• Member
• November 12, 2026

Software Based Advertising A Srijan Sanchar Foresight 2026 to 2035

Srijan Sanchar · Shailendra Jaiswal · 12 November 2026

Summary

Software-based advertising is moving from humans buying screens to AI agents negotiating with AI agents. In 2026 advertising is a $1.3 trillion market growing at 8.9%, and Alphabet, Meta and Amazon take 57.6% of it outside China. AI now runs campaigns end to end, chatbot answers are becoming the fastest-growing ad surface, and commerce ads have overtaken television. This inquiry began with InMobi, whose decision engine, sell-side agents and Glance surfaces show the selling side is already agent-ready. Read through IEF v3, the history of advertising reveals one tension recurring at every stage: whose interest the system serves. The forecast is a bifurcation between platform-loyal and person-loyal futures.

Two questions decide the future to 2035: whom the AI works for, and whether the rules are open. Today’s walled autopilot is most likely, at about 40%. The personal agent economy, where people’s own agents post needs and collect verified offers on public rails, is the most valuable, at about 15%. Assistant gatekeepers and an open machine market lie between. The central insight is order: open the rules before personalising the agent, because intelligence on the device that stays loyal to sellers becomes manipulation, and assistants built inside closed platforms become new toll gates.

Content now behaves like advertising, since both must reach the same receiver through the same agent. No single payment model wins. Hybrid tiers dominate to about 2030, after which fluid choice rises, letting people pay, watch or share at any moment through an attention wallet holding money, attention, data and contribution. Content is also fragmenting into clips and chapters, and ads are becoming serials told across many creators. Value therefore moves to the orchestrator, the algorithm that picks, sequences, prices and pays. The orchestrator becomes the new publisher, and value concentrates at both ends of the smile curve: original work and the relationship with the receiver.

The value of a moment is attention multiplied by capability to purchase and by trust. Access is moving from feeds to phone assistants between 2027 and 2032, because they are the first gate to see both intent and ability to pay. This creates risks of attention inequality, wealth guessing and exploiting children, which calls for attention neutrality. Structural Threshold Foresight places the ad-funded free model in a pre-threshold state, five to ten years from breaking. It also signals receiver assurance: independent proof that privacy, children and fair conduct are protected. This is wider than cybersecurity, because it guards against legitimate actors behaving badly, not only attackers.

For Srijan Sanchar this is a thought-leadership opportunity grounded in its Dharma gate. Twelve novel claims are ready to publish, from “the ad becomes the accepted offer” to “attention neutrality”. Recommended actions stay exploratory while the threshold is uncertain: a Dharma Assurance standard, a child-safe toy pilot with TAI before India’s children’s data rules apply in May 2027, a fluid-choice pilot on Katha Setu, and a convening of industry, regulators and child-safety voices. India’s UPI, consent managers and DPDP Act make it well placed to lead.

At a glance

Indicator (2026)	Value	Source
Global ad revenue	$1.3 trillion, +8.9%	WPP Media
Share taken by Alphabet, Meta, Amazon (outside China)	57.6%	WPP Media
Ads in AI search	$5.1B in 2026, $100B+ forecast by 2030	WPP Media
Ad plans’ share of new Netflix sign-ups in ad markets	Over 60% (Q1 2026)	PPC Land
Advertiser-side verification (DoubleVerify)	$748M revenue, 2025	Company results

The core foresights.

1.         Advertising becomes negotiation. The unit shifts from the impression to the offer an agent accepts.

2.         Two questions decide the future: whom the AI works for, and whether the rules are open. The walled autopilot is the default (about 40%); the personal agent economy is the highest-value path (about 15%).

3.         Open the rules before personalising the agent. Device intelligence without loyalty to the person is a forbidden zone.

4.         Content and advertising converge. Fluid choice through an attention wallet becomes the leading payment model.

5.         The orchestrator is the new publisher. Algorithms weaving fragments and serial ads capture most of the value.

6.         Access moves to OS-level assistants (about 2027–2032).

7.         The ad-funded free model is pre-threshold (5–10 years).

8.         Receiver assurance becomes a major function, wider than cybersecurity.

Claims Srijan Sanchar can stake: Edge without loyalty is forbidden · The ad becomes the accepted offer · Best execution comes to advertising · India can do for advertising what UPI did for payments · The orchestrator is the new publisher · Ads become serials, not spots · The fragment needs a royalty, like the song · Every fragment is an ad for the next · One auction for every moment · Attention neutrality · Assurance is wider than cybersecurity · Children need protection in content, not only in ads.

1. Starting point: InMobi through IEF v3

InMobi has built the sell side of agentic advertising. IEF v3 shows the open fork is whether software advertising stays loyal to platforms or turns loyal to the person.

What InMobi has built (2025–26). The Mobius decision system makes 25T+ predictions a day across 200B+ ad opportunities and leaves a “decision trace”. A sell-side agent built with Scope3 opened mobile and CTV supply, including Glance’s ~300M active devices, to AI buying agents. A conversational agent on Buyer Hub serves traders. InMobi acquired MobileAction, and its founders bought out SoftBank ahead of an expected India IPO. What remains open: no agent represents the person, consent is a cost rather than a priced asset, and nothing proves an agent acted for its principal.

IEF step	Result
Phenomenon and objective	Software advertising after agent-to-agent buying; horizon 2027–2031; act on theses at medium probability or higher
Thesis inventory	22 theses across five tiers, with 2 axis candidates, 4 matrix seeds, 3 quadrant flags and 1 forbidden zone
CCN genealogy	Five nodes (banners → search → programmatic → privacy turn → agentic); one polarity recurs: whose interest the system serves
Constraint field	C1 auditable unit of exchange; C2 tie the ad to expressed intent; C3 verifiable supply path; C4 revocable, purpose-limited consent and no child targeting; C5 every autonomous decision traceable with human override
Axes	Axis 1: Agent loyalty (platform → advertiser → mutual → person). Axis 2: Intelligence locus (cloud → on-device)
Matrix reading	Two clusters with a gap; forbidden cells at the device end of the platform and advertiser rows: edge without loyalty is forbidden
Forecast state	Bifurcation: platform cluster ~55%, person cluster ~30% by 2031
Target cell	Consent-priced intent exchange: the person’s on-device agent posts intent windows, sellers’ agents bid with verifiable offers, a share of value funds content
Trajectory	Loyalty shifts before locus: best-execution receipts → consent-artefact signals → negotiated offer protocol → target

2. Global context in 2026

Advertising is a $1.3 trillion market growing faster than the economy. Three firms take most of it, and AI is both its growth engine and its biggest disruptor.

Fact (2026)	What it tells us
Global ad revenue grows 8.9% to $1.3T, its highest share of GDP since 1999	AI-run performance ads drive growth
Alphabet, Meta and Amazon take 57.6% outside China	Software platforms now own advertising
Commerce ads reached $178B in 2025 and overtook TV	Ads move to the point of purchase
Generative-search ads: $5.1B in 2026, $100B+ by 2030	AI answers are the fastest-growing ad surface
ChatGPT ads launched Feb 2026; $100M+ annualised in six weeks; no ads to under-18s	Chatbots adopt ads, with stated limits
Meta aims to automate ads end to end	The media buyer is automated away
Rival agent-buying standards: AdCP (2025) and IAB Tech Lab’s AAMP (2026)	Agent trading is real; the rules are contested
Google kept third-party cookies and retired most Privacy Sandbox tools (Oct 2025)	Privacy moves to law and device
US court (Sept 2026): Google keeps AdX but must open its ad tech, under six years’ oversight	Regulators open the plumbing rather than break up platforms

 

#	Force reshaping advertising	Pulls toward
1	AI runs the whole campaign	Platforms
2	Answers replace search results	AI owners
3	Commerce absorbs advertising	Platforms with purchase data
4	Agents trade with agents	Undecided: depends on the rules
5	Screens multiply: TV, lock screen, car, glasses	Device makers
6	Privacy moves from browser to law and device	People, if law holds
7	Regulators open the plumbing; children’s rules tighten	Challengers and publishers
8	Trust becomes scarce as AI content floods feeds	Whoever can prove authenticity

3. Four futures to 2035

Two questions decide the future: whom does the AI work for, and are the rules open or closed? The walled autopilot is the most likely; the personal agent economy is the most valuable, and it is reached by opening the rules first.

Four futures to 2035: who the AI serves and how open the rules are

•           Walled autopilot (about 40%): the default path. Ads get cheaper and better targeted, but a few firms set prices and publishers lose income.

•           Assistant gatekeepers (about 25%): assistants from phone makers and chatbot firms shop for users, but sellers pay a toll to be seen.

•           Open machine market (about 20%): open protocols and antitrust remedies let many sellers compete, but agents still serve the seller.

•           Personal agent economy (about 15%): people’s own agents negotiate on open rails, needing open protocols, consent law and trusted public infrastructure, all of which India has in other sectors.

4. Ten foresights with timing

#	Foresight	Roughly by	Confidence
1	The media buyer becomes a supervisor	2027	High
2	AI answers become the biggest new ad surface ($100B+)	2030	Medium-high
3	Agents trade with agents by default	2029	Medium-high
4	The ad becomes the accepted offer	2030	Medium
5	Best execution comes to advertising	2030	Medium
6	Free content needs new funding (AI licensing, ad revenue shares)	2030	Medium
7	Personal shopping agents go mainstream	2031	Medium
8	Proof of authenticity earns a premium	2031	Medium
9	Consent becomes a priced, portable asset	2033	Low-medium
10	Public rails for advertising emerge, like UPI	2035	Low-medium

Foresights 1–3 are under way and favour platforms; 4–7 are the turning point; 8–10 decide whether people and small firms gain power, and depend on choices made before 2029.

5. Content and advertising converge: who pays for attention?

Content now behaves like advertising: it must reach a receiver, and soon the receiver’s agent decides what gets through. No single payment model wins. By 2035 the leading model is fluid choice: at any moment the receiver, or their agent, pays with money, attention, data or contribution. Ad-supported plans were 57% of new US streaming sign-ups in Q1 2025, and Netflix’s ad plan took over 60% of Q1 2026 sign-ups in its ad markets.

Model	How the receiver pays	Outlook to 2035
Pay for content	Money	A stable premium layer; subscription fatigue limits it
Watch ads, get content free	Attention and data	Largest by reach, but each ad earns less
Hybrid tiers	The receiver picks a plan	Dominant to about 2030
Fluid choice	Per piece or per minute, in the moment	Rises after 2029 where instant payment rails exist; India well placed

Other models will coexist: micropayments (₹1 per story), paying with consent, your agent paying creators from purchase commissions, AI licensing, commerce-funded content, patronage, public and CSR funding, earning to view, and paying with contribution.

The attention wallet. Each person holds four currencies: money, attention, data and contribution. Their agent pays for each piece of content in the cheapest acceptable currency, within limits the person sets. Content will have to pitch like an ad, and ads will have to deliver value like content. Children’s content stays ad-free in every model.

6. Fractional content, serial ads and the orchestrator

Content is breaking into fragments, and one brand message is split into episodes woven across many creators’ content. This needs coordination, so value moves to whoever orchestrates: the algorithm that picks, sequences, prices and splits the payment. The orchestrator is the new publisher.

Fractional content and serial ads: the orchestrator in the middle

Shift	What changes	Who gains, who loses	Roughly by
Fractional content	The unit of consumption and payment becomes the fragment	Gains: snackable formats. Loses: long-form makers unless valued per fragment	Under way
Serial ads across content	Campaigns become episodes measured by whether the story completed	Gains: brands with story skills. Loses: single-spot sellers	2028–30
More coordination	Sequence, frequency, rights and payment splits agreed across parties	Gains: protocols, clearing houses	2028–31
Value shifts to algorithms	The orchestrator keeps most of the margin	Gains: orchestrator owner. Loses: creators’ bargaining power	Already visible

If the platform’s algorithm orchestrates, control concentrates further; if the person’s own agent orchestrates on the device, ads can be woven in without cross-app tracking and creators paid directly. The shift needs fragment royalties (like IPRS for songs), serial-ad rules, orchestrator accountability and creators’ right to refuse unsuitable placements.

7. Attention × capability to purchase × trust: who controls access?

The scarce resource is qualified attention. Value of a moment ≈ Attention × Capability to purchase × Trust. Content, its own promotion and others’ ads now sit on one spectrum, from pure content, through content advertising itself (“every fragment is an ad for the next”), others’ ads inside content and brand-made content, to ads good enough to be chosen. One auction decides every moment: a fragment bids with engagement, an ad with money, a self-promotion with future subscriptions.

Gate to access	Controlled today by	Contest to 2035
The person	Attention, time, consent	Gains power only with their own agent and enforced consent
Device and OS	Apple, Google, phone makers	Strongest gate as assistants move into the OS
AI assistant	Chatbot and phone-assistant makers	The new front door: sees intent and capability together
Feed and orchestrator	Video, social, streaming platforms	Powerful, increasingly reached through assistants
Search and discovery	Search engines, AI answers	Shrinks as answers replace visits
Payment and identity rails	App stores, card networks, UPI	Decides whether fluid choice is possible
Network and bundles	Telecom operators	Strong in India
Rules	Competition, consent, child-protection law	Decides whether gates stay open and neutral

Control shifts toward OS-level assistants, roughly 2027–2032; in India, super-apps and telecom bundles may hold the gate. Risks include attention inequality, capability scoring, monetising children through parents, blurred labels and closed gates. The remedy proposed is attention neutrality: no gate may favour its owner’s content or ads without disclosing it.

8. Receiver assurance: wider than cybersecurity

As agents decide what reaches people, a new function becomes necessary: independent assurance that receivers are not tracked beyond consent, children are not influenced, and algorithms behave as they claim. We expect it to become a major function by about 2030 (high confidence); whether it stays independent is open (medium confidence).

India’s DPDP Rules were notified on 13 Nov 2025; consent managers register from 13 Nov 2026, and fiduciary and children’s obligations apply from 13 May 2027. The EU’s DSA Article 28 bars profiling-based ads to known minors. Yet a 2026 audit of TikTok found formal compliance alongside promotional content profiled to minors 5–8 times more strongly than formal ads to adults. Today’s verification industry (DoubleVerify $748M in 2025; IAS $530M in 2024) protects the advertiser, not the receiver.

#	Assurance domain	Question it answers
1	Privacy and consent	Was data used only within consent, and did withdrawal take effect?
2	Child safety	Were children kept out of profiling and persuasion, in ads and content?
3	Disclosure	Was every paid, sponsored or AI-inserted placement labelled?
4	Agent conduct	Did the agent act for its principal and leave a trace?
5	Orchestrator neutrality	Did the gate favour its owner without saying so?
6	Non-manipulation	Were grief, debt, illness and addiction signals left unused?
7	Fairness	Were ad load, offers and prices fair across income and language?
8	Provenance	Is the content authentic and traceable?

Assurance evolves in three phases: compliance assurance (2026–28), behavioural assurance testing what receivers actually saw (2028–31), and continuous machine-speed assurance with device attestation (2031–35). It could settle as captured (platforms certify themselves), regulatory (solid but slow), or independent and person-centred, the only outcome in balance. Liability, verifiability and an independent payer decide which. Receiver assurance turns Srijan Sanchar’s Dharma gate into a category it can lead: Dharma Assurance.

9. Framework lens

Six Srijan Sanchar frameworks were applied alongside IEF v3. They added rigour, timing and an action discipline, and confirmed the direction more than they discovered it.

CRRAAAS. The four gates map onto the value formula: Ability is capability to purchase, Trust is permission, Agent is who controls access, Mode is the format. The ten operators are what orchestrators now automate. Two-layer design, a machine-legible core plus a human layer, becomes the new SEO.

Structural Threshold Foresight. The ad-funded free model is pre-threshold (5–10 years): absorptive capacity eroding, legitimacy decay at stage 3 of 4, narrative frame qualifying, elite consensus strained (the swing layer).

Balanced Value Chain (Smile Curve). AI commoditises the middle of the chain; value moves to original work and to the orchestrator and relationship.

Smile curve: content and advertising value chain, today vs 2035

HEM. Only the personal agent economy keeps Integrity, Impact and Vitality in harmonic balance.

Cross-domain isomorphisms (qualitative). Electricity grids (open access), music collecting societies (fragment royalties), stock exchanges (fair ordering, best execution), UPI (public rails), spectrum allocation (ad-load limits, a protected band for children), and financial audit (assurance as a profession).

Thought–Action Network. With the threshold 5–10 years away, hold mostly exploratory and provisional actions and avoid irreversible lock-in.

What the frameworks added. STF and TAN added the most new substance; the Smile Curve sharpened the value-shift finding; CRRAAAS was partly confirmatory; HEM mostly relabelled. Scores were made after a view formed, so a blind, framework-first test would measure their contribution properly.

10. Action plan, Dharma guardrails and signposts

#	Action	Who	Intentionality	When
1	Publish the thought-leadership claims and the assurance note	Srijan Sanchar	Committed	Now
2	Convene adtech, OEMs, consent managers, auditors, creators and child-safety experts	Srijan Sanchar	Provisional	Early 2027
3	Draft a Dharma Assurance standard	Srijan Sanchar	Provisional	Q1 2027
4	Best-Execution Receipt standard	Indian adtech, agencies	Provisional	Mid-2027
5	Child-safe toy discovery and assurance pilot	TAI members, Srijan Sanchar	Exploratory	2027
6	Katha Setu fluid-choice and per-chapter pilot	Katha Setu	Exploratory	2027
7	Consent-artefact intent signals on DPDP rails	Consent managers, a sell-side platform	Provisional	2027–28
8	Negotiated Offer Protocol	OEMs, exchanges, open networks	Provisional	2028–29
9	Fragment royalty clearing house	Creators, platforms, collecting societies	Exploratory	2028–30
10	Avoid lock-in or exclusive deals until the threshold is clearer	All	Avoid as irreversible	Ongoing

Dharma guardrails, built before misuse.

•           No profiling or targeting of children, in ads or content; no monetising children through parents.

•           No use of grief, debt, illness or addiction signals.

•           Always disclose the AI and whom it serves; label every paid or AI-inserted placement.

•           Paid placement never changes what an AI states as fact.

•           Consent revocable at once, with the withdrawal logged.

•           Every autonomous decision leaves a receipt and allows human override.

•           Agents tested for tacit price coordination; attention neutrality at every gate.

•           Equal offer quality across languages; ad-load limits for those who cannot pay.

•           Assurance answers to the receiver, tests behaviour not paperwork, and never needs more data than the harm it prevents.

Gate question for every initiative: would the person’s own agent, acting in good faith, accept this, and would the receiver endorse having been reached this way?

Signpost	If it appears, it means
OS makers let assistants filter or refuse ads	Personal agents are arriving
One agent-trading protocol wins clear adoption	The open machine market is forming
Chatbot ad revenue passes $10B a year	Answer-engine ads are mainstream
A regulator requires proof that buying agents acted for their client	Best execution is coming
Consent managers carry shopping or content intent	The personal agent economy is opening
An open split over who pays creators for AI answers	Threshold imminent (1–5 years)
A compliant platform is shown still profiling children through content	Behavioural assurance becomes mandatory

11. Sources

•           WPP Media, This Year Next Year 2026 Midyear Forecast, Jun 2026 — https://www.wppmedia.com/news/report-this-year-next-year-midyear-2026

•           PPC Land, WPP Media sees AI as advertising’s $1.3 trillion growth engine, Jun 2026

•           PPC Land, ChatGPT ads hit $100M in six weeks, Mar 2026

•           Influencer Marketing Hub, Meta’s plan to reinvent digital ads with AI, Jun 2025

•           PPC Land, AAMP and AdCP overlap on 13 functions, Aug 2026

•           PPC Land, Chrome kills most Privacy Sandbox technologies, Oct 2025

•           Tech Policy Press, Antitrust remedies that spared Google’s ad tech business, Sep 2026

•           DestinationCRM, InMobi introduces Mobius, Oct 2026; Business Wire, InMobi conversational AI agent, Aug 2026; Adweek, InMobi and Scope3 sell-side agent, Jun 2026; Forbes, InMobi’s MobileAction deal, May 2026; IPOCentral, InMobi’s buyback before IPO

•           eMarketer (Antenna data), Streaming growth driven by ad tiers; PPC Land, Netflix Q1 2026 results, Apr 2026

•           S.S. Rana & Co., DPDP Rules 2025, Nov 2025; Digital Services Act, Article 28

•           KInIT, The DSA’s Blind Spot: algorithmic audit of minor profiling on TikTok, ACM FAccT 2026

•           DoubleVerify full year 2025 results, Feb 2026; Digiday on ad verification, Mar 2025

•           Srijan Sanchar: IEF v3.0; CRRAAAS; Structural Threshold Foresight; Balanced Value Chain; HEM; ICDIDF; Thought–Action Network (srijansanchar.com/blogs)

•           Online version with live links: https://claude.ai/artifact/NDtcBvvyfnTpAdCibnQhsy

Probabilities, timings and framework scores are Srijan Sanchar analytical judgements. Glance reach figures are company claims. This document is not legal or financial advice.