Introduction
Jason Barnard, in a wide ranging final instalment of his AI authority series, has issued a clear warning to companies that treat search optimisation as an optional marketing channel. In a piece first published on Search Engine Land, Barnard argues that the new commercial battleground is not only on the public web but inside the engines and the agents that now act on behalf of buyers. His thesis is simple and brutal: when an agent buys for a customer, that agent becomes a client you must satisfy directly. See Barnard’s article for the full argument. (Author: Jason Barnard, source: https://searchengineland.com/ai-salesforce-selling-your-brand-481027)
Why SEO is now a business infrastructure, not just marketing
Barnard uses a Russian doll metaphor to describe the stack companies must reach: at the core sits traditional SEO and the web index; around that sits the large language model layer that reasons and converses; and on the outside sit agents with direct access to commerce systems. The implication is straightforward. The stronger your SEO foundation and your entity data, the more reliably every engine and assistant will recognise and recommend you. That means the work formerly labelled as marketing becomes engineering: product pages, pricing, checkout, customer success and post-sale evidence must all emit machine readable signals so an agent can transact without human intervention.
This is not speculative. Large language models such as ChatGPT and search indexes like Google form the intelligence and information layers most assistants pull from, while knowledge graphs provide the verification layer. Businesses that leave their entity home underpopulated or their structured data incomplete will pay for it in lost recommendations. For practical guidance on structured data, see Google’s developer documentation. For a commercial perspective on AI visibility tools, vendor pages such as Semrush’s AI SEO overview are useful to benchmark where you stand.
The funnel flips – build from the bottom up
One of Barnard’s most useful reframes is the funnel flip. Marketers traditionally build top down because their audience travels from awareness to purchase. In the age of assistive engines and agents the correct build direction is bottom up: lock your entity home, prove your credibility, then make yourself deliverable. Agents reason from the decision point backward; they will only recommend brands that clear understandability and credibility gates.
That flip is operationally significant. It forces businesses to prioritise the warmest traffic first: protect your brand SERP so users who search your name encounter no doubt; secure third party corroboration so agents have evidence they trust; and finally, win topical territory higher up the funnel. Barnard calls these leak points three taxes – doubt, ghost and invisibility – and they translate directly into lost conversions and hidden churn.
The agentic spectrum and the measurement gap
Barnard introduces two complementary ways to measure the shift. The delegation boundary tracks what share of a single buyer journey a machine handles. The agentic spectrum looks at what share of your total clientele is being handled by agents. Most firms have not quantified either. That is a strategic failure.
Without measurement you cannot prioritise. If your agentic share is small you may patch brand SERPs and structured data. If the share is already large you must reorganise product, pricing and fulfilment so agents can close sales cleanly and reliably. The practical step for any CEO or marketing director is to instrument the stack, measure agentic transactions, and treat the resulting data as commercially binding feedback.
Third party proof beats first party claims
A central, and often misunderstood, point in Barnard’s piece is the hierarchy of evidence. First party claims on your site are just that, claims. Second party corroboration moves the needle a little. Third party proof – independent journalism, analyst reports, customer-written reviews and partner testimonials – is the heaviest currency in the new sales economy. Agents trust what they cannot be seen to have been handed by the vendor.
That has real operational consequences. Collecting third party evidence requires real-world activity: structured customer interviews, sustained PR campaigns, analyst engagement and the discipline to publish machine readable proof from post-sale processes. This is not a content week task. It is an enduring function that needs budget, governance and KPIs.
Practical checklist for Irish businesses that want to win
1. Entity Home first – ensure your canonical brand page contains unambiguous facts, persistent identifiers and complete structured data. Treat it as your single source of truth.
2. Secure the brand SERP – monitor what engines say about you and remove hedging language and ambiguity before a buyer sees it. Use neutral test accounts and tracking tools.
3. Harvest third party evidence – design customer success processes so outcomes are documented in ways machines can read. Set a quarterly cadence to convert case studies into machine readable formats.
4. Measure agentic share – instrument transactions to identify which come from agents and how often they fail to complete a sale. Use that metric to prioritise engineering work.
5. Prioritise revenue-first taxes – fix doubt on warm traffic first, then the ghost tax against near-purchase comparisons, then invisibility at the top of the funnel.
6. Align product and commercial systems – price, inventory, fulfilment and return processes must be machine friendly. Agents will reject merchants with brittle or opaque flows.
A critique and a policy note
Barnard is right to push the organisation-level argument, but the policy debate matters too. Some commentators call for heavy regulation of assistant behaviour or of the major platforms that control knowledge graphs and indexes. That would be a political minefield and risks freezing innovation. A more constructive approach for firms is pragmatic adaptation. Invest to make your business legible to machines and insist on contractual and technical standards for any third party agents you integrate with. Regulation may come later; in the meantime failure to adapt is a self-inflicted commercial penalty.
Why this matters to Ireland
Irish companies sit at a moment of opportunity. Our economy is export oriented, digitally sophisticated and small enough that strategic shifts can scale quickly. Firms that treat Barnard’s prescriptions as operational priorities will convert a disproportionate share of global agentic transactions because adapting early buys both market share and bargaining power. For Irish B2B exporters and SaaS vendors, the cost of inaction is direct: agents will prefer competitors who present clean proofs and reliable checkout. That is a competitive problem that can be fixed by engineering, not pleading.
Conclusion
Jason Barnard’s final piece in the series reframes SEO as infrastructure, not an add-on. The salesforce of the internet is already at work in Google, ChatGPT and other assistants, and unless you train it to sell for you, it will sell for someone else. The prescription is organisational: measure the agentic share, shore up the entity home, harvest third party proof and build product systems so an agent can close a sale as reliably as a human. Read the original analysis for the full framework and series context (Jason Barnard, source: https://searchengineland.com/ai-salesforce-selling-your-brand-481027).
