Why We Decided to Sell Addition to R/GA

The creative industry has split into two camps. Most are treating AI as a productivity booster—sprinkling magic AI pixie dust onto existing processes to show efficiency gains and cost savings. This approach is appealing because it's practical and immediately useful. You don't have to change how you work, and the benefits are clear. But while necessary (if you don't use AI to work faster and more efficiently someone else will), it's only half of the equation. This industry's fixation with this approach risks a race-to-the-bottom that puts legacy agencies at odds with tomorrow's next AI startup or newly released model capability.

The alternative is harder: designing AI systems that orchestrate foundation models from the ground up. This creates new value by solving problems in ways that weren't possible previously. But it also requires a fundamentally different approach and requires clients to change the way they engage and operate. It's infinitely harder, but in the long run it will be much more disruptive as we enter the intelligence era. This is the bet we've been making since Rick and I started the company four years ago, and as we began exploring a potential acquisition with R/GA earlier this year, we discovered it's a vision we share.

My history with R/GA goes back to the beginning of my career. As a young producer in Minneapolis, I was obsessed with their Nike ID launch. I applied for a job, and when I eventually heard back from them I had already taken a role at CP+B. Luckily for me, CP+B's interactive department was run by R/GA transplants who schooled me in the "R/GA process" and how to understand software systems from the bottom up. It's also where I first worked with Tiffany Rolfe, R/GA's CCO who I've now known and collaborated with on-and-off for 20 years.

When Rick and I founded Addition, Tiffany was one of the first people we shared our work with. We stayed in touch over the years, sharing R&D and demoing the AI systems we were creating for clients. When R/GA spun out of IPG earlier this year, we began exploring a deeper partnership.

The alignment was natural. R/GA has been pioneering creative systems for decades. Take Nike ID, for example—they transformed a basic shoe customization tool into a creative system that turned customers into product designers. The system orchestrated user choices across colors, materials, and personal messages into a seamless experience that encouraged experimentation and self-expression. As generative AI has emerged, they've applied this systems thinking to AI with "Abandoned Nights" for AB-InBev, generating personalized videos for cart abandoners, and the new Androidify platform that uses AI to create custom Android characters. It's the same fundamental approach we've been developing: orchestrating technology, creativity, and data to create new experiences, not automate old ones – a philosophy and approach that we've gotten to dial in over the past four years.

Through painstaking trial and error, we've learned that the systems that produce truly creative outputs—the kind that don't scream "this was made by AI"—require creative orchestration. Instead of one magic prompt, you build dozens of inference functions working together, each one leveraging the specific strengths of different models and techniques. The AI system we created for Google Shopping, for example, combines location data, search insights, image generation, and content evaluation into a single system that creates hundreds of unique, contextually relevant ads. A similar system for Google Cloud MLB used a slightly different approach to AI orchestration including use of classical AI and ML models, and live display of data via web wrappers. Small changes in technique, sequencing, or model selection can dramatically impact results.

What's surprising is that this fundamental approach has remained consistent even as the underlying technology rapidly evolves. Techniques like retrieval augmented generation (RAG), chaining, and agents are architectural strategies we've deployed across hundreds of projects. The specific inference techniques evolve, and the models get better, but the orchestration principles stay the same.

This approach also demands new types of talent. We've had to invent roles like AI Content Design—a hybrid of prompt engineering, vibe coding, and content strategy. Engineers become creative writers. Creative writers become engineers. Most importantly, everyone needs to stay insatiably curious, hands-on, and adaptive to new capabilities that keep emerging. Teaming up with R/GA and their global network of talent, spanning strategy, design, creative and technology, enables us to explore new configurations in the 'human' part of our AI system design process, unlocking new solutions and offerings for our clients, and enabling us to take on bigger challenges.

As AI systems become more powerful, they make it possible to achieve scale without compromising on quality or creativity—and with far less operational complexity.

The X factor is how AI Systems can unlock new strategies for how "scale" shows up in the world to consumers. For years, scale has been a dirty word in our industry—synonymous with generic, templated, mass-produced content that people actively avoid. AI systems make it possible to achieve scale in a way that feels more intelligent—through personalization, localization, cultural relevance, and real-time responsiveness to the world around us. Scale that ultimately feels more human.

We see a future where, together with R/GA's global infrastructure and enterprise relationships, we can transform what 'scale' means and how it shows up in the world in a more intelligent and human-centric way.

In the intelligence era, brands don't have to choose between boutique expertise and enterprise scale. The teams that thrive won't be the ones with the biggest AI budgets or the most powerful AI tools. They'll be the ones that understand how to orchestrate machine intelligence into systems that solve real problems in ways that weren't possible before.

R/GA's newly independent status—after a management and Private Equity partnership led to its exit from IPG earlier this year—means we can move with the speed and agility this work demands, without holding company constraints. Addition remains an independent studio within this structure, but with access to global talent and enterprise capabilities when the work demands it.

The most important work in AI isn't about building better tools—it's about building better futures. And it's something we can do even better together.

Addition is an applied AI studio for modern brands.

Visit Addition.ml to learn more.