IAB Australia today published a report warning that agentic AI has moved from speculative concept to operational reality for search marketers, and that the infrastructure most brands rely on was not built to handle it. The report, released on June 18, 2026 by the IAB Australia Future of Search Working Group, runs 14 pages and provides a framework that agencies and brands can use to prepare for a search environment where the click may no longer be the transaction's first step - or any step at all.

From response to execution: what actually changed

The report's central argument is precise. According to IAB Australia, agentic AI marks a shift from AI as a "response engine" that finds information to an "execution engine" that completes tasks. The distinction is not semantic. A large language model answers a question. An agentic system interprets intent, breaks the request into sequential steps, takes action across multiple environments, verifies the outcome and reports back - with limited human input at any stage.

The practical consequence for search marketing is stark. Discovery, consideration and purchase - the three stages that traditionally define the purchase funnel and the three stages that justify search advertising budgets - can now collapse into a single sequence of seconds, driven by an agent acting on behalf of the user. The IAB Australia report describes this as a compression of the traditional funnel: an agent browsing on behalf of a consumer does not pause at each stage waiting for a human decision.

This is a continuation of themes the IAB Australia Future of Search Working Group addressed in its previous paper on instant answers and discovery, which PPC Land covered in December 2025 [https://ppc.land/iab-australia-report-finds-instant-answers-reshape-search-measurement/]. That earlier paper examined how zero-click search experiences were already pushing marketers to abandon traffic-based metrics. The new report goes further. Where the December 2025 paper addressed what happens when humans stop clicking through, today's report addresses what happens when agents handle the entire journey without human clicks at any point.

The agentic stack: four protocols doing different jobs

The technical architecture underpinning agentic AI in search is made up of four emerging protocol layers that the report maps in detail, each solving a distinct problem.

Agent-to-Agent Protocol (A2A), developed by Google, handles how agents from different platforms find each other, establish trust and collaborate. It provides a standardised way for one agent to pass a task to another, handles authentication so sensitive data such as payment credentials is handled securely, and allows what the report calls "delegation" - one agent hiring a specialist agent rather than attempting everything itself.

Model Context Protocol (MCP), developed by Anthropic along with OpenAI and others, covers how agents connect to external data and tools. It simplifies the connection between an agent and an outside database or API, eliminating the need to write separate integration code for each data provider. PPC Land has covered MCP extensively in the context of advertising technology: the Ad Context Protocol launched in October 2025 was built directly on MCP and brought nine core tasks for autonomous campaign management to the ad tech sector.

NLWeb, developed by Microsoft, addresses how agents interact with existing website content. Rather than inferring what a website contains from a summary - which, as the report notes, creates a risk of hallucination - NLWeb allows the agent to pull information directly from the source's database. The result is more accurate interpretation and the ability for websites to expose specific "actions" to agents, such as booking, tracking a shipment or adding an item to a cart, without the agent needing to locate a button on a screen.

Universal Commerce Protocol (UCP), a Google-led initiative, handles agent-to-commerce transactions. According to IAB Australia, it lets an agent talk directly to a store's backend to add items to a cart without any clicking, links loyalty programs so the agent automatically applies member discounts, and standardises how payment and shipping information is passed between agent and retailer - turning what the report describes as "a 10-minute manual checkout into a 2-second confirmation."

According to IAB Australia, the agentic stack "is still in its infancy" but specific approaches are consolidating as the most likely building blocks. The most traction so far has been in defining how agents talk to each other, transact on behalf of users and are governed. Google's Universal Cart, which works across merchants and services and can be added to while browsing Search, chatting with Gemini, watching YouTube or reading Gmail, is cited in the report as a concrete recent deployment in this space. The agentic ad tech week roundup published today by PPC Land confirms the pace of infrastructure development: DoubleVerify launched a cognitive AI engine with autonomous execution capabilities on the same day, and LiveRamp opened its data platform to third-party agent builders.

Consumer behaviour: 60% of Australians expect to use agents daily

The report draws on several external data sources to establish the scale of the consumer shift. According to a Microsoft study cited in the report, 58% of consumers are ready for AI agents to replace some of their traditional website search, particularly for tasks like shopping. A separate figure from the same study shows that 68% of respondents say they either like using automation for simple tasks or "love it" for making life easier and more efficient.

The Australian-specific numbers are higher. According to an Adobe report cited in the IAB Australia document, 60% of Australian respondents said they expect to be using agentic AI in their daily lives. Gartner's annual report adds a forward-looking dimension: according to the report, Gartner predicts that by 2028, 70% of customer service journeys will begin and be resolved in conversational AI through third-party assistants built into mobile devices.

But consumer readiness and consumer trust are not the same thing. The report is careful to keep them separate, and the gap between them is one of the document's more instructive data points. According to an XM Institute study cited in the report, only 14% of respondents in Australia trust organisations to use AI responsibly. When asked what they are most concerned about if companies use AI to automate customer interactions, the top two responses were lack of humans to connect with, cited by 61%, and misuse of personal data, cited by 56%. Furthermore, 66% of respondents say the ability to manually override or stop an automated action is critical.

The report frames this as a structural constraint on the agentic opportunity, not a temporary sentiment problem. Consumers are willing to delegate low-stakes tasks to agents. They want transparency about what decisions are being made on their behalf and a mechanism to intervene. Brands that do not build those mechanisms into their agentic infrastructure will face a trust deficit they cannot correct through messaging alone.

Discoverability replaces awareness at the top of the funnel

One of the most significant structural shifts the report identifies is what happens to brand awareness in an agentic world. Traditional awareness - being remembered by a human consumer who then decides to search for a brand - loses relevance when the agent makes the discovery decision. The report introduces "discoverability" as the replacement: can an agent find a brand, read its product information accurately and evaluate it against the task it has been given?

According to IAB Australia, agents evaluate brands based on existing data points and user preferences, not on emotional associations or creative impressions. The emotional factors that differentiate brands for human consumers "are reduced if not eliminated entirely for more factual and credential-based criteria." A brand that is not structured for machine-readability will not be considered, regardless of creative quality or media investment.

This connects directly to the question of organic search. According to the report, before pursuing Generative Engine Optimisation (GEO) or Answer Engine Optimisation (AEO), organisations need strong SEO and website fundamentals in place. Both Microsoft and Google, according to the report, emphasise that success in AI-driven search is built on core principles: delivering high-quality and relevant content, ensuring fast and accessible experiences across devices, maintaining strong security and demonstrating depth or authority in a specific area. Generative search experiences use techniques including retrieval-augmented generation (RAG), grounding and query expansion to surface the most relevant and up-to-date content.

Visibility in the agentic era increasingly depends on being "agent-ready," which the report defines as using structured and semi-structured data such as schema.org, JSON-LD and product feeds so AI systems can accurately interpret and surface content. This sits within a broader GEO strategy that ensures discoverability not just in traditional search but across chat-based and AI-powered experiences.

The IAB Australia December 2025 report introduced the E-E-A-T framework and NLWeb as relevant technical tools for visibility. Today's report extends that analysis into paid search and commerce, adding product feed infrastructure to the list of urgent priorities.

For search marketers running paid campaigns, the report's section on paid search contains the most technically specific guidance. The analysis contrasts traditional product feeds with what agentic search requires, and the differences are substantial.

A traditional product feed is built to drive a click to a site. Its success metric is click-through rate. Titles are written to be keyword-rich and attention-grabbing. Attributes such as size and colour are treated as optional. Inventory is updated daily or hourly.

An agentic product feed has a different primary goal: to execute a conversion inside a chat or have an agent complete the purchase on the user's behalf. The success metric shifts to what the report calls Agentic Conversion Rate (ACR). Titles need to be simple and factual rather than keyword-led. Every attribute is mandatory. Inventory must be updated in real time via API sync, because an agent checking availability cannot act on yesterday's data.

The difference in title strategy is particularly consequential. An agent does not respond to persuasive or emotive language. It reads factual signals. A product title that is keyword-stuffed and catchy may perform well in traditional search but may confuse or mislead an agent attempting to match a product against a user's stated requirements. Plain and complete attribute data is not a nice-to-have for agentic commerce - it is the prerequisite for being included in any consideration set.

New metrics: the four Cs replace CTR

The report proposes a measurement framework specifically designed for agentic activity, structured around four stages. Traditional metrics such as click-through rate, bounce rate and session-based conversion rate are described as becoming irrelevant for evaluating agentic commerce, because the human who would generate those signals may not be the one taking the action.

Consideration asks whether the brand was discovered and included in the agent's shortlist. The suggested metric is the consideration set or shortlist rate. Confidence asks whether the interaction earned the user's trust in the agent's recommendation. The suggested metric is recommendation acceptance rate. Completion asks whether the agent fulfilled the user's intent. The suggested metric is agent conversion rate. Continuation asks whether the session deepened brand loyalty or led to the next logical action - a metric that will vary by objective.

According to IAB Australia, new measurement approaches should sit alongside existing methods such as incrementality testing and marketing mix modelling rather than replacing them entirely. The priority sequence the report recommends starts with auditing machine-level discoverability, then modernising product and content data, then setting measurement priorities based on business model - commerce and retail brands focusing on real-time product, pricing and availability data; publishers focusing on content quality, structure and machine discoverability; product and service brands focusing on completeness and clarity of service information.

Trust as infrastructure: the audit log requirement

The report's section on consumer trust goes beyond data privacy language into operational specifics. According to IAB Australia, building trust in the agentic era means recording not just the agent's actions but also what the report describes as the "prompts, decisions, internal state changes, and intermediate reasoning" that led to those behaviours. This is essential for auditability, root cause analysis and maintaining consumer confidence when errors occur.

The operational implication is that brands deploying agents need audit logs that capture reasoning, not just outputs. The 66% of Australian consumers who say the ability to manually override or stop an automated action is critical translates into a specific infrastructure requirement: a mechanism that lets users pause or reverse an agent's decision before or after it is executed.

The report also identifies proactive service recovery as a trust-building mechanism. Spotting a delayed shipment and issuing a refund or discount before the consumer contacts support is given as an example of how agents can demonstrate their value by acting in the consumer's interest unprompted.

For the advertising and marketing technology sector, this trust infrastructure connects directly to the broader regulatory environment. PPC Land has covered the intersection of agentic AI and data governance across multiple jurisdictions: Spain's data watchdog published a 71-page guide on agentic AI and GDPR compliance in March 2026, and UK regulators warned in April 2026 that agentic AI is already present in production systems and needs oversight frameworks now.

Collaboration across marketing systems

The report identifies one more dimension of the agentic shift that does not fit neatly into the consumer-facing discussion: how agents change internal marketing operations. According to IAB Australia, agents will help marketers and agencies run media more effectively, save time on tactical tasks, understand data across ads, analytics, feeds and merchant centres, and reduce the workload of moving data between tools. Agents running across systems will be able to talk to each other and cut the time currently spent extracting data from one platform and re-entering it in another.

The IAB forecast for 9.5% US ad growth released in January 2026 cited agentic AI as a central driver, with two-thirds of advertisers already concentrating on agentic systems for ad buying and campaign execution. The operational efficiency argument in the IAB Australia report aligns with that picture: the investment rationale for agentic tools inside marketing teams is not only consumer-facing but also operational.

Six priority actions for agencies and brands

The report's closing section organises its recommendations into six priority actions. The first four are rated high priority.

The first action is to identify where agent-led interactions are most relevant - specifically, to map the customer journey and find where an agent is likely to act for the user at discovery, comparison, purchase or post-purchase support. For paid media, modernising product feed infrastructure is positioned as the first step, because agent-led conversion depends on it.

The second action is to set priorities based on business model. Commerce and retail brands should focus on accurate, real-time product, pricing and availability data. Publishers and content owners should focus on content quality, structure and machine discoverability. Product and service brands should ensure their information is clear and complete enough for agents to evaluate correctly.

The third action is to start building consumer trust through transparency and control mechanisms, including audit logs, override options and proactive service recovery.

The fourth action is to audit machine-level discoverability - checking structured data such as schema.org, JSON-LD and product feeds for accuracy and completeness, testing how the brand is represented across major AI search and chat interfaces and fixing gaps.

The fifth action, rated medium-to-high priority, is to modernise product and content data: upgrading product feeds to meet agent requirements with complete attributes, plain factual titles and real-time inventory via API rather than daily batch uploads.

The sixth action, rated medium priority, is to monitor leading markets such as the US and UK, which receive earlier access to platform releases and AI-powered search features, and to factor local market conditions into any adaptation plans.

Why this matters for the marketing community

The IAB Australia report matters for search marketing professionals because it puts numbers and an operational framework on a shift that has until now been discussed largely in abstract terms. The 14% trust figure for Australian consumers is a specific constraint. The 60% daily agentic AI usage expectation is a specific timeline signal. The contrast between keyword-rich traditional feed titles and plain factual agentic feed titles is a specific technical requirement. The four-stage measurement framework - Consideration, Confidence, Completion, Continuation - is a specific replacement for CTR-based reporting.

PPC Land has tracked the infrastructure side of the agentic shift across multiple sectors over the past year. The IAB Tech Lab Summit in May 2026 brought nearly 400 advertising and technology professionals together to address agentic AI standards, publisher monetisation and machine-to-machine advertising infrastructure. Optable published a six-pillar self-assessment tool on June 16, 2026 for publishers to evaluate their readiness for agentic advertising. HUMAN Security's 2026 State of AI Report, covered by PPC Land in April 2026, found that automation is now growing eight times faster than human web traffic.

What the IAB Australia report adds to this picture is the consumer-side and search-specific dimension, written specifically for the agencies and brands that buy search advertising rather than for the technology vendors building the infrastructure. The report is available to download from the IAB Australia website.

Timeline

  • December 4, 2025 - IAB Australia's Future of Search Working Group publishes its previous report on instant answers and zero-click search, introducing E-E-A-T and NLWeb as tools for AI search visibility.
  • January 28, 2026 - IAB publishes its 2026 Outlook Study forecasting 9.5% US advertising growth, with agentic AI cited as a central driver and two-thirds of advertisers focusing on autonomous campaign execution. PPC Land coverage.
  • March 1, 2026 - Spain's AEPD publishes a 71-page guide mapping GDPR risks of agentic AI, citing MCP and A2A as the protocols enabling systematic connectivity between agent systems. PPC Land coverage.
  • April 1, 2026 - UK regulators warn that agentic AI is already operating in production systems and that oversight frameworks are needed now. PPC Land coverage.
  • April 21, 2026 - HUMAN Security expands Agentic Visibility capabilities to marketing and commerce organisations, reporting that automation is growing eight times faster than human web traffic. PPC Land coverage.
  • May 20, 2026 - Google launches Ask Advisor at Google Marketing Live 2026, a unified Gemini-powered agent spanning Google Ads, Analytics, Merchant Center and DV360. PPC Land coverage.
  • May 21, 2026 - Google SVP Nick Fox, in an interview covered by PPC Land, confirms that Google's agentic commerce infrastructure is designed to reduce friction in discovery and checkout while keeping the purchase decision with the human. PPC Land coverage.
  • May 28, 2026 - IAB Tech Lab Summit brings nearly 400 advertising and technology professionals together to address agentic AI standards, publisher monetisation and machine-to-machine ad infrastructure. PPC Land coverage.
  • June 16, 2026 - Optable publishes a six-pillar self-assessment framework for publishers to evaluate agentic advertising readiness, noting that only 20% of marketers had begun using AI agents as of September 2025. PPC Land coverage.
  • June 18, 2026 - IAB Australia Future of Search Working Group releases "Agentic AI: What Does It Mean for Search Marketers and How to Prepare?" - the source document for this article.

Summary

Who: The IAB Australia Future of Search Working Group, comprising agency and industry representatives.

What: A 14-page report titled "Agentic AI: What Does It Mean for Search Marketers and How to Prepare?" covering the technical architecture of agentic systems, consumer trust data specific to Australia, implications for traditional paid and organic search journeys, a new four-stage measurement framework and six priority actions for agencies and brands.

When: Published on June 18, 2026.

Where: The report addresses the Australian market specifically for consumer trust and behavioural data, while the technical framework and protocol stack it describes - A2A, MCP, NLWeb and UCP - are global.

Why: Agentic AI compresses the purchase funnel from a multi-step human journey into a seconds-long automated sequence. Traditional search marketing infrastructure - product feeds built for clicks, metrics centred on CTR, content structured for human readers rather than machine parsing - is not compatible with this model. The report was produced to give search marketers a practical roadmap for adapting before the technology and consumer adoption curves make the gap impossible to close.