Base Design is a global network of creative studios with outposts in New York, Brussels, Geneva and Melbourne. It works on strategy, design and technology projects for clients like 12 matcha, Bellerose and Théâtre du Châtelet.
Here Mirek Nisenbaum, partner and digital director at BaseNYC, explains how they use AI, and how he feels about it.
You can see all the articles in this series here.
Broadly speaking, are you excited for how AI will change the design industry, or nervous?
Excited, without hesitation. Design has always evolved alongside new tools, from the printing press to Photoshop to Figma, and AI is another chapter in that story.
What excites me most is its potential to push us into unexpected territories, to help us reframe what concept and craft can mean today.
Of course, there are risks– homogenisation, over-reliance, ethical pitfalls, even skills atrophy. I see those as challenges to approach with eyes wide open.
Our job is to bring discernment, taste, and critical thinking to the table, ensuring AI pushes us beyond the common denominator rather than dragging us down to it.
Do you have an agreed policy around AI as a business?
Yes, and it was one of the first things we worked on when AI tools began entering creative workflows.
We wanted clarity for our teams and reassurance for our clients, so we built a policy around transparency, human oversight, and protecting client IP.
The process was collaborative – we convened leads from cross-functional teams, mapped out the risks and opportunities, and then tested the guidelines against real project scenarios.
The outcome is a living policy that sets guardrails around tool use, security, privacy, documentation, and contract language.
When did you realise AI was going to have an impact on design?
At first it was gradual, seeing designers and artists experiment with GenAI tools a few years back. Results were unexpected, mostly funny, sometimes uncanny, but it was becoming clear that something special was beginning to happen.
Our Brussels team used GenAI as the core concept and execution for a campaign, and we could see its creative potential, even in its rough early stage.
I saw designers begin experimenting with LLMs for language in decks, research, insights, and today some of our creatives are already “vibe coding” tools for design ideation and prototyping.
This progression is making it clear to me that we are starting to see a fundamental shift in how design is practiced.
Have you undergone any AI training, either as a studio or individuals?
We’ve approached training as a mix of structured workshops and bottom-up experimentation.
We kicked off the process by bringing together a cross-disciplinary group of AI champions, a few people from every Base studio, representing different areas of expertise. Some were advanced early adopters, while others were just coming on board with curiosity and an open mind.
This group was tasked with identifying needs and blockers, and with activating adoption and experimentation across studios and disciplines.
We ran internal sessions focused on mapping workflows and pinpointing where AI could accelerate work and where it risked making it generic, from admin to creative.
We partnered with several AI-focused companies to dig deeper into workflows and gain first-hand experience with tools, guided by the teams building them. We also invited outside facilitators to introduce prompting techniques and best practices, helping our teams approach AI as a true creative partner rather than just a set of new tools.
Our hope is that the impact will be cultural as much as practical – people feeling more confident, more critical, and more willing to experiment.
How do you use AI in the studio’s creative process?
It’s the challenge we are currently navigating.
I see AI’s real promise in exploration, helping us push our thinking, test ideas in multiple directions, simulate scenarios, and prototype quickly.
The easiest uses today are more basic – asset production or workflow assistance, they don’t challenge us conceptually.
We always felt that words are as important in our creative work as images. Now with AI, the function of language as a model of the world we are building has become both amplified and more accessible for everyone on the team.
We can think deeper, surface obvious ideas quickly, use them as bounce-off points, and uncover unexpected ones.
Do you think clients care if/how you use AI in your work?
Yes, but in different ways. Some clients are curious and want to understand how AI might accelerate their own processes, while others are primarily concerned about IP protection.
What’s consistent is that transparency helps. When we explain our policy, how AI supports but never replaces judgment, it builds trust, and in some cases even strengthens our advisory role.
Do you use AI for any non-creative aspects of running your business?
We have an integrated engineering group, Base Digital, and unsurprisingly that team has been among the most active and immediate users of AI as a coding assistant.
They’ve seen direct gains in productivity, while also becoming early experts in agentic workflows, systems we can now apply both to our internal processes and to client projects.
Beyond engineering, we’ve been piloting AI in business development, proposal drafting, and internal knowledge management. For example, we’ve connected our shared drives and Slack channels to a private GPT, allowing teams to quickly surface insights from precedents, case studies, and research materials.
It saves time, accelerates onboarding for new team members, and ensures knowledge doesn’t get lost.
Operationally, AI also streamlines meeting notes, summarises transcripts, and highlights insights we might otherwise miss. In this capacity, it’s already proving highly effective.
Beyond the best known tools, what is one AI tool that you would recommend to other design studios?
We’ve been testing and piloting a lot of tools.
A few of the favourites currently are – NotebookLM for research and learning, Granola for note taking, Node based tool Flora for design and multimodel experimentation, well-known ones Midjourney (be aware of image rights) and Adobe Firefly (when copyright is a must) for images, Runway for video.
We also use Flux Kontext, Fal, Krea and Cursor for coding, Magnifique and Topaz for image upscaling.
Each has its niche, but what keeps us on our toes is how quickly the ecosystem is evolving.