Rajesh Jain

Applications to Marketing

This essay builds on my past writings and inputs from AIs (Claude and ChatGPT).

World models are rapidly emerging as the next frontier of artificial intelligence. Unlike large language models, which predict the next word in a sequence, world models predict what will happen next in the world—they build internal simulations of cause and effect. They imagine, forecast, and plan. If LLMs are powerful librarians of human knowledge, world models are flight simulators for decision-making. And when viewed through the lens of marketing, their impact could be nothing short of transformative.

For decades, marketing has been built on pattern recognition. We segment customers into groups, examine historical patterns, and optimise campaigns based on what has worked before. At its core, marketing has been fundamentally backward-looking: the past determines the next best action. World models flip this on its head. They move marketing from remembering to imagining—from projecting history to simulating futures.

1. From Customer Segmentation to Individual Simulation

Today, personalisation is rooted in similarity: “customers who bought X also buy Y.” With world models, similarity becomes irrelevant. Instead of predicting behaviour because other customers behaved that way, the AI models the internal dynamics of each specific customer and simulates how that individual will behave in different conditions. This is methodological individualism applied to marketing—understanding each customer as a unique decision-maker, not as a member of a segment. It is no longer “people like her”—it is her.

2. From A/B Testing to Counterfactual Experimentation

Instead of running experiments on customers, brands can run experiments inside the twin. A marketing team can test hundreds of interventions—timing, cadence, offer structure, creative style, channel mix—safely in simulation. The winning strategy reaches the customer only after the twin has proven it. This dramatically reduces AdWaste and profoundly raises confidence in execution. You’re not spending to learn; you’re learning to spend.

3. From Pattern Matching to Causal Persuasion

Current marketing predicts that a runner will click on trainers because other runners did so before. World models understand why she runs—whether for stress relief, social belonging, or competition—and generate interventions that align with her motivation. Timing becomes as important as messaging; triggers become as critical as offers; subtle nudges outperform overt pitches. The power lies not in showing the right product but in activating the underlying decision driver.

4. From Owned Channels to Adaptive Environments

Emails, apps, and websites become dynamic worlds that adjust themselves based on moment-to-moment interaction. A NeoMail is no longer a static message—it becomes a live space that responds to each user’s micro-behaviour, shaping content, rewards, and recommendations automatically. This shift turns the revenue tax paid to ad platforms into investment in owned media that learns and improves. A channel becomes a continuously adjusting environment rather than a delivery mechanism.

5. From Designed Journeys to AI-Orchestrated Futures

Today’s customer journeys are flowcharts—courses built by humans and refined through hindsight. Tomorrow’s journeys will be discovered, shaped, and adjusted by the model itself in real-time. Every customer receives the journey most likely to maximise lifetime value—not because it was built manually, but because the model has simulated thousands of futures and chosen the best one. Marketing becomes proactive navigation, not reactive optimisation.

6. From Cost Centre to Profit Engine Through Predictive ROI

World models change the role of marketing measurement. Dashboards today describe the past. World models allow marketers to preview the future—predicting churn before it happens, forecasting which Rest customers can be recovered economically, and projecting the impact of expenditure on attention and retention. This makes outcome-based pricing models like Progency feasible: you can guarantee results because you’ve already seen them work in simulation. Marketing becomes a profit engine with measurable, predictable returns, not a reporting function justifying past spend.

The Core Transformation

In short: LLMs answer questions; world models answer consequences. And that single shift moves marketing from reactive communication to proactive orchestration of customer behaviour. When combined with innovations like AI Twins, NeoMails, Co-Marketer agents, and the war on AdWaste, world models may herald the most consequential marketing transformation in decades—the shift from marketing to memories to marketing to imagined futures.

The question is no longer “what worked before?” but “what will work next?” That’s the difference between looking in the rear-view mirror and having a windscreen that shows you the road ahead before you drive it.

**

I asked the AIs to bring this future to life via a story about Ria (customer) and Maya (marketer).

Digital Twins

I next asked Claude to discuss Digital Twins in the context of World Models.

Digital twins have been around since NASA used simulators to troubleshoot Apollo 13 in 1970, though the term wasn’t coined until 1997. By the 2000s, manufacturers were creating virtual replicas of jet engines and factory floors to monitor performance and predict maintenance needs. Today, roughly 75% of businesses employ digital twins in some capacity, from monitoring turbines to simulating entire cities.

But here’s what most digital twins are today: sophisticated data visualization systems. They mirror current states, track historical performance, and run rule-based simulations. A digital twin of a wind turbine displays real-time performance data and can simulate the impact of changing wind speeds based on pre-programmed physics equations. Useful, certainly. Transformative? Not quite.

The Static Twin Problem

Current digital twins have a fundamental limitation—they’re built on static models. Whether it’s a Digital Twin of an Organization mapping business processes or a Digital Twin of a Customer representing buyer behaviors, these systems operate on fixed rules and historical patterns. They can tell you what happened and predict outcomes based on what has happened before, but they struggle with genuine novelty.

A customer digital twin today might predict that Sarah will buy running shoes because she bought running shoes before. It’s pattern matching wrapped in fancy visualization. The twin doesn’t understand why Sarah runs, what motivates her purchases, or how her behavior might evolve as her life circumstances change. It’s a database with a better interface.

Enter World Models

World models change the equation fundamentally. Instead of rule-based simulations, they learn the underlying dynamics of environments. They don’t just store facts about how customers behave—they develop an internal understanding of what drives that behavior, similar to how humans build mental models of cause and effect.

Imagine a customer digital twin powered by world models. Rather than simply mirroring past purchases and predicting repeat behavior, it would understand the customer’s decision-making process. It could simulate how that customer would respond to different marketing interventions, not because similar customers responded that way in the past, but because it has learned how this customer actually thinks and decides.

The twin could run counterfactual scenarios: “If we change the email cadence, how does Sarah’s engagement evolve over three months?” “If we introduce this new product category, which existing customers will find it relevant based on their underlying preferences, not just their purchase history?”

The Convergence

The convergence of world models and digital twins represents a shift from backward-looking analytics to forward-looking simulation. Digital twins have always promised the ability to “test before you deploy”—but with static models, you’re testing against the past, not the future.

World models bring three critical capabilities to digital twins:

1. Dynamic understanding rather than static rules. The twin doesn’t just know that customers like personalized emails—it understands how personalization affects decision-making and can predict when it becomes intrusive.
2. Genuine prediction rather than pattern matching. The difference between “customers like Sarah usually do X” and “Sarah will likely do X because of how she perceives value and makes decisions.”
3. Causal reasoning rather than correlation. World model-powered twins can distinguish between what drives behavior and what merely correlates with it, enabling true experimentation and optimization.

What This Means for Marketing

For marketers, this convergence opens a new frontier. Customer journey optimization today means analyzing past journeys and tweaking touchpoints. With world model-powered digital twins, you could simulate entire journeys before customers take them, testing not just messaging variations but strategic interventions—timing, sequencing, channel choices, offer structures.

The digital twin becomes less like a dashboard and more like a flight simulator—a place to test piloting strategies in realistic conditions before committing resources. You’re not just looking at customer data differently; you’re creating an environment where you can experiment with customer dynamics.

The question is no longer “what did customers do?” but “what would customers do if we changed the system?” That’s the shift from digital twins as mirrors to digital twins as laboratories.

**

My previous essays on Digital/AI Twins and Marketing:
Marketing’s AI Triad: Large Customer Model, Digital Twins, Co-Marketer
AI-Powered Digital Twins: Marketing’s Marvel
AI Twins: Digital Customer Representations That Will Transform Marketing
Digital Twins in Marketing: Magical Minions
AI Twins: The Future of Marketing Intelligence
AI Twins in Action: Daily Allies for Smarter Marketing and Meaningful Connections

Comparison

I asked Claude to compare LLMs and World Models.

The AI landscape is witnessing a fascinating debate: are Large Language Models the path forward, or do we need something fundamentally different? To understand this, we need to grasp what each approach actually does—and more importantly, what it cannot do.

What LLMs Actually Do

Large Language Models are trained to predict the next token in sequences of text, learning statistical patterns across billions of words. Think of them as extraordinarily sophisticated pattern-matchers. When you ask ChatGPT a question, it’s not “thinking” in any conventional sense—it’s generating the most probable continuation of your prompt based on patterns it absorbed during training.

This gives LLMs surprising proficiency in diverse tasks like translation, summarization, and dialogue generation. They can write code, explain complex concepts, and even seem to reason. But here’s the catch: their understanding of phenomena like gravity or thermodynamics is inferential, derived solely from linguistic patterns rather than grounded in physical reality.

The hotly contested question is whether LLMs develop genuine “world models” internally. Some research suggests language models can develop rudimentary world models for simple concepts in their internal representations, like tracking game states in Othello. But prominent researchers argue that instead of coherent world models, today’s generative AIs learn “bags of heuristics”—scores of disconnected rules that don’t form a consistent whole.

How World Models Work Differently

World models are neural networks that understand the dynamics of the real world, including physics and spatial properties. Rather than predicting the next word, they predict what will happen next in the physical world, modeling how things move, collide, fall, interact and persist over time.

World models take inspiration from mental models humans develop naturally—our brains form abstract representations from our senses into concrete understanding of the world around us. A baseball batter has milliseconds to swing—shorter than the time for visual signals to reach the brain. The reason we can hit a fastball is our ability to instinctively predict when and where the ball will go based on our internal world model.

These models consist of multimodal input processing, temporal prediction that forecasts future frames from historical data, and latent representation learning that operates on abstract state spaces. This allows AI agents to “imagine” different scenarios before acting, planning by simulating future state sequences for different strategies.

The Fundamental Difference

The key distinction lies in grounding. LLMs derive their power from statistical patterns in textual data, while world models seek to emulate causality and the laws of physics through learned simulations. An LLM “knows” about gravity because countless texts mention it. A world model understands gravity because it has observed thousands of hours of objects falling and learned the underlying physics.

World models trained on photos, audio, videos, and text create internal representations of how the world works and can reason about the consequences of actions. This matters enormously for applications like robotics, autonomous vehicles, and any AI that must act in physical space rather than just generate text.

What This Means Going Forward

Yann LeCun has argued that within three to five years, world models rather than LLMs will dominate AI architectures. Whether that timeline holds, one thing is clear: text prediction alone has limits. Acting reliably in the physical world requires more than next-token prediction—it requires maintaining state, predicting dynamics, and supporting counterfactual reasoning for planning and control.

We’re likely heading toward hybrid systems that combine both approaches—language understanding from LLMs with physical reasoning from world models. The question for marketers is: which capabilities matter for your applications?

Review – 3

ChatGPT:

In artificial intelligence, world models refer to an AI system’s internal representation of how the world works – essentially a simulated reality inside the machine’s mind. A world model enables an AI agent to predict future states of its environment and imagine potential scenarios without directly experiencing them. This concept has roots in cognitive science: humans are thought to carry a mental model of the world, allowing us to anticipate outcomes and plan our actions. As system dynamics pioneer Jay Forrester described, “The image of the world around us, which we carry in our head, is just a model… [one] has only selected concepts, and relationships between them, and uses those to represent the real system.”. In other words, we do not mentally simulate every detail of reality, but we hold an abstract model that captures the essential structure of our world. In AI terms, a world model is typically a generative model of the agent’s environment. Formally, researchers define a world model as a system that, given the current state of the world and an action taken by the agent, can predict the next state. For example, if the current state includes the position of objects and the agent’s status, and the action is “move forward,” the world model should predict what the new positions and statuses will be after that action.

A world model is all about learning an internal model of the world’s dynamics so that an AI can understand, predict, and plan within its environment. This involves learning abstract representations (latents), using self-supervised prediction as a training signal, capturing causal relationships over time, and leveraging the model for imagination and planning. The end goal is an AI that doesn’t just react reflexively, but “knows” how the world works and can reason about what to do – much like humans who foresee outcomes using mental models.

Consider this vivid analogy from a recent essay: An LLM is like a brilliant autocomplete that can imitate understanding by stitching together patterns in language, but it has no “imagination” beyond what it’s seen in text. A world model, by design, imagines – it simulates situations that perhaps never occurred exactly the same in its training data. It’s the difference between reading about the world versus experiencing a simulated mini-world. One paper framed it elegantly: “LLM simulates the next word in human language… [A world model] simulates the next world.” [Source]. Each has its strengths – LLMs leverage vast knowledge encoded in human text, whereas world models provide an internal sandbox for truly understanding and navigating an environment. The consensus emerging is that next-generation AI systems likely need both: the knowledge and linguistic capabilities of LLMs plus the grounded predictive planning of world models. Indeed, researchers have begun experimenting with combining the two, for example by training language models with objectives that encourage predicting missing text in a way analogous to JEPA (to imbue them with a bit of world-model-like consistency) [Source].

Claude:

Kenneth Craik proposed in 1943 that minds construct “small-scale models” to anticipate events before they occur. Philip Johnson-Laird’s mental models theory (1983) suggests we reason by mentally simulating scenarios rather than applying formal logic. The predictive processing framework in neuroscience argues the brain constantly predicts sensory input, updating beliefs only when predictions fail.

World Models translate these cognitive principles into AI architecture. They comprise nine interconnected components: latent spaces that compress high-dimensional observations into 32-64 dimensional representations filtering out unpredictable noise; predictive coding that forecasts future latent states from current state plus action; model-based reasoning enabling deliberative Mode-2 thinking versus reactive Mode-1 responses; causal structure encoding cause-effect relationships; simulation capabilities for forward prediction and counterfactual “what if” analysis; memory systems maintaining both short-term hidden states and long-term associative knowledge; planning mechanisms using gradient-based optimization or search-based methods like Monte Carlo Tree Search; self-supervision through contrastive or non-contrastive learning from unlabeled data; and embodied agent capabilities closing the sensorimotor loop from perception through planning to action.

This architecture provides what LLMs fundamentally lack: sample efficiency through imagination rather than trial-and-error (humans learn to drive in 20 hours versus millions of RL episodes for pure model-free learning), common sense grounded in intuitive physics about gravity and momentum, planning and reasoning through forward simulation of consequences, generalization by capturing underlying dynamics that transfer across tasks, safety through virtual testing of dangerous scenarios, and computational efficiency by predicting abstractions not pixels. [Source]

Review – 2

WSJ: “Today’s AIs are book smart. Everything they know they learned from available language, images and videos. To evolve further, they have to get street smart. That requires “world models.” The key is enabling AI to learn from their environments and faithfully represent an abstract version of them in their “heads,” the way humans and animals do. To do it, developers need to train AIs by using simulations of the world… There’s an almost unanimous belief among AI pioneers that world models are crucial to creating next-generation AI. And many say they will be critical to someday creating better-than-human “artificial general intelligence,” or AGI… While the type of AI that makes large language models and ChatGPT possible gets all of the attention right now, it’s world-model-based AI that is gaining momentum in frontier research and could allow technology to take on new roles in our lives… LLMs appear to be taking over some functions in white-collar jobs already, and AIs that are world-model smart could allow them to take over yet more. Blue-collar work has been relatively safe. But as AI developers go ever deeper on world models, robots could start applying for jobs as truck drivers, plumbers or caregivers.”

WSJ: “Yann LeCun has been telling anyone who asks that he thinks large language models, or LLMs, are a dead end in the pursuit of computers that can truly outthink humans… “We are not going to get to human-level AI just by scaling LLMs,” he said.”

Fei-Fei Li: “Today, leading AI technology such as large language models (LLMs) have begun to transform how we access and work with abstract knowledge. Yet they remain wordsmiths in the dark; eloquent but inexperienced, knowledgeable but ungrounded. Spatial intelligence will transform how we create and interact with real and virtual worlds—revolutionizing storytelling, creativity, robotics, scientific discovery, and beyond. This is AI’s next frontier…Building spatially intelligent AI requires something even more ambitious than LLMs: world models, a new type of generative models whose capabilities of understanding, reasoning, generation and interaction with the semantically, physically, geometrically and dynamically complex worlds – virtual or real – are far beyond the reach of today’s LLMs… For spatial intelligence, I define world models through three essential capabilities: 1. Generative: World models can generate worlds with perceptual, geometrical, and physical consistency…2. Multimodal: World models are multimodal by design…3. Interactive: World models can output the next states based on input actions.”

Daniel Ince-Cushman: “In the evolving landscape of artificial intelligence, two prominent paradigms have emerged although World Models remain lesser known. One is the Large Language Model (LLM), a linguistic construct trained to predict and generate human-like text. The other is the World Model, a computational framework focused on simulating environments and enabling decision-making through internalized representations of external dynamics…World models…are designed to learn representations of environments, often through multi-modal sensory data or interactive simulation. These models typically consist of perception modules (e.g., vision encoders), memory or state encoders, and predictive components that simulate future states given current inputs and actions. These systems embody fundamentally distinct conceptions of intelligence. The LLM derives its power from statistical patterns in textual data, while the world model seeks to emulate causality and the laws of physics through learned simulations. The utility of a world model lies in its ability to enable planning. Through internal simulation, an agent can forecast outcomes and select optimal actions without direct environmental interaction.”

Gilbert Traverse: “The most valuable enterprise AI investments going forward will likely focus on integrating these two paradigms: leveraging LLMs for human-centric interfaces and general knowledge, while deploying increasingly sophisticated World Models to drive the complex planning, optimization, and action needed for real-world automation in areas like manufacturing, logistics, and autonomous operations. Strategic foresight demands an understanding of both, allowing organizations to capitalize on the linguistic power of today’s AI while building the causal foundation for tomorrow’s intelligent agents.”