When AI Speaks, Design Matters!
- comevis
- 14. Nov. 2025
- 3 Min. Lesezeit
Technology is Becoming Simpler, Design is Becoming Crucial
Many people are initially enthusiastic about AI, but then become disappointed. 🙎 Not because AI is bad, but because it interprets language and does not read minds. 💭 If you phrase things unclearly, you get generic results. If you describe things precisely, you get substance. The decisive factor is not the choice of model, but the design: setting goals, clarifying requirements, and defining quality.
Design beats technology. 💪Technology is developing rapidly. Impact is created through design. Those who formulate goals clearly, consciously shape preferences, and treat communication like a product deliver reliable results.
AI Functions Like a New Colleague
A helpful analogy: Think of AI as a new colleague. 👩💻 She is motivated and capable, but she doesn't know your context. Good leadership means explaining goals, frameworks, and expectations in an understandable way, giving examples, and allowing questions. Bad leadership assumes prior knowledge and prevents questions. AI reacts in the same way. Clear tasks produce good results. Implicit assumptions lead to frustration. Communication thus becomes a key skill, 🔑 both with people and with machines.
The Invisible Default Settings of the Tools
Another point that many overlook: Tools such as ChatGPT come with default settings.⚙️These consist of predefined prompts, training data, and interaction patterns. They shape the results, even if they are invisible.👀
That's why the first question at the start of any project should be: Do these default settings suit my project, or do I need to actively design them?
This includes consciously choosing the database, defining the system prompt, and clearly specifying the desired response behavior.
❌ Default Settings
Unknown training data
Standard prompt
Preset response pattern
✅ Custom Design
Own database
Defined system prompt
Customized response design
Response Design as a Success Factor
Successful applications show that it's not just the size of the model that matters, but also comprehensible responses. An example: ChatGPT is convincing because the response pattern is familiar and clear.
It follows a sequence: first a confirmation, then the content, followed by a follow-up question. 🔄 This pattern reduces cognitive load and makes the interaction pleasant. 😊 This is design work, not coincidence.
Three Steps for Greater Impact
What does that mean in practical terms? Three points help in all areas, whether product, marketing, HR, sales, or service:
🎯 Clarify the goal. What exactly are we working toward, and what is explicitly not the goal? Focus beats feature lists.
🔍 Check default settings. Don't blindly accept the database and system prompts; instead, version, test, and adapt them.
📝 Briefing as you would a new team member. Goal, context, examples and counterexamples, rules, definition of done. Explicitly allow questions.
“AI can do a lot—but it only has an impact when we design it. Design is not an accessory here, but the decisive control instrument.”
Simon Strauß, comevis
Project Manager & CX/UX Designer
Why AI Needs Leadership and Brands Need Design
Those who manage AI clearly not only get better answers, but also consistent brand experiences. The key lever: clear communication creates a consistent customer experience – across language, voice, and sound. 🔊
This is exactly where our approach comes in. Under “Brand & CX Building – audibly better!”, we combine audio brand management, brand voices, and conversational design into an integrated understanding of communication.
comevis empowers marketing, CX, service, and AI managers to make communication clear, recognizable, and impactful - across all channels.
Because strong AI interactions don't come from technology alone, but from conscious design. 🧠🎨
👉 Impact happens when AI is guided. comevis is here to help you on this journey.
