Digiday reported on 10 August 2026 that Zoom is buying creator partnerships explicitly to move its visibility inside AI search results, and that Zoom's own CMO says the link between the two cannot yet be measured. Kimberly Storin, Zoom's chief marketing officer, told Digiday it is "not a direct correlation at this point" while maintaining that influencers matter to a brand's presence and credibility inside AI-generated answers.
That admission is the news. An enterprise brand is funding a creator programme against a hypothesis it cannot yet prove, on the record, with a named executive attached. Most spend justified by AI search visibility so far has been described in vaguer terms by people who were not willing to say the measurement is missing.
What Digiday reported on 10 August 2026
Digiday, in a piece bylined Kimeko McCoy published on 10 August 2026, reported that Zoom's summer campaign centres on Nayeema Raza, a journalist, filmmaker and host of the "Smart Girl Dumb Questions" podcast, who has more than 30,000 followers across Instagram and YouTube. Zoom is sponsoring a six-part podcast series with Raza and has booked her to keynote its annual conference in October. Financial terms were not disclosed.
Alongside the creator work, Zoom has partnered with Profound, a generative-AI marketing intelligence platform, to analyse which sources the models pull from and to adjust its marketing and communications accordingly. Zoom is also participating in the ChatGPT advertising pilot. Digiday reported all three strands as parts of the same effort to be present where AI answers get assembled.
The article cites Semrush's AI Visibility Index for the underlying asymmetry between engines: ChatGPT draws on an average of 15.4 preferred sources, against 3.3 for Google Gemini, with ChatGPT skewing toward Wikipedia and Reddit while Gemini skews toward YouTube, Wikipedia and e-commerce sites. Digiday also noted Priceline shifting social spend toward influencers and creators, announced in June.
Storin's on-record admission is the centre of the story
Kimberly Storin, Zoom's CMO, told Digiday on 10 August 2026 that the relationship between creator partnerships and AI search visibility is "not a direct correlation at this point". She said in the same breath that influencers matter to presence and credibility inside AI search results.
Those two statements sit in tension and should be left in tension rather than resolved. A CMO is spending against a mechanism she believes in and cannot yet demonstrate. That is a normal position in an emerging channel and an unusual one to state publicly, because the pressure in marketing communication runs the other way, toward implying attribution that does not exist.
The honest framing for anyone reading this as a template is that Zoom is running an experiment with real money and no readout. Nothing in the Digiday piece establishes that the creator work has moved a single citation. No before-and-after data was published, no citation-lift figure was given, and no timeline for when Zoom expects to be able to measure it was reported. Any mechanism described below, including the one in the next section, is the hypothesis being tested, not a finding.
The Semrush numbers: 15.4 preferred sources against 3.3
Digiday cited Semrush's AI Visibility Index in its 10 August 2026 report for the source behaviour of the two largest assistants. The figures describe how many preferred sources each engine typically draws on and which kinds of sources each leans toward.
| Semrush AI Visibility Index, as cited by Digiday | ChatGPT | Google Gemini |
|---|---|---|
| Average number of preferred sources drawn on | 15.4 | 3.3 |
| Source types the engine skews toward | Wikipedia and Reddit | YouTube, Wikipedia and e-commerce sites |
The gap between 15.4 and 3.3 is the practical core of the story, and it is a structural difference rather than a quality judgement about either engine. Digiday did not publish the methodology behind the index, so the figures should be read as a directional description of source breadth rather than as a precise constant.
What a narrow source set does to a visibility budget
The mechanism this implies, and it is inference rather than something the source stated, runs on scarcity. If Gemini typically draws on about 3.3 preferred sources for an answer, then each of those slots carries enormous weight, and the source types Gemini favours are YouTube, Wikipedia and e-commerce sites. Presence in those specific places is worth disproportionately more in that engine than presence anywhere else, because there is almost no room in the answer for anything else.
ChatGPT's average of 15.4 preferred sources works the opposite way. A wider draw means no single source dominates, and a brand mentioned across many independent third-party places has more chances to be one of the fifteen. That favours distributed earned coverage over any one owned asset, and the source types it skews toward, Wikipedia and Reddit, are places a brand cannot simply buy its way into.
Split a generative engine optimisation budget along those lines and the two engines pull in different directions. Owned content and structured, factual reference material serve the narrow-source engine; volume and spread of independent third-party mentions serve the wide-source engine. A creator partnership is one of the few tactics that plausibly touches both, since it produces third-party commentary and, on YouTube, sits inside a source type Gemini reportedly favours. That is the argument for the buy. It is not evidence that the buy works, and Digiday published none.
A 30,000-follower host is a different kind of buy
Nayeema Raza has more than 30,000 followers across Instagram and YouTube, which Digiday reported on 10 August 2026 as part of describing the campaign. By the standards of a reach-led influencer programme at an enterprise software company, that is a small audience.
What the selection appears to be optimising for is something other than impressions. The reasoning this suggests, and Zoom did not spell it out in the coverage, is that a six-part podcast series and a conference keynote generate a body of substantive, attributable commentary about a topic, in formats that produce transcripts, descriptions, write-ups and quotations. Those are text artefacts that end up on the open web with a named human attached to them. Follower count does not describe that output at all.
If selection criteria really are shifting this way, the screening question changes from how many people will see this to what durable, citable material will exist afterwards and where it will sit. Podcasts, long-form video and conference talks score differently on that question than a high-reach single post does. This remains reasoning about an unmeasured hypothesis, and the source did not report any selection framework that Zoom used.
The measurement problem, stated plainly
Storin said the correlation is not direct at this point, and no vendor named in the Digiday piece was reported as closing that gap. Zoom's partnership with Profound is described as analysing which sources the models pull from so that Zoom can adjust its marketing and communications, which is a visibility and source-tracking function.
Separating what a marketer can observe today from what remains unmeasurable is the useful discipline here. Observable, in principle: whether a brand appears in AI answers for a given set of prompts, which sources an engine tends to cite for those prompts, whether a specific piece of coverage exists and is indexable, and referral traffic that arrives with an assistant as the source. Not observable, on the evidence in this article: whether a given creator partnership caused a change in citation frequency, how long any such effect takes to appear, and how any of it converts.
The gap between those two lists is where every claim about AI search attribution currently lives. Anyone presenting a causal chain from creator spend to citation to revenue is presenting a model, not a measurement, and Zoom's own CMO declined to present one.
Zoom is not the only advertiser moving spend this way
Digiday reported on 10 August 2026 that Priceline announced in June it was shifting social spend toward influencers and creators. The piece placed that alongside Zoom's campaign as evidence of a broader move rather than a single company's bet.
Two named advertisers do not establish a trend, and Digiday did not report a reason for the Priceline shift, so it should not be assumed to share Zoom's AI-visibility rationale. What the pairing does establish is that budget is moving toward creators at companies large enough for the shift to be announced, which is the observable part.
What the source did not say
The absences in this story are as important as the content, because the story is about a hypothesis rather than a result.
- No financial terms. The value of the six-part podcast sponsorship and the October keynote booking was not disclosed.
- No citation-lift figures. Nothing was published on whether Zoom's presence in AI answers has changed at all.
- No before-and-after data. There is no baseline and no measurement period reported.
- No timeline. Zoom did not say when it expects the correlation to become measurable, or what would count as proof.
- No methodology behind the Semrush AI Visibility Index. The 15.4 and 3.3 figures were cited without how they were derived, what prompt set was used, or over what period.
- Nothing about Thailand or Southeast Asia. The article contains no regional data at all.
Everything in this article beyond those reported facts is analysis, and it is marked as analysis where it appears.
What this means for Thai marketers
The source has no Thai or Southeast Asian data. Nothing in it describes how assistants behave with Thai-language prompts, which sources they favour in Thai, or whether Thai creator coverage carries the same weight as English-language coverage. Assuming the Semrush source-breadth figures transfer to Thai queries would be an assumption, not a finding, and the honest position is that nobody has published the Thai equivalent.
What does transfer is the structure of the decision. A Thai brand considering creator spend for AI visibility reasons is making the same unmeasured bet Zoom is making, with less published evidence behind it, and should size it accordingly. Treating it as a test with a defined budget and a defined observation window is defensible. Treating it as a proven channel is not.
The one place the Thai market may differ in the brand's favour is source scarcity. If an engine draws on a small number of preferred sources and comparatively little authoritative Thai-language material exists on a given topic, the bar to becoming one of those sources is lower than it is in English. That is reasoning from the reported source-breadth asymmetry, not something the source measured, and it argues for building substantive Thai-language reference material as the first move in a generative engine optimisation programme rather than as an afterthought.
What to check before you copy this play
The useful version of this story is a set of questions, not a tactic to adopt.
- Write down what you would accept as proof that creator spend moved your AI search visibility. If you cannot define it, you are in the same position as Zoom, which is fine as long as it is stated in the plan.
- Establish a baseline before spending. Record which prompts your brand appears in today and which sources the assistant cites for them, because without a before, there is no after.
- Separate the two engine behaviours in your thinking. Reported source breadth differs by an order of magnitude between the assistants, so a single visibility target for all of them hides the actual work.
- Check whether your creator output leaves durable text behind. A post that disappears from a feed and a transcript that stays on the open web are different assets for this purpose.
- Cap the experiment. An unmeasured hypothesis deserves a test budget and a review date, not an open line.
- Keep the claim honest internally. Reporting creator spend as an AI visibility investment is defensible; reporting it as a proven citation driver is not, and the CMO in this story did not.
That review is mostly documentation, and it tends to expose how much of a content marketing plan is already aimed at answer engines without anyone having written down how it would be judged.
Frequently asked questions
Does creator content actually improve AI search visibility?
The source does not establish that it does. Zoom's CMO said explicitly that it is not a direct correlation at this point, and Digiday published no citation-lift data, no before-and-after figures and no timeline. Zoom believes influencers matter to presence and credibility inside AI answers and is spending on that basis, which makes it a hypothesis under test rather than a demonstrated mechanism.
Why do ChatGPT and Gemini need different approaches?
Because the number of sources they draw on differs sharply, according to the Semrush AI Visibility Index as cited by Digiday: an average of 15.4 preferred sources for ChatGPT against 3.3 for Gemini. A narrow draw concentrates weight in a few places, which Semrush reported as YouTube, Wikipedia and e-commerce sites for Gemini, while a wider draw rewards being mentioned in many independent places. How to act on that is inference; the figures themselves are what was reported.
Is any of this relevant to Thailand yet?
The source says nothing about Thailand or Southeast Asia. There is no regional breakdown, no Thai-language prompt data and nothing about how assistants source Thai content. The structural argument about source scarcity can be reasoned about locally, but no figure in this story should be quoted as applying to the Thai market.
Do I need a tool like the one Zoom is using?
Not necessarily, and the source does not recommend one. Digiday reported that Zoom partnered with Profound, a generative-AI marketing intelligence platform, to analyse which sources the models pull from. The underlying task, checking which prompts surface your brand and which sources get cited, can be started manually at small scale before any platform decision is made.
How small an audience is too small for this kind of partnership?
The source offers one data point rather than a rule: Zoom built a campaign around a host with more than 30,000 followers across Instagram and YouTube. That is small for a reach buy at an enterprise brand, which suggests the selection was made on something other than audience size. Digiday did not report Zoom's selection criteria, so the reasoning about why remains inference.
Where this leaves generative engine optimisation
What makes this story worth reading is not the tactic. It is that a named CMO at a large software company said on the record, in Digiday on 10 August 2026, that she is spending against an effect she cannot yet measure. That is a more accurate description of the state of AI search visibility work than most of what is being sold around it.
The correct response is not to dismiss the spend or to copy it. It is to build the baseline that Zoom's admission implies is missing, keep the experiment sized like an experiment, and be precise internally about the difference between presence you can observe and causation you cannot. The full Digiday report is available here.
If you are being asked to justify budget on AI search visibility grounds, Relevant Audience can help you set up the measurement baseline first, so a GEO programme is judged on what it can actually show.
