Three years ago, in an MxMIndia column (The Diffusion of AI: What do the marketing model predict? – Feb 16, 2023), I attempted to apply some classical marketing models to the coming diffusion of Artificial Intelligence. At that time, ChatGPT had just been released and AI was beginning to escape the laboratories of technology giants into public consciousness. The central question I explored then was simple: how would AI diffuse through markets and society?
Looking back from 2026, the answer is clear: the diffusion has been faster, deeper and structurally more disruptive than most technology marketing models anticipated. But something else has also become evident. While classical technology marketing models help explain how AI spreads, they do not fully explain how AI creates value. For that, we must turn to a different framework in marketing theory: Service-Dominant Logic (SDL).
To understand the marketing of AI today, we must combine the diffusion models with the SDL perspective. In my earlier column I referred to the Theory of Product-Form Strategy (PFS) which proposes that innovators must decide whether to market their technology as know-how, a component or a system. At the time, I suggested that the two leading AI innovators—OpenAI and DeepMind—were likely to follow different paths.
In the intervening three years, this prediction has broadly held.
OpenAI has largely followed the know-how and component route, licensing its models through APIs and platforms. Its partnership with Microsoft embedded AI capabilities across Microsoft’s software ecosystem—from Office to GitHub to cloud services. Alphabet, through DeepMind and Google AI, has pursued the system strategy, embedding AI into end-to-end platforms across search, advertising, cloud infrastructure and enterprise solutions. A third important player has emerged with a distinct positioning: Anthropic. Anthropic’s strategy sits somewhere between the two approaches. Through its Claude models and its partnerships with Amazon and others, Anthropic has positioned itself not merely as a model provider but as a trust-centred AI infrastructure layer for enterprises. In marketing terms, Anthropic has emphasised safety, reliability and alignment as its primary value proposition. Where OpenAI’s strategy emphasised rapid diffusion through developer ecosystems, Anthropic has emphasised trust and governance as differentiators, appealing particularly to enterprise customers navigating regulatory and reputational risks around AI.
Meanwhile, the broader AI ecosystem has expanded dramatically. Today the competitive landscape includes:
- OpenAI and Microsoft
- Google / DeepMind
- Anthropic
- Meta with open-weight models
- Amazon through AWS AI services
- Apple with on-device AI
- Nvidia as the dominant AI infrastructure provider
- A rapidly growing universe of specialised AI start-ups
The result is not a single AI market but a stack of AI markets, ranging from foundational models and chips to enterprise tools and consumer applications.
The second framework I used earlier was Geoffrey Moore’s Technology Adoption Life Cycle, particularly the challenge of “crossing the chasm” between early adopters and the early majority. In 2023, AI appeared to be at the early adopter stage. By 2026, the picture looks very different. Generative AI tools have been adopted by hundreds of millions of users globally. Enterprises across sectors—from finance and retail to healthcare and education—are integrating AI into workflows. This suggests that AI has not merely crossed the chasm; in some domains it is already entering what Moore called “the tornado”, the phase where adoption accelerates dramatically as market standards begin to emerge.
Three forces have driven this rapid diffusion: First, the cloud infrastructure required to run AI models already existed. Second, the consumer interface—smartphones and browsers—was already ubiquitous. Third, AI applications offer immediate productivity gains in knowledge work. Unlike earlier technologies, AI did not need to build an entirely new distribution infrastructure.
However, diffusion models still leave a crucial question unanswered. What exactly is the product that is diffusing? Is AI a product, a platform, a capability, or something else entirely?
Traditional marketing frameworks assume that firms produce value and consumers purchase it. But the AI economy increasingly contradicts that assumption. The reason becomes clearer when viewed through the lens of Service-Dominant Logic (SDL)
In an earlier column, (Operand and Operant Resources and the Emergence of Service-Dominant Logic (SDL) Marketing -Oct 27.2022), I wrote about the shift in marketing thinking from operand resources to operant resources.
Operand resources are tangible resources—land, minerals, machines—on which an operation is performed. Operant resources are resources that act on other resources: knowledge, skills and competencies. The key insight of SDL is that value is not created by firms alone. Instead, value emerges through the interaction of operant resources across a network of actors.
Customers themselves become operant resources because they actively participate in creating value. Few technologies illustrate this principle more clearly than AI. Unlike traditional products, AI systems improve through use. The prompts users write, the workflows they design, the feedback they give and the applications they build all shape the value of the system. In other words, AI users are not merely consumers of value—they are co-producers of value. A programmer using GitHub Copilot, a marketer crafting prompts for generative design tools, or a doctor using AI diagnostic assistance is actively extending the capability of the system.
The real economic value of AI therefore emerges not from the algorithm alone but from the interaction between human expertise and machine intelligence. This is pure Service-Dominant Logic.
If AI users are operant resources that co-create value, an important question follows: who captures that value? Today, much of the economic benefit generated through user interaction accrues to the AI platforms themselves. User prompts, workflows and feedback often help refine the system, improving its capabilities for all users. But as AI ecosystems mature, a new possibility emerges.
Platforms may begin to recognise and reward value-adding users—whether developers, enterprises or individual creators—whose innovations significantly expand the capabilities of the system. Such mechanisms would need to operate within strict safeguards to protect confidentiality, intellectual property and privacy. Yet if designed carefully, AI platforms could diffuse insights generated by users across the ecosystem while ensuring that the originating user is appropriately credited or compensated. In effect, the AI economy may evolve toward a model where value created by operant resources is shared more explicitly across the network.
Seen through the SDL lens, AI is not a product but a service ecosystem. AI firms do not deliver finished value. Instead, they make value propositions—capabilities that users integrate into their own processes. The enterprise user who embeds AI in a supply chain workflow, the designer who uses AI to generate creative variations, the educator who builds AI-driven tutoring systems—all become participants in the value creation process.
This also explains why AI markets have evolved toward platform models rather than traditional product markets. Platforms allow a network of developers, companies and consumers to act as operant resources, continually expanding the value of the system. The extraordinary growth of the AI ecosystem over the past three years—from prompt engineers and AI start-ups to enterprise integrators—reflects precisely this SDL dynamic.
The diffusion of AI is therefore not simply the spread of a new technology. It is the emergence of a new value creation architecture, where humans and machines increasingly collaborate as operant resources. In this sense, the future of AI may be less about the triumph of machines over humans and more about the co-evolution of human and artificial intelligence.
And for marketers, that means the challenge is no longer just to sell technology It is to design ecosystems where intelligence—human and artificial—can continuously create value together.
hardraincafe.blog 