Who Owns AI Discoverability Inside Your Organisation

I have watched this pattern play out in enterprise organisations for twenty years. A new discipline arrives, nobody owns it, and the people who should be worrying about it rank it lowest on their list. Two years in, someone on the board asks why the brand looks the way it does on the surface everyone has moved to, and the finger-pointing starts.

The latest version is AI discoverability. The signals a model picks up about you: the sitemap it reads, the schema it parses, the crawler it waves through or blocks by accident, the narrative it stitches from the live site plus the three subdomains your business units quietly launched since 2019. There is work to do. There is, in most organisations, nobody doing it.

Jodi Cerretani, CMO at WordPress VIP, put a number on it on the record. In the second of two webinars WordPress VIP ran with 97th Floor this year, only 5% of her audience reported a clear owner. In the first session, the same audience had ranked ownership as the lowest of their concerns. The gap between what needs owning and who thinks about owning it is this piece’s subject, and it is a procurement problem before it is a technical one.

If you are a CMO or a digital lead reading this because you suspect no one in your organisation can answer “who owns this”, you are already ahead of 95% of that webinar audience. A low bar. Let us set a higher one: what does it look like to actually hold the role, and what does your supplier need to build to make it holdable at all.

The question a buyer should be asking their team this quarter

Cerretani’s framing is uncomfortable in the right way. In session one, the line I would hand to any executive who still reads AI discoverability as traffic optimisation: “Your content can be live, it can be on brand, it can be perfectly written, it can be beautiful. But it can be functionally invisible to the systems that your buyers are now engaging with first.” The qualifier that matters is first. For a growing share of considered purchases, the model is the first impression. The website is the second.

The size of the shift is now reported outside the webinar circuit. G2’s 2026 buyer research, a March 2026 survey of 1,076 B2B software buyers and decision makers, found 51% now start research with an AI chatbot more often than with Google, up from 29% in the 2025 report. Sixty-nine percent chose a different vendor than they had planned on chatbot guidance. Eight in ten said chatbots accelerated the decision. Source: G2’s own report, March 2026, not the webinar’s rounded restatement of it.

Set this next to the ownership poll and the question answers itself. The surface that is now the first impression for a majority of your buyers has no accountable owner inside your organisation. The catch-22 for the leadership team: we have moved the top of the funnel to a channel we do not staff.

The contradiction between the two polls is the post

The WordPress VIP hub page frames the discipline as Visible, Structured, Optimized: understand how AI currently represents the brand, strengthen the content and technical foundations, then measure and refine. A sensible loop. The problem is that it is a loop, and loops need an operator.

In session one, broadcast 19 August 2026, Cerretani and Paxton Gray, CEO of 97th Floor, polled their audience on top concerns. Visibility, content and brand representation ranked highest. Ownership across teams ranked lowest. Paxton’s read, at twenty-two minutes in, was generous: “maybe everyone feels they own it, which would be a great problem to have.”

In session two, broadcast 22 September 2026, Cerretani ran a second poll. Twenty-five minutes in, she stopped and read the result out loud: “it was interesting that only 5% said that they had a clear owner in place.” That is the one number from either session that an executive should carry out of the room. The thing almost nobody has, and almost nobody is worried about not having.

They don’t have a single point of accountability, a directly responsible individual, to borrow an Amazon term.

Jodi Cerretani, CMO, WordPress VIP, session two of the enterprise AI discoverability series, 22 September 2026

Cerretani borrowing Amazon’s DRI label does more than name-drop a framework. The DRI pattern scales inside large organisations because it collapses a diffuse “the team” into a person. When someone asks who owns AI discoverability, “marketing and content and SEO and the web team together” is a wrong answer dressed as collaboration. It is the answer that produces the 95%.

Seven functions can hold the role, and that is actually the point

Cerretani’s second answer, around twenty-two and a half minutes into session two, is more useful than her first, because it names the people who could reasonably hold the role rather than inventing a new one. Her list: marketing owns the narrative, content owns the questions buyers actually ask, SEO maps the query fan-out, the web team owns templates and information architecture, engineering owns performance, security owns crawler policy, governance owns the standards. Any of them, she said, can carry accountability. What matters is that one of them does.

If you are evaluating where to put the role inside your own organisation, the useful test is which function will not pass the brief back to another at the first scope dispute. Marketing is the natural home for the narrative question but defers to engineering on crawler access. Security owns crawler policy decisively but not content pruning. The web team owns templates but is rarely resourced to run a quarterly re-audit. Governance has the right remit on paper and the slowest cadence in practice.

My read, after watching this role fail to settle in a dozen enterprise estates: the honest answer is marketing with a formal secondment from engineering. Primary KPI is brand sentiment inside the models; secondary is crawler coverage and schema integrity. The second KPI exists because the first is lagging and nobody can wait a quarter to find out they are regressing. That is a structural decision, not a job description, and it is the kind of decision an agency should be helping you arrive at before anyone writes the first content prune.

The lever nobody has cracked yet, and why that makes ownership heavier not lighter

The honest admission in the series is Paxton’s, twenty-four minutes into session one: “it’s one thing to see are we showing up, and then it’s another thing to see is showing up having an impact on the bottom line. And that is a big gap that I don’t think there’s really great answers for right now. There’s a lot of correlation and not causation.” Later, around twenty-nine minutes in, he reached for a billboard analogy worth taking seriously. You cannot measure a billboard against a direct conversion. You can measure brand search, direct traffic, and conversion rate on visitors who already know the name, and watch those move after you change what the billboard says.

Cerretani undercut her own product on the same point, sixteen minutes into session one: “current measurement tools, mine included, here at WordPress VIP, they really measure humans because they run on JavaScript in the browser. And AI crawlers don’t execute JavaScript.” A candid line from a vendor. The implication is that the KPIs the owner of this role will be judged on are, today, approximations of approximations. The absence of a clean metric is a reason the role needs a settled owner, not a reason to defer appointing one.

Session three of the series airs on 20 October 2026 and is dedicated to measurement. I will come back to this piece afterwards. The point I would make to a board this quarter: measurement will catch up on an industry timeline, and the ownership question needs to be settled before it does, because the person running the role in the quarter the metrics clarify will be the person judged on the baseline.

What a real WordPress or WooCommerce build looks like when the role is held, not just assigned

The gap between “we named someone” and “someone is actually doing the work” is where the platform earns its keep. If AI discoverability sits with marketing but every quarterly audit requires an engineering ticket to run a crawler log query, the role will quietly revert to engineering until the next reorganisation. What follows is the shape of the build we would put behind the role on an enterprise WordPress estate, mapped against the four technical layers that have to serve it.

Layer one, content modelling. Model the entities once, as fields on structured post types, not as free-text blobs inside a block editor. For a product catalogue, that is name, canonical description, specifications, availability, pricing, reviews, support policy, each a field, each surfaced as JSON-LD by the template. Model it twice, as free text plus a schema block the editor can forget to fill in, and you will lose a quarter arguing about which page said what to the model. On a 2,000 to 10,000 SKU WooCommerce catalogue, the first build is four to six weeks including taxonomy repair.

Layer two, template-level schema and sitemap integrity. Every page template emits the right JSON-LD, one H1, a clean heading hierarchy and a canonical URL the sitemap actually lists. Audit cadence quarterly, which means a job on the build and not on a human. The regulatory-compliance plugin Cerretani described in session one, quietly rejecting around 15% of ChatGPT requests because it treated AI crawlers as human visitors from blocked countries, is the kind of defect that lives inside this layer and is invisible until someone pulls the server logs. A log-driven crawler visibility check that runs quarterly and surfaces anomalies to the owner’s dashboard is three days of work, and pays for itself the first time a mu-plugin or a WAF rule changes.

Layer three, editorial governance that lives in the tool. Checks in the publish flow, not in a review document. Mandatory schema fields on the structured post types. A naming guard that catches the “WordPress VIP” versus “enterprise WordPress” collision Cerretani named in session two, where two names for the same thing are read by AI as two things. Pre-publish validation that fails loudly, not a style guide the content team is expected to remember. On WordPress: capability checks, pre-save filters, custom validation on post save, enforced through must-use plugins so a theme switch cannot deactivate them. One to two sprints for a mid-market estate, and the highest-leverage item on the list.

Layer four, the operational cadence. Cerretani said in session one that WordPress VIP re-runs the baseline, fix and retest cycle about every 90 days. Close enough to feel the business pressure, far enough apart that fixes have time to land. The owner needs a quarterly surface: last baseline, delta since last quarter, open fixes, their blockers, who is working on them. On a real WooCommerce estate this is a dashboard, not a deck, and it has to be fed by data the stack already collects: server logs for crawler coverage, schema validation from a nightly job, sitemap state as a diff against the previous quarter, and sentiment pulled from a scripted prompt run against the three or four models that drive the measurable share of referrals today. Build time, four to six weeks on top of the layers above.

Total realistic effort on a mid-market multi-brand estate, WordPress with WooCommerce behind the catalogue: eight to twelve weeks for the first full pass, two to three days per quarter thereafter. The quarterly cost is low on purpose. If holding the role requires a full-time hire to re-run audits, the role will not survive the first budget cycle. If it is a dashboard the owner can read in forty minutes, it will.

Cerretani’s platform line, at six minutes fifty into session two, is the one I would read back to any procurement team weighing WordPress VIP against a managed host or a self-managed enterprise stack: “do you have a website software that can allow you to do this automatically, or are you dependent on your human resources to make sure to do all these things? If the answer is the latter and not the former, you might be in trouble.” Agree with the thrust, read it carefully: the platform does not hold the role for you. It makes the role holdable. That is the WordPress VIP argument stated plainly, and the agency’s argument. Platform is the floor. The person running the role is the ceiling.

If you are closer to the 95% than the 5% today, the first move is not an audit, it is a conversation at leadership level about where the role sits. We run that conversation and come out with a named owner, a measured baseline across your estate, and a shortlist of the three fixes worth putting in the next quarter. On an estate we would plausibly be asked about, that is a two-week engagement, and it is the engagement that makes the rest of the work land.