UN-ADVERTISING
What happens when AI unbundles advertising, marketing & PR into data and code — and re-bundles them around proof?
“Pervasive structured data is inevitable, which means value creation lives in the irreducible.”
Receipts are the New Attention
Advertising and promotion exist because human attention is scarce. What should marketers know when your next customer isn't a human you attract and persuade?
The semantic web uses structured data — shared vocabularies foundational to AI and blockchains — that transforms simple search into novel automated services. We’re in a new reality, where earning the buy box is a formula: Context Completeness × Trust.
I argue that agentic commerce reduces everything except desire, taste, identity, belonging, status, meaning, and fandom—to math (expressed as data and code). You used to build trust. Now you’ll sell receipts (attestation and proof).
Agents replace human browsing with matching that happens upstream of human attention and point of sale: Matching & Fit = Alignment of Human Context with Product Context. This shift collapses the sales process from weeks to moments.
The New Rules
The sales funnel collapses because agents, (machine automations), evaluate and act rather than browse. This change in sequence, speed, and matching-for-fit shifts value from the Web2 standard of buying attention to controlling what AI systems surface, recommend, and transact. Rich Data = Structured Data + Context. Rich Data is the signal that shapes what gets seen and selected.
Revenue Models
Big AI needs revenue models. See OpenAI’s new ad service, here. To me, it looks like Chrome-with-kickbacks. I take the contrarian view that switching costs are so low, incentives so changed, and open-source model costs decreasing that LLM competition self-corrects towards transparency and unmediated results.
Context. Context. Context.
Think about Agentic Commerce as interacting with thick (lots) or thin (few) variables called “context”. A thin-context agent is a lot like search that can parse simple queries, but know nothing about the person who typed it. Thick-context agents can consider, compare, and qualifiy constraints to match for symmetry and fit. Thick context enables task automation and advanced agentic services.
In this new world, you win by matching product and buyer context. What do I mean by that? Agentic shopping uses rich, structured data to understand buyer requests holistically, as scored needs, wants, constraints, and weighted fit. In other words, agentic shopping needs context. These “intent graphs” let agents deliver sophisticated matches from high-intent search. Example: “Durable, on-trend trail-running shoes, under $150, delivered by Friday.” The winner isn’t the brand that ranks for “running shoes.” Products that win the buy box carry context-aware human desire and performance constraints as data.
The Math of Intent Graphs
Context Completeness × Trust
Matching and Fit = Human and Product Context Alignment
Rich Data = Structured Data + Context
Thick Context Shifts Your To-Do List From Discovery to Delight
Imagine what thick, holistic product and buyer context…constructed from intent, preferences, taste, history, past purchases, constraints, daily interactions, and fitness-for-purpose could do? Thick context delivers measurably better matching and fit. It is also a form of insurance against expensive disputes, returns, and mistakes.
The Marketing Hinge
Agent-driven buyer context and thick product context are the new decision vectors. But, brands only control product variables. The move: Make structured product context an always-on task to be ready for the moment agents discover, compare, and purchase. If I’m right about self-correcting LLM transparency and trust, legacy marketing should retire funnel, click, auction, and view interventions in favor of always-on context development. www.economy3.org.
Context Commerce and The New Data Value Chain
I argue for a new management vector I’m calling a Data Value Chain.
If advertising, marketing, & PR unbundle into data and code, and re-bundle around attestation and proof…
If agentic commerce rewards thick product context (expressed as math and code)…
If countries with 42% of global GDP require structured sustainability data (product context)…
If brands must prove marketing claims (with dynamic product context)…
If product value chains work best when aligned with data value chains…doesn’t that mean we we should recognize that data value chains are a valid management vector?
Isn’t this where agentic commerce, the regulatory environment, and the real economy all meet?
Aren’t compliance, climate, culture, and performance the new advertising?
Aren’t they also data value chains?
Don’t data value chain records build the proof layer?
Isn’t the only question what data structure to use?
On Advertising
New global advertising regulations view greenwashing and false claims as consumer fraud — an idea sure to spread. On 27 September 2026, the EU’s Empowering Consumers Directive prohibits generic environmental claims — “eco-friendly,” “green,” “sustainable” — unless the seller substantiates claims dynamically. The rule binds any company, anywhere, selling to EU consumers. Penalties reach 4 percent of annual turnover. Compliance requires dynamically provable receipts.
The EU’s Green Deal has a suite of regulations that require structured sustainabililty data governed by Digital Product Passports (DPPs). California, and 42% of global GDP require sustainability data, too.
AI search volume is another driver. More than half of Google searches now return an AI overview. McKinsey estimates that 20 to 50 percent of open-web search traffic converts to AI search by 2030. More than 90 percent of advertisers already use AI to plan and buy.
Performance is a data state of freshness, robustness and the ability to meet customer expectations. and the measurable outcome — efficacy, durability, conversion — evaluated as a checkable result rather than a delivered impression. “Thirty percent faster” stops being a tagline and becomes an attested figure.
Machine Buying, Autonomous Delegation, and Agent Revenue Models
Agents use five variables to drive buying and earn revenue. Anything you’d once stage across funnels needs to align with agent evaluation and purchase.
Capability Variables
Ranking
-Sponsored options shown alongside ranking for context completeness × trust
-Bidding to rank higher among provable options
Knowledge — from training data.
Retrieval — thick or thin context
-Pays for better data and verification
Relationship Variables
Governance — platform policy.
Commerce — merchant integrations and payment incentives.
-Pay to be the default
-Affiliate commissions that edge out more qualified options
I make the case that influenced shopping agents won’t work for long because of competition, low switching closts, and transparency.
Disclosed influence is likely to become the standard.
Signed data and low switching costs reveal agent bias that creates self-correcting feedback loops due to lost business and trust.
Anyone can check the receipts, so most delegation doesn’t hide bias.
Undisclosed shopping manipulation is no better than Chrome with a kickback. It is likely to be exposed and competed away.
The Defense: Irreducible Meaning and Trust
Marketers need to distinguish meaning that can be automated from what survives automation—the irreducible. Rich product context drives brand value at the bottlenecks we already identified: desire, taste, identity, belonging, status, meaning, and fandom. Human-verified provenance drives sales value at data value chain touchpoints. When receipts are the new attention, the omnichannel playbook — campaigns, ads, funnel nurture — should pivot to structured data receipts.
Sell proof to Red Oceans. Sell meaning in Blue Oceans.
Pervasive structured data is inevitable, which means value creation lives in the irreducible. What should marketers do when the factual half of advertising compresses to provable claims? Structured data is the new minimum, but staying there keeps you in the land of Porter competition and red oceans.
Products contend on measurable attributes — cost, performance, footprint — and the agent selects on the evidence. Advantage belongs to the seller with the stronger verified record. I’m advising brands to head for irreducible meaning by building it into the product’s structured data context, not added by marketing later. In practical terms that means merging product, brand, audit, and agent readiness in longer planning cycles. We’re close to launching Economy3. Book a discovery call: scott@economy3.org.
Implications
Agents now drive half of web traffic. Your biggest customer is likely to be a machine that values only math and performance.
Agents run on structured data. You control only the product half.
Defensible brands are irreducible: desire, taste, identity, belonging, status, meaning, fandom. Everything else migrates to attestation, proof, and performance.
Brand recall and omnichannel presence fade into the agentic selection process.
Regulation and AI incentives make climate, compliance, culture and supply-chain tracing the new advertising;
Context thickness × verification density win the buy box.
Data value chains build the proof layer.
Agent forms and revenue models vary. Transparency and low switching costs self-correct for honest results.
Don’t bring new tools to a dissolving Web2 world.
Lost trust rarely returns.
Selection runs on AI evaluation of structured data, verified claims, and measurable outcomes — not on attention or bidding.
Sending thanks for inspiration and leadership to:
Howard Yu
Sangeet Paul Choudary
https://substack.com/@platforms
and https://reshufflebook.com/
Caroline Rothwell-Gerstein
https://substack.com/@caroconsulting
Economy3 solves climate, compliance, and culture-as-advertising for companies of every size. Write: Scott@Economy3.org.