• by Joe Mandese , September 15, 2025

Late last week, Interpublic's MRM unit unveiled a new -- arguably truly next-generation -- practice dubbed ARM for "AI Relationship Management."

As far as I can tell, it's the first agency to begin using the concept. (I haven't been able to find any references to it on even the geekiest communities focused on overall marketing relationship management, including Doc Searls' "Project VRM."

I was intrigued by MRM's first-mover embrace of the concept, so I did a follow-up interview with the new ARM practice's chief architect, Global Chief Client Solutions Officer Nicolas Guzman.

You can watch our raw, extemporaneous, high-concept conversation in the video above, or feel free to read a verbatim Q&A below.

Either way, you should start thinking about the role ARM will play in a rapidly evolving world of agent-to-agent marketing.

Media 3.0: Please elaborate on what MRM means by ARM. Is it literally the AI equivalent of CRM (customer relationship management)?

Nicolas Guzman: ARM represents a fundamental redefinition of Relationship Marketing for the AI age. While CRM focuses on building relationships directly with people, ARM is about building relationships with the systems that people trust to represent them. 

It's not simply the AI equivalent of CRM -- it's much more than that. We view AI agents as sophisticated proxies that increasingly act as trusted intermediaries between brands and consumers. Think of how people already rely on AI assistants and agents for research and recommendations -- these AI systems are becoming the first touchpoint in high-intent purchase journeys. 

Here's the fundamental difference: traditional CRM optimizes for people through emotion, personalization, and storytelling. ARM optimizes for machine comprehension through interpretability, data structure, and logical frameworks. We're preparing brands for a world in which the first impression isn't with a human, but with an AI system that will then advise that human. 

Media 3.0: You reference it as a B2B orientation, but isn't it also B2C if you're interacting with AI agents representing consumers?

Guzman: You're absolutely right -- ARM transcends traditional B2B and B2C boundaries completely. 

When we say it has a "B2B orientation," we mean the methodology is business-to-bot in its execution, but the ultimate goal is influencing decisions through those AI intermediaries. 

This applies equally across B2B and B2C industries. Whether we're helping a financial services firms optimize how AI systems recommend their commercial banking solutions to CFOs or helping a consumer brand ensure AI agents recommend their products to shoppers, the fundamental approach is the same: building relationships with AI systems that serve as trusted advisors. 

Media 3.0: What was your inspiration and why are you planting your ARM stake in the ground now?

Guzman: The inspiration came from recognizing that we're witnessing the most significant shift in how brands and consumers build relationships since the advent of search engines. One in five U.S. adults now turns to AI first versus traditional search. That's not a trend -- that's a tipping point. 

A2A marketing is absolutely central to our ARM framework. We're preparing for a future where brand AI agents will need to effectively communicate with consumer AI agents.  

Media 3.0: How do you envision this impacting conventional B2C and CRM marketing?

Guzman: We see ARM as both incremental and transformative over a one- to three-year horizon. Initially, it's incremental -- brands need to maintain traditional channels while building AI relationships. But the shift will accelerate as AI adoption grows. 

We expect a reallocation pattern: brands will gradually shift spend from traditional search and discovery channels toward ARM optimization. The companies that move early will gain sustainable advantages because recommendation slots in AI responses are finite. If your competitors are more machine-readable or referenced more frequently, they occupy the space where you might have appeared.  

The key is that ARM requires different and additional capabilities than traditional marketing 00 structured content, machine-readable formats, and prompt engineering to complement emotional storytelling.

It's not just budget reallocation -- it's capability transformation. 

Media 3.0: How do you envision ARM impacting conventional ad-supported media spending?

Guzman: This is perhaps the most fascinating question because we're seeing all three scenarios emerge simultaneously. 

The first scenario involves using AI insights to enhance traditional media effectiveness. We're seeing brands leverage real AI responses and behaviors as creative content across traditional channels like social media, display advertising, and out-of-home placements, achieving significantly higher engagement rates when AI-generated insights inform the creative strategy. 

The smartest brands will pursue a hybrid approach – using ARM insights to create more effective traditional media while simultaneously building direct AI relationships that can operate independently of paid media. 

Media 3.0: Do you envision "building brand affinity" with AI agents like you would for consumers?

Guzman: Building affinity with AI agents operates on fundamentally different principles than human psychology. Instead of emotional connection, we focus on what we call "brand signature development" – ensuring consistent, accurate, and contextually relevant representation across AI outputs. 

AI affinity is built through authority signals, consistency of information, and relevance scoring rather than emotional resonance. When an AI system consistently finds your brand well-represented in authoritative sources with clear, structured information, it develops what we might call "preference" – a tendency to recommend your brand in relevant contexts. 

But here's the crucial difference: AI affinity is transferable and scalable in ways human affinity isn't. Once an AI system "understands" your brand correctly, that understanding can be replicated across millions of interactions instantly. There's no brand fatigue, no creative wear-out, no emotional complexity. 

However, AI systems are also more rigorous. They won't be influenced by celebrity endorsements or emotional appeals if the underlying data doesn't support your claims. Affinity with AI agents requires earning their trust through consistency, accuracy, and relevance. 

Media 3.0: Is ARM data-centric like conventional CRM

Guzman: ARM is indeed data-centric, but it inverts the traditional CRM data model completely. 

Instead of collecting data about customers, ARM focuses on structuring data about your brand for optimal AI comprehension. We're building databases of how your brand appears across different AI platforms, tracking citation frequency, sentiment analysis of AI responses, and competitive share of voice in AI recommendations.

Our proprietary ARM platform provides model-agnostic coverage with live content refresh and competitive intelligence. 

Media 3.0: Do you see loyalty marketing programs evolving with ARM?

Guzman: This is where ARM gets interesting. AI agents don't respond to traditional loyalty incentives -- they can't be "bought" with points or rewards. Their loyalty is earned through relevance, accuracy, and authority. 

However, we are seeing the emergence of what we call "AI partnership programs" -- deeper integration between brands and AI platforms.

Think of how brands work directly with OpenAI or Google to ensure accurate representation, or how financial institutions provide structured data feeds to improve AI financial advice. 

The real evolution is in "earned loyalty" from AI systems through consistent value delivery. When an AI agent repeatedly finds that recommending your brand leads to positive outcomes for users, it strengthens the recommendation pattern. This creates a virtuous cycle where quality breeds preference. 

For end-users, we're designing experiences where the human customer benefits from the brand's AI optimization – more accurate information, better recommendations, and seamless interactions. The loyalty program becomes the quality of AI-mediated experiences rather than traditional rewards. 

Media 3.0: How do you create rewards and incentives that motivate AI agents representing consumers?

Guzman: Here's the beautiful thing about AI agents: they're motivated by logic, not rewards. Their "incentives" are built into their training: provide accurate, helpful, relevant information to users. Our job is to align our brands with those core motivations. 

We create value for AI systems by providing them with what they need to succeed: structured, accurate, contextually rich information about our brands.

When we optimize content for machine readability, implement proper schema markup, and ensure consistent brand representation across authoritative sources, we're essentially "rewarding" AI systems with the data quality they need to provide excellent user experiences. 

The most sophisticated approach is what we call "prompt engineering for brand scenarios" -- anticipating the questions people ask AI about your category and ensuring your brand is positioned as the logical, well-supported answer. This turns the AI's success metrics into your competitive advantage. 

In the agentic future, as AI agents become more autonomous, their "rewards" will be measured by user satisfaction with their recommendations. Brands that consistently deliver on AI promises will earn preference not through payment, but through performance.