"Who's running point on AI visibility for 2026?" Ask that in a planning room and watch it empty out. The meeting gets scheduled. The invite goes out. Three people decline, two accept and show up without an opinion, and the quarter closes with the line item still sitting in a parking lot.

Being named and cited inside ChatGPT, Perplexity, Gemini, and Google's AI surfaces is now a channel in its own right. The budgets are moving to prove it: a Gartner survey of 401 marketing leaders found CMOs are allocating 15.3% of marketing budgets to AI, while total marketing spend sits flat at 7.8% of company revenue. The money is there. What's missing is a name next to the work.

The SEO Lead Says It Already Lives in Their Backlog

The SEO team's version of the story is reasonable on paper. They already own structured data, entity work, internal linking, and the authority stack that AI engines lean on. If a model is pulling a sentence from the site into an answer, the thinking goes, that's a crawl-and-citation problem, same muscle on a new surface.

The trouble is scope. SEO's measurement model ends at the SERP. AI visibility starts after it: the prompt, the model's chosen sources, the paraphrased answer the user actually sees, and whether the brand shows up by name inside it. Hand the work to SEO and it lands on a team whose dashboards don't show the outcome.

TechBullion's coverage of the 2026 planning cycle makes a longer case for why AI visibility needs its own budget line, and it's worth reading before the next budget meeting.

The PR Lead Treats It as a Byproduct of Earned Coverage

PR's version is almost the opposite. Get the brand quoted in the trades, mentioned in analyst notes, and referenced in the publications the models trust, and citations will follow. There's truth in that, since editorial sources carry weight, but it leaves two things undone.

First, PR is pitching stories, not engineering what a model retrieves. Second, nobody on a PR team is checking, week over week, which prompts return the brand and which return a competitor.

The coverage happens. The machine's reading of that coverage is somebody else's job, and nobody has that job.

The Product Team Says It's a Site and Schema Problem

Product and web engineering have their own claim: the model reads pages, pages are shipped by product, therefore the work belongs in the product backlog. Schema markup, glossary pages, FAQ blocks, canonical entity definitions, all of it ships through them.

What product won't do is write the content strategy, pitch the external sources, or own a KPI that reads "share of answers." They'll ship what's specified. Nobody is writing the spec.

The CMO Assumed One of the Three Already Had It

The failure mode is slow and almost invisible. Each function has a plausible claim and a partial contribution, so the CMO signs off on three overlapping workstreams and reads the overlap as coverage. What it really produces is duplicated tool spend, contradictory measurement, and no single person who can answer whether the brand's share of citations moved this quarter.

A governance guide from GrowByData frames the fix well: name a single accountable owner for the outcome, give a program lead the measurement, and let each channel own its own fixes against a shared scorecard. Turf is beside the point; one name has to appear next to the number.

Decide This Before the Quarter Closes

The planning calendar does not forgive indecision here. A few concrete calls separate a funded program from another declined invite in January:

• Name the owner, not the committee. Pick one leader accountable for share-of-answer across the engines that matter to your buyers. A committee is a decline in disguise.
• Fund it as its own line. Ring-fence the budget rather than carving slivers from SEO, PR, and content. Reallocating inside existing lines leaves you with the same ambiguity you started with.
• Define the win before the spend. Agree on the prompt set, the engines in scope, and what "cited" means in your category. Without this, every quarterly review turns into a debate about the definition.
• Separate the work from the tool. A monitoring subscription is not a program. Decide what gets fixed when the dashboard turns red, and which team ships the fix.

The invite keeps getting declined because each team is right about the piece they already own and wrong about the whole. Someone has to accept that the whole is now a job. Do it before Q4 planning closes, or do it in Q2 after a competitor has already been named in the answer your buyer typed.