So, you go to your Global Brand Hub, where you find several 100+ PDF documents full of various guidelines. You search for “colour” and find a matrix that says “Use Pantone 299C for print, #00A3E0 for digital.” The agency used #00A4E0. Is that a typo? Or a holdover from another guideline deck, “Digital-First Brand Refresh_v3_FINAL.pptx,” from last quarter?

You Slack a senior director, but they're in back-to-back meetings. You email the brand compliance alias and get an auto-reply: “We will respond within 48 business hours.” But the agency is pinging. The media slot is booked. It’s a simple, 10-second question that has blocked a time-sensitive asset.

This seemingly small frustration is actually a symptom of brand governance chaos—a massive, hidden tax on your speed, budget, and morale. Fixing this chaos requires more than just a clearer guide or a better folder structure; the real solution is to evolve from static repositories to dynamic, intelligent agents capable of delivering a single, correct answer instantly.

Agentic architectures help solve for relevant retrieval.

For the last decade, improving access to information largely meant adding a better search box to static, file-based brand hubs. However, a search box only fetches documents; it still forces the user to do the work of finding the answer within those documents. This is the critical failure point in the 4:00 PM panic scenario described above. A better solution is to move from a static repository to a system powered by orchestrated agents that retrieve data.

Unlike a search box, an agentic solution can perform tasks on behalf of the users. It can understand context, like knowing who you are (for example, a brand manager) and what you're working on (in our example above, a digital display). From there, it can reason, retrieve information across multimodal sources (PDFs, databases, websites), verify accuracy, resolve data conflicts, and compose an answer. 

It doesn't give you ten blue links to sift through; instead, it offers a single, definitive, reference-backed response. If conflicting data appears—for example, Marketing_v1.pdf says #000000 but Poster_v1.pdf says #000011—the agents use context, role and logic to determine the most accurate answer. If a clear resolution isn't possible, it flags the conflict so the user can make an informed decision: “The correct hex code for digital-first applications is #00A3E0. The #00A4E0 value is an outdated code from the Q1 refresh.” 

This shift is a powerful new driver of enterprise efficiency today. In fact, agentic assistants like this move beyond being passive tools for answering questions, evolving into Brand Intelligence Systems that retrieve accurate brand data, enforce multi-modal compliance, and generate on-brand, multi-modal content at scale.

Proving a measurable lift in efficiency for over 1,800 users. 

We recently partnered with a global technology leader facing this exact challenge. With thousands of employees and partners across the globe, they needed to provide instant, reliable and source-attributed answers to brand questions, at scale, as their MVP. We designed and deployed a bespoke, enterprise-grade AI assistant powered by orchestrated agentic workflows using a tailored Retrieval-Augmented Generation (RAG) architecture.

This orchestration ensures two things. First, the agents don’t hallucinate. Second, the system understands context (who you are, your role and your task) to deliver the right answer, often by combining multiple verified sources. 

The impact of our solution was immediate. Within four months of its rollout, over 1,800 unique users were interacting with the assistant per month, with engagement trending positively. More importantly, we proved we were solving the slow bleed of inefficiency. User sessions became measurably more efficient, dropping from an average of 1.64 messages to just 1.41, because they were getting the right answer, faster, on the first try.

Crafting solutions that integrate into real workflows. 

There are plenty of off-the-shelf and decent tools already available for building a chatbot. The real challenge lies in building bespoke systems that integrate seamlessly into daily workflows, also known as complex enterprise integrations, that are also secure, reliable, highly accurate and personalized.

This is essential not just for performance, but also for compliance, data protection and user adoption. When the experience feels like a natural extension of how teams already work, that’s when transformation sticks.

However, connecting every piece of the puzzle requires a holistic approach. The development of our solution for this specific client was rooted in a single, unified team that covered everything from initial strategy and user pain point understanding to UX and design. This was made possible by engineering teams who orchestrated models to deliver secure answers at scale, all while our QA and delivery teams ensured everyone remained focused on achieving enterprise-grade outcomes.

In practice, this meant that our strategy team mapped pain points, the AI Core team built datasets and evaluation frameworks, the UX team distilled complexity into intuitive experiences, and engineering ensured scalability, resilience and security.

The hardest part? Balancing accuracy, latency and cost across deep enterprise-grade system integrations. It took many iterations over the past three years to achieve team maturity. Key project members had already worked on five similar conversational AI deployments across industries, and that collective experience was crucial. The learning curve has been steep but transformative.

A unified system built by a holistic team frees creativity.

The 4:00 PM panic is just the surface symptom of deeper inefficiencies that off-the-shelf tools can’t fix when accuracy, latency and cost all matter. 

True success comes from integrating bespoke AI systems seamlessly into the creative process—not as add-ons, but as enablers. This is what a holistic approach delivers in practice: a unified system where strategic insights, intuitive design and enterprise-grade engineering work as one. It is this system that ultimately solves the hidden tax of brand friction, giving your most valuable creative people back their time, budget and energy to focus on the work that actually moves your brand forward.

At the heart of it all is the user. We are driven every day by the goal of building next-generation AI interfaces that are not only intuitive and meaningful but also truly smart. For brands and enterprises seeking to achieve this same level of clarity and efficiency, our bespoke agentic AI architecture can be fully tailored. It adapts to your unique workflows and data environments while respecting all governance requirements, empowering your teams with intelligent systems designed precisely around your needs.

At a glance:

Marketers can use AI to create differentiated and effective work by moving beyond generic tools and building a proprietary system trained on their own unique brand intelligence. Here’s how:

• True AI-powered differentiation comes not from the tools themselves, but from training them on your unique historical data, audience insights, and strategic knowledge.
• The key is to develop a custom agentic workflow—a network of specialized AI agents—that acts as a proprietary, intelligent system that understands and enforces your brand.
• This approach transforms your marketing archive from a static resource into a dynamic, compounding advantage, creating a defensible piece of intellectual property that competitors cannot replicate.

How can we use AI without losing our brand's unique voice?

When everyone is rushing to adopt the same AI tools with the same capabilities, how do you create work that actually stands out? Creative differentiation has long been a critical imperative for brands, which is taking on a new urgency as creative workflows become more automated—at the risk of becoming more homogenized in the process. So, the answer to the differentiation dilemma doesn't lie in mastering the same off-the-shelf tools as competitors. Instead, it lies in building a proprietary system fueled by a company's unique data, intelligence, and history.

As we move from the excitement and potential of the early hype cycle to the messy reality of implementation, many brands are finding that while generic AI tools are powerful, they can lack the brand-specific nuance required for high-stakes marketing if they aren’t trained on your brand’s unique intelligence. The initial thrill has been replaced by the practical challenge of making AI outputs not just faster, but better and—most importantly—distinct.

How can my brand's data create a competitive advantage in AI?

In overcoming these challenges, we must move our focus from the AI model itself to the fuel that powers it. When brands use the same public models with similar prompts, they will inevitably get similar, generic results—a race to the bottom that creates content not worth people's time.

True differentiation comes from the quality and uniqueness of your inputs. The most valuable assets your brand possesses are your own historical performance data, your nuanced audience insights, your risk tolerance, and your strategic knowledge. Altogether, these make up your brand’s DNA. More than just quantitative data, brand DNA includes the qualitative elements that are harder to measure but essential to your brand’s identity: your unique point of view, your strategic decisions about which markets to enter (and which to avoid), and your specific interpretation of complex industry regulations. These are the inputs that contain your competitive edge.

So, how do you activate this unique brand DNA, transforming this treasure trove of data from a static archive into a dynamic, intelligent engine? The path to doing so lies in building a custom, agentic workflow.

What is an agentic workflow for marketing?

A marketing-focused agentic workflow operates as a network of specialized AI agents, all trained on your specific brand ecosystem to work in concert. These agents function like a bespoke, in-house team of experts that thinks and acts like your brand, but with the capacity to operate at limitless scale.

At the heart of this system is the brand model. Training this model involves feeding the system your entire history: your visual identity system, your tone-of-voice playbooks, your compliance policies, and your competitive positioning. We enrich this with the very language from your past high-performing campaigns and the specific nuances of your regulatory landscape. The model learns not just the rules, but the spirit of the brand. This in turn delivers consistency, ensuring every piece of AI-assisted work is instantly and recognizably on-brand.

Within the agentic workflows we build for our clients, we deploy specialist agents to support the core brand model. For example, our Insights Agent surfaces relevant trends and conversations, enabling brands to move at the speed of culture and engage authentically with relevant trends. Meanwhile, Strategy Agents act as powerful accelerators for growth, helping brands identify opportunities and activate new audience segments by analyzing market data through the lens of their unique strategic goals. Insights that in the past could take one of our CPG clients 18 months to identify and activate are now reviewed and responded to in close to real time.

How does an agentic workflow become a defensible asset?

This approach transforms AI into a strategic asset that serves as a powerful engine for differentiation. While competitors rent access to generic models and compete on similar prompts, you’ve built distinct intellectual property that produces creative work no one else can.

This "brand brain" is a living system that grows smarter and more effective with every campaign it analyzes and every data point it absorbs, a virtuous cycle that turns your brand’s history into compounding advantage. In practice, this means the agents work in concert to uncover unique creative opportunities. An Insights Agent might identify a burgeoning cultural trend through the specific lens of your brand's DNA, surfacing an angle invisible to your competitors. This is passed to the Strategy Agent to propose a micro-campaign, which is then validated by the brand model to ensure it aligns perfectly with your tone. This interconnectedness is what allows you to move on distinct cultural moments with speed, relevance, and safety.

What’s even better is that insights from a successful campaign automatically inform the strategy for the next one. The nuances of your highest-performing creative become ingrained in your brand model’s DNA, making the entire system better at finding the unique intersections between what your brand stands for and what the culture is talking about, leading to increasingly differentiated work over time.

Where should we begin in building a custom AI workflow?

In the age of AI, the most durable competitive advantage won't come from a tool you can buy, but from an internal capability you build. Activating your unique data through a custom agentic system is how you make your own intelligence scalable, turning your brand’s history into a strategic asset.

Making this operational and strategic shift is a significant undertaking, and every brand is at a different stage of its journey. That’s why we engage with our clients flexibly. For some, it begins with advisory to navigate the change. For others, we stand up managed agent services to take the burden from in-house teams. And for those ready to invest in self-service, we help execute full custom builds. Regardless of where you start, the most critical step is recognizing that your brand’s history is the most powerful dataset you own.

As part of NVIDIA’s OpenUSD Insiders livestream series, our SVP of Innovation, Susan Foley; our VP, Global Head of Technology, Peter Altamirano; and our VP, Computational Creativity & Innovation, Emrah Gonulkirmaz, dove into AI-driven marketing, content creation and the future of agentic workflows. In conversation with NVIDIA’s Jamie Allan, Director of AdTech & Digital Marketing Industries, and host Edmar Mendizabal, they explored practical use cases for digital twins and NVIDIA Omniverse—complete with hands-on tips, real-world client examples and advice for organizations eager to embrace these innovations.

If you missed it, you can watch the full session below or keep reading for the key takeaways.

Digital twins are the new foundation for creative scale.

Unlocking business value today means taking control of your creative assets and processes. Digital twins—the hyper-accurate virtual models of products, characters and spaces—are quickly becoming the bedrock of that approach. Allan set the stage for the session by explaining, “A big part of what we’ve been doing is [figuring out] how to evolve the content supply chain for marketing content and ads. A lot of that is founded in creating digital twins of products, whether it’s a car or a shampoo bottle. That’s where the power of OpenUSD comes into play.”

Digital twins are built using open standards like OpenUSD—an open-source framework and file format for describing, composing and interchanging 3D scenes and assets—and applications that are developed with platforms like Omniverse. They serve as the single source of truth for everything from product imagery to complex industrial simulations, allowing businesses to rapidly iterate, test changes virtually and deliver new products or updated assets in a fraction of the traditional time. As Altamirano said, “You can optimize layout, workflows, asset creation and test and simulate your processes far faster than in the real world—no matter if you’re updating a retail shelf, visualizing packaging or piloting new robotics workflows.”

That said, what usually holds organizations back isn’t a lack of understanding of the benefits of digital twins, but simply not knowing where to start. Foley advised, “Start small, scale up as you prove value, and we’ll help you migrate services to compute so you can build your own moat with owned intelligence.” Pro tip: Don’t let complexity slow you down. Today’s SDKs, libraries, templates and demo projects make it easier than ever to get started quickly with no need to build everything from scratch.

Modular, agentic workflows mean AI is now your creative partner.

The future of creative production isn’t simply generating more and more images and text with AI. It’s about orchestrating a system where specialized AI agents collaborate across the full pipeline. A standout example is our experimental AI-generated campaign for PUMA, where every stage—from initial script and storyboard to animation and editing—was orchestrated by AI agents using Monks.Flow, our professional managed service powered by AI. 

Thanks to NVIDIA NIM microservices and node-based orchestration enabled by Monks.Flow’s Pathways framework, AI agents can swap out models or creative roles as needed. Pathways uses self-learning AI to autonomously manage, optimize and adjust workflows in real time. For example, we can switch from one generative model for texturing to another for background imagery without disrupting the flow. 

Crucially, this entire process was anchored to a high-fidelity digital twin of the PUMA product, built using NVIDIA Omniverse libraries. “We started by importing a precise 3D model of the sneaker created in the OpenUSD format into NVIDIA Omniverse USD Composer,” explained Altamirano. This virtual product served as the foundational source of correctness for every subsequent creative step. 

The process didn’t stop there: synthetic data generated from the USD-based model was used to train and guide AI agents tasked with upholding the correct cinematic style throughout the film. That way, the team could control vital aspects required to meet brand guidelines, such as camera angles, lens choices and product accuracy.

As Emrah put it, “After bringing the product to the previsualization stage in Omniverse, we gained full control over the generation process.” Then, using Pathways, a network of specialized AI agents orchestrated animation, editing and scene composition, maintaining brand consistency throughout. 

How to get started and scale fast.

Whether you’re a marketer, a manufacturer looking to modernize, or a developer curious about AI-driven workflows, the new ecosystem emphasizes accessibility. “Get your first Omniverse setup done and you’ll see how reusable and scalable it really is,” said Altamirano. “Start with a prototype, prove the value, then expand. This approach works for retail, hospitality, even real estate.”

For organizations not sure where to begin, the advice is clear:

• Start small and show early results. Quick wins build stakeholder buy-in and reveal practical value.
• Invest in training not just for engineers, but for your whole team. NVIDIA Omniverse Blueprints, free Deep Learning Institute courses, and a vibrant developer community enable rapid learning and onboarding.
• Embrace open, modular platforms. This lets you change direction, upgrade AI models, and keep workflows on the cutting edge without locking yourself into monolithic systems.

Ultimately, modern creative innovation isn’t about one-off experiments; it’s about embedding intelligence, agility and modularity into the heart of your business. Digital twins anchor accuracy and scale. Agentic AI workflows make creativity collaborative and customizable. The path to scalable, AI-driven content creation has never been clearer. 

In fact, the concept that echoed through every session was the rise of the AI agent. Far from being just another chatbot, Salesforce’s powerful agentic technology, Agentforce, is positioned as a new face for the brand—a concierge, a personal shopper and a problem-solver, all powered by a company's unique data and content. This vision has the potential to reshape everything from customer service to a marketer's daily workflow.

So, let’s unpack the key announcements that build toward this vision. We’ll explore what it means to redefine customer interactions through AI agents, look at the underlying platform changes in Data Cloud and Marketing Cloud that make it possible, and contextualize how this could transform your day-to-day reality.

AI agents are transforming brand communication into a dialogue. 

For years, digital marketing has largely been a monologue where brands broadcast messages and hope customers are listening. The vision presented at Connections signals a definitive move away from this dynamic. The future showcased by Salesforce is one of dialogue, where every touchpoint becomes an opportunity for a meaningful, two-way conversation orchestrated by AI.

This marks the end of the "do not reply" email. Imagine sending a promotional email with a curated set of product recommendations. Instead of that being the end of the interaction, it becomes the beginning. A customer could reply directly to the email with, “I like those pants, but do you have them in blue?” and Agentforce, acting as a personal shopper, would respond with available options, check inventory and even help complete the purchase. This transforms a static campaign into a dynamic, personal shopping experience.

This shift also forces us to rethink the purpose of content. We’ve traditionally viewed content as a destination; the goal was to get a visitor to land on a blog post or a product page. In the new paradigm, content becomes fuel for the AI engine. That blog post about new product features is no longer just for human readers; it’s a critical input that teaches AI Agents how to intelligently discuss those features with a customer. Every article, every product description, every FAQ becomes a knowledge source that makes the brand’s AI smarter and more helpful. 

Ultimately, this leads to a new front door for your website. AI Agents will become the primary way visitors interact with a brand online, moving beyond static navigation and simple search bars. It will be a conversational interface that can answer complex questions, make nuanced recommendations and even take action on the user's behalf, like booking a meeting or making a reservation.

A unified platform makes all your data ready for AI.

This conversational future requires an immense amount of power running behind the scenes. The strategy presented at Connections rests on two pillars: unifying the marketer’s experience with the Marketing Cloud Next platform and making all data AI-ready with major enhancements to Data Cloud.

One of the key enablers of this vision is the forthcoming Marketing Cloud Next platform, built on Salesforce Core. It’s designed to solve the problem of platform fragmentation that has long challenged marketers who navigate disparate tools for different channels. Instead, the new vision is one of consistency, where the experience is seamless from one tool to the next. The HTML editor for building a web campaign will look and feel just like the one for creating an email. The same powerful product recommendation engine will be used across every channel, from web to email to mobile. And a single, unified AI brain—Agentforce—will power every interaction. This consistency will not only streamline workflows but also make it significantly easier to cross-train teams, allowing them to become true cross-channel experts.

Of course, any AI is only as good as the data it’s trained on. This is where the enhancements to Data Cloud become critical. Salesforce showcased features designed to supercharge the process of getting data ready for AI agents and recommendation engines.

• A renewed focus on data foundations: While new features are exciting, they amplify the need for a solid data foundation. Without it, brands face common challenges like low-quality data that erodes trust, data sprawl across disconnected systems, and limited access that prevents teams from acting on insights. Getting the foundation right is the prerequisite to unlocking the true power of AI.
• Smarter, broader data resolution: We’re familiar with identity resolution for customer profiles, but a key enhancement is the extension of these resolution rules to all Data Model Objects (DMOs). Imagine being able to resolve your entire product catalog, ensuring clean, consistent product data for your AI to use.
• Making unstructured data usable: One of the most powerful announcements was the ability to "vectorize" unstructured data. In simple terms, this is a kind of technological magic that allows AI to understand the meaning and context of content without needing manual tags. The full text of a blog post can now be used as if it were perfectly structured data, allowing the AI to mathematically determine which parts are most relevant for answering a specific question.
• Secure, seamless collaboration: To further break down data silos, Salesforce announced Data Cloud-to-Data Cloud zero-copy connectors. This will allow different business units, or even different companies, to join their Data Cloud instances together. These “data clean rooms” enable data sharing for richer insights in a completely privacy-compliant way.

AI will become an essential co-pilot for every marketer.

So, what does this all mean for the marketer? Their role will evolve from a manual builder to something more akin to a creative director. Using Agentforce, a marketer will be able to generate campaign ideas, customer journeys and even certain components of creative assets simply by providing a strategic brief. Crucially, this output is fully editable. The marketer remains in complete control, but their time is freed up to focus on strategy, creative oversight and optimization, rather than the nuts and bolts of campaign setup and creative production.

Monks is also building tools that empower marketers to take their ability to scale content to the next level. Monks.Flow can be integrated into Salesforce to supercharge a marketer's ability to build net new creative assets that power these journeys and are incorporated into these HTML templates to support better, smarter dynamic personalization.

Collaboration will also become more fluid. These AI agents can be invited directly into Slack channels, acting as a new kind of teammate. Teams will be able to brainstorm with the agent, review its proposals in real-time and collaborate on campaign elements in the same digital space where they conduct the rest of their work.

Finally, this changes how we approach analytics. Instead of digging through complex dashboards and manually cross-referencing reports, marketers will be able to use AI dashboards to get instant, natural-language summaries of campaign and channel effectiveness. This boosts the productivity of not just marketers, but data scientists as well, allowing them to focus on deeper strategic questions instead of routine reporting.

Brands can prepare now for this conversational shift.

The message from Salesforce Connections was clear: the future of marketing is a shift from monologue to dialogue. This approach is built on the pillars of a unified platform that simplifies execution, smarter data that makes every piece of content valuable and AI co-pilots that empower marketers to work more strategically and creatively.

The key takeaway for brands isn't simply to "buy more AI." It's to start thinking differently today. How can we prepare our content to be AI fuel? How can we structure our data to be ready for these new conversational experiences? Those who begin asking these questions now will be the ones best prepared to lead in the conversational future. The time to prepare is now.