As a creative agency specialising in tech, we help established companies and startups (including those in AI/ML) shape their brands. Naturally, we were curious: could AI develop a brand with minimal human input? The Human Feedback Foundation, a non-profit dedicated to advancing open-source, human-centred AI, was the perfect opportunity to put this idea to the test. Along the way, we encountered interesting challenges and surprises, revealing unexpected twists in our AI-driven branding experiment.
We started with a premise: AI would handle 80% of the process, with our team guiding, providing feedback, and making final selections. Using ChatGPT for ideation, Dall-E and MidJourney for visuals, and Grain for transcription, we set out to explore AI-driven branding.
Our Process:
1. Creating the Brief
We began by having ChatGPT generate interview questions instead of using our standard questionnaire. The results were promising—we even noted a few questions to add to our core document.
Typically, we conduct brand interviews in person, using the questionnaire as a flexible guide rather than following a strict script. While this works well with people, it proved challenging for GPT. When we uploaded the unstructured Zoom transcript, GPT reduced Human Feedback’s mission to an “experimental project exploring brand creation using AI.” 🤦♀️
To resolve this, we had the client fill in the questionnaire manually. With clear responses, GPT produced a well-organised, insightful brief that became the foundation for brand development.
2. Concept Development
The next step was concept development. Usually it’s a multi-day team effort involving research, brainstorming, and refining ideas for client presentation. We used GPT to generate high-level ideas and picked the best ones.
The initial concepts—AI Co-Pilot, Echo, and Human Blueprint—focused on collaboration, essential for creating ethical, human-centred AI. While Echo resonated, Co-Pilot and Blueprint felt too conventional, prompting a shift toward concepts that emphasise diversity and open-source values. The client highlighted the need for open AI, cultural diversity, and viewing dialogue as an input/output process, setting the stage for a more refined approach.
Refined Concepts
In response to client feedback, we developed and tested four new concepts. This time, we used AI to evaluate each concept’s audience appeal, risk of clichés, and alignment with the brief—tasks typically handled by an art director. Once again, the best ideas were sent to the client for their feedback.
Final Concept
Ultimately, the client chose the combined Infinite Dialogue/Open Network concept, which blends continuous interaction, cultural diversity, and open-source values. With GPT’s help, we refined the ideas further and crafted final prompts for MidJourney to develop the visual identity elements, creating a solid foundation for the brand’s look and feel.
At this stage, we arrived at some solid ideas that became the basis for the visual identity, successfully guiding the AI-assisted design process.
3. Creating Visual Identity
With the chosen concepts, we moved into the visual identity phase, focusing on logo creation and colour palette development.
Logo Creation with AI
Initially, we let Dall-E and MidJourney generate logos based on minimal instructions, producing results that felt stock-like.
We refined the output by adjusting prompts to spell out qualities essential for professional logos, such as simplicity, versatility, and scalability.
Idea №1: “Enhanced Probability Dialogue Brand Concept”
This option merges Infinite Dialogue and Open Network, visualising AI’s dynamic, interconnected nature. Using metaphors like fluctuating graphs and evolving connections, it highlights the complexities of AI learning while appealing to tech professionals through themes of continuous human-AI interaction and collaborative diversity.
Client Feedback: “We like the dynamic feel, but it could use more clarity. The graph elements should better reflect human contributions without feeling too technical.”
Idea № 3: “Open Source Community Brand Concept (Revised from Open Source Synergy)”
This idea centres on community in AI development, shifting from “synergy” to collective contribution. The logo uses symbols of collaborative workspaces, forums, and open-source contributions to represent shared efforts in advancing AI.
Client Feedback: “These aren’t harsh but lack clear meaning; the second feels incomplete, and the third option seems too generic.”
Idea №3: “Shift the Distribution Network Concept”
This concept blends the idea of a network with the dynamic shifts in power and probability within AI. Rather than a traditional mosaic, the logo uses elements like shifting graphs, variable node sizes, and evolving connections to illustrate the diverse contributions shaping AI development.
Client Feedback “Thin lines and complex shapes generally don’t work well for logos. The first design feels off due to its symmetry, which doesn’t convey the “shift the distribution” idea—it seems controlled and locked, more like a decorative key. The other two miss the concept entirely, while the last is simpler but still too generic.”
Colour Palette Experiments
We also tested AI Color’s capabilities for colour palette generation. Although AI produced some interesting combinations, they often required human refinement for originality.
Sample Prompts for Colour Palettes
Key Visual Experiments in Midjourney and DALL-E
We also generated key visuals using Midjourney and DALL-E to explore AI-driven concepts. Midjourney excelled in expressive, abstract imagery, while DALL-E offered more precision and control for realistic, structured visuals. Both tools produced compelling images, but human refinement was often needed to achieve originality and brand alignment. Despite using specific styles, prompts, and references, we didn’t reach a result adaptable enough to serve as a core brand asset.
Designer’s Feedback
Abai:
Besides generating logos, we had the chance to test tools for creating colour palettes and key visuals. While the colour palettes inspired future work, the key visuals didn’t add much value—they were attractive but lacked meaningful content. AI is here to stay, and I see it as an accelerator for generating new ideas. Like any accelerator, AI will eventually blend into the background, becoming an invisible tool rather than a replacement for human creativity.
Challenges and Human Intervention
After multiple rounds of AI-generated designs, the client made a new request: bring in a human designer to fully capture their vision. In an ironic twist, our experiment—intended to push AI to its creative limits—ended up needing the very human touch we’d set out to minimise. Unexpectedly, bringing humans back in the loop aligned closely with the Human Feedback Foundation’s mission, reinforcing the value of human touch in AI-driven branding.
That said, the AI-generated visual identity, however, served as a strong reference for the final branding.
To see how this experiment shaped the final brand explore the full case study here.
Conclusion: AI as a Creative Partner, Not a Replacement
Our creative director joked that working with AI is like working with an intern who isn’t a designer but knows Photoshop. While AI excels at brainstorming (though some of the ideas can feel somewhat surface-level) and speeding up design, it lacks the nuanced insight needed for a truly unique brand identity. Our experiment showed that AI can suggest ideas, but human creativity remains essential in refining them.
Elena Yunusov, Executive Director, Human Feedback Foundation:
We’re grateful to Oddbee for the opportunity to participate in this unique experiment. We’ve learned that while AI can be a fun creativity amplifier, the human-in-the-loop approach is essential to getting the most from AI systems
The Future of AI in Branding
As AI tools continue to evolve, they hold exciting potential for reshaping the branding landscape. With their ability to quickly generate ideas, create initial visual assets, and streamline parts of the design process, AI tools are becoming indispensable partners in creative work. However, our experiment highlights an important truth: while AI can lay the groundwork and offer a fresh perspective, it’s the human touch that transforms these concepts into memorable, resonant brands. For now, AI serves best as an invaluable assistant—an enhancer of human creativity rather than a replacement. Looking ahead, we see a future where AI and human ingenuity work hand in hand, pushing the boundaries of what’s possible in branding. The challenge will be finding the right balance, ensuring AI complements the artistry and intuition that only humans can bring to the table (at least for now).
