through thousands of small design decisions. Each choice reflects what the brand considers more important. Over time, these choices shape how the brand looks, feels, and behaves.

This happens quite intuitively: most branding experts create this hierarchy without giving it a name and clients often approve it without knowing it exists. This is why many customers experience its effects without noticing the values hierarchy behind it.

This has always been part of brand strategy. But what has changed today is how much it matters. A brand’s values have always been a promise to its audience. The issue with AI products is that this promise is tested every time someone interacts with the product.

The AI product brief is different.

Brand strategists and designers have always worked on products that affect people. The values embedded in the brand of a financial institution, a healthcare provider, or a government agency have real consequences for the people those organizations serve. The profession has developed ethics frameworks, user research methodologies, and accountability practices that reflect the seriousness of that responsibility.

AI products, however, present a different kind of responsibility, not just a greater one.

For example, when a brand designer works on the identity of an AI assistant, they do more than build a visual system for a product. What they are doing is establishing the terms of an ongoing relationship between a system and the people who will interact with it. That system will answer people’s questions, influence how they understand particular subjects, help them make decisions, and, in many cases, influence choices that affect their health, finances, relationships, and work. And the brand has a big role to play in the definition of this hierarchy of values. With AI products, it is often established long before the first interaction between the AI assistant and the general public. 

Every brand decision contributes to these expectations. For example, the name chosen for the product conveys what kind of system people are interacting with. The visual identity suggests behavior they can expect. The tone of voice communicates how the system will treat them.

These choices go far beyond aesthetics and influence how millions of people interpret what the AI says and does. Brands that communicate warmth via their identity create an expectation of care. The ones that communicate authority are expected to be knowledgeable. Those that communicate transparency are trusted to remain honest, even when the system is uncertain.

AI products don’t work like traditional products. We test our expectations of AI assistants and agents directly and repeatedly through every interaction. Every answer, recommendation, and decision becomes an opportunity for the system to either fulfill or break the promise its brand has made – and brand owners have little control over how the product will react. 

So the question for branding professionals building AI brands is: are they fully aware of the brand promises they are making when designing the brand’s strategy and identity?

The hierarchy in practice

Let’s take the example of an AI assistant that is branded to feel warm and approachable. The branding team chose a soft visual language, a conversational tone, and a friendly brand name. The overall impression is that the system is on your side. That brand has established a specific values hierarchy. And while those priorities feel like aesthetic choices when you’re making them, in reality they function as governance once the system is deployed.

In this scenario, imagine the AI gives a perfectly accurate answer one day, but the way it responds feels cold or indifferent. Even though the information is correct, the user might still feel let down because the experience doesn’t match what the brand led them to expect. If the brand has built its identity around warmth and care, people will naturally expect those qualities to come through in every interaction. When they don’t, it feels like a broken promise. Users don’t really separate the brand from the AI they’re interacting with, so to them, they’re the same thing.

The same dynamic works in reverse as well. When an AI brand claims to be highly capable and confident, people expect it to have the answers. If one day the system admits it doesn’t know something or expresses uncertainty, users may see it as a weakness rather than a sign of honesty. 

It can be interesting to look at real-life examples, and how they work on their hierarchy of values: 

• Anthropic published Claude’s constitution, a document that explains how the system should handle situations where values conflict. The constitution states that Claude should be safe, ethical, compliant with Anthropic’s guidelines, and genuinely helpful. What makes this interesting is that Anthropic openly acknowledges these values won’t always align and explains how the system should decide when they don’t. Many brands establish this kind of hierarchy without ever making it explicit, but Anthropic chose to clarify it and make it public.

• OpenAI has taken a somewhat different approach with its Charter, which sets out the company’s broader mission and principles, including building AI that benefits humanity, preventing serious harm, and prioritizing long-term safety over short-term profits. From a branding perspective, though, we can see that little of that language comes through in ChatGPT’s identity. While the Charter focuses on responsibility, safety, and accountability, the product brand focuses more on being capable, accessible, and easy to use. This creates a gap between what the organization says it stands for and what the product leads people to expect, even if users aren’t really articulate it.

Neither company has completely resolved this tension, and perhaps that’s the bigger point. As branding experts working with AI products, we need to approach AI brands differently, with the aim of reducing the gap between what AI brands promise implicitly and how AI systems behave. In the end, brand promises are one of the most important aspects of branding.

Whose hierarchy

The values hierarchy in any brand system reflects the judgment of the people who built it. It is a reflection of their assumptions, intuition, and ideas. None of that is universal. It’s just their perspective built into a system and applied at scale.

This is manageable for most brands offering traditional products or services. The brand serves a defined audience in a defined context. The gap between the designer’s assumptions and the actual user’s experience is real, but it’s contained.

For AI products, the situation is completely different. One brand system governs interactions across every country, every cultural context, every relationship to institutions, and every understanding of what it even means to ask for help. 

What signals trustworthiness to one user may mean something else entirely to another. The hierarchy that feels natural to a design team in one city is now running interactions in hundreds of cities that team has never been to.

The standard brand audit asks whether the brand is consistent, distinctive, and whether it’s reaching the right audience. What I find dangerous in the context of AI products is that those questions don’t consider values hierarchy and often don’t reveal whose assumptions got built in, or whose experiences weren’t considered when it was set. The profession really needs different questions for brand audit briefs.

What the brief actually requires

Brand designers working on AI products are building some of the most consequential identity systems ever. Closing the gap between the profession’s frameworks and the ones this moment requires means asking questions standard briefs don’t include.

When it comes to AI brands, most briefs skip the most important question, which is: “When the system’s values conflict, which one wins?” It’s important to be as concrete as possible and think through all real-life scenarios. For example: “When being helpful means not being entirely honest, what should happen?” Every AI product faces that moment, and the brand needs to have an answer for each most potential situations.

The second most important aspect to consider is accountability: Who set the hierarchy, and how? The people in that room are making decisions for users who were never consulted, and who probably don’t even know a hierarchy exists.

The third is following through. Brand teams should focus on testing how AI assistants behave in real situations. For instance, does the product keep the promises the brand makes? If the brand communicates warmth, is it warm? If it promises transparency, does it admit when it doesn’t know? That gap goes beyond being a branding problem and affects the organization as a whole.

The question the profession should be asking

Brand designers have spent years learning how to embed values in things. For example, a typeface can signal warmth or authority, a color’s temperature can set the emotional tone, and a name can set expectations for users. These decisions happen before anybody has interacted with the product or read a single word about the brand. 

That expertise is exactly what this moment needs, but it requires more honesty about what the hierarchy actually contains – and more accountability for what it’s promising. The stakes are higher than the profession has ever faced.

The values hierarchy has always been part of the brand’s strategy and identity. For most of the profession’s history, it was possible to leave it implicit. But AI products change that. Now, the profession that understands how that hierarchy is built should be the one asking what it should contain.

That conversation is overdue.