All issues
Discovery Digest · 9 June 2026
Issue 03. The week AI search gave users the off switch
Three stories this week point in the same direction: users are getting more control over AI in search, the company behind the biggest AI search product is preparing to go public, and Google's underlying model infrastructure keeps moving fast. Here's what your team needs to know.
01 · AI Search / GEO
Bing lets users disable AI Copilot responses in search results
- What
- Microsoft has launched a preview feature that lets Bing users turn off AI Copilot responses in search results entirely. The mechanism works two ways: a browser extension for Chrome and Edge with a simple on/off toggle, or appending '-ai' to any search query to strip AI answers from that result. Microsoft's President of Search, Jordi Ribas, confirmed the feature on X, framing it as a direct response to user demand for choice. The move follows a measurable uptick in DuckDuckGo usage, which Microsoft has attributed in part to users wanting results without AI-generated content. This is a preview, and Microsoft is gathering feedback before deciding on broader integration.
- When
- Announced 9 June 2026. Available now as a preview via browser extension for Chrome and Edge. No confirmed date for full rollout.
- How it shifts discovery
- This matters for GEO strategy in a way that is easy to underestimate. If a meaningful share of Bing users opt out of Copilot responses, the reach of AI-cited content on Bing shrinks proportionally. Your team may already be tracking citation frequency in AI answers as a GEO metric. Now you need a second dimension: what percentage of your target audience is likely to use the opt-out? Bing's market share is smaller than Google's, but it skews towards enterprise and older demographics, and those audiences may be more inclined to turn AI off. On the paid side, Microsoft has noted that disabling AI responses could affect ad visibility alongside AI-generated content, so PPC teams should watch impression and CTR data for any shifts. What I'd do: flag this to whoever owns your Bing GEO and paid search reporting, and set up a segment to monitor Bing organic and paid performance over the next four to six weeks to catch any signal early.
- Questions to ask
- Are we tracking AI-citation impressions on Bing separately from organic click data, and do we have a baseline before the opt-out feature scales?
- Does our target audience on Bing skew toward users who are likely to disable AI features, and how should that change our content and bid strategy there?
- If Bing rolls this out broadly and Google follows with a similar option, how does our GEO playbook need to change to retain visibility in traditional results?
- Sources
02 · Industry / AI Infrastructure
OpenAI files a confidential S-1 with the SEC
- What
- OpenAI has submitted a confidential draft S-1 registration statement to the US Securities and Exchange Commission, the first formal step toward an initial public offering. The company published the announcement itself on 8 June 2026, noting it expected the filing to leak and chose to get ahead of it. OpenAI was explicit that timing is not fixed and may be some time off, citing activities that are easier to complete as a private company. The statement is classified as a notice under Rule 135 of the Securities Act and does not constitute an offer of securities.
- When
- Announced publicly by OpenAI on 8 June 2026. IPO timing described by OpenAI as undecided and potentially 'a while' away.
- How it shifts discovery
- This is a structural story, not a product update, but it has real downstream implications for anyone building strategy around ChatGPT Search or AI-answer citation. A public OpenAI means quarterly earnings scrutiny, revenue targets, and investor pressure on product decisions. That changes the incentive calculus around how aggressively ChatGPT Search surfaces citations, whether it introduces ads, and how it balances user experience against monetisation. The most likely near-term effect is that OpenAI accelerates commercial features to demonstrate revenue growth ahead of listing. For teams running GEO programmes targeting ChatGPT citations, that could mean new ad surfaces appearing around cited content, or shifts in which content types get favoured. What I'd do: note this as a long-horizon signal rather than a this-week action item, but start asking your GEO strategy questions now: if ChatGPT Search introduces paid placement or shifts citation logic under commercial pressure, is your content positioned to hold organic citation, or are you relying on a model that assumes the current ranking signals stay stable?
- Questions to ask
- If ChatGPT Search introduces advertising or changes citation logic under IPO-related commercial pressure, how exposed is our current GEO strategy?
- Are we tracking which content assets earn ChatGPT citations today, so we have a baseline if the model's behaviour shifts post-IPO?
- Should we be diversifying our AI-answer visibility across Perplexity, Gemini, and Copilot rather than concentrating on ChatGPT?
- Sources
03 · Google Search / AI Overviews / GEO
Google I/O 2026 recap: AI Mode, Gemini in Search, and what it means for your team
- What
- Google I/O 2026 took place in late May 2026 and served as the primary venue for Google's search and AI product direction for the year. The Developer Keynote and main I/O Keynote are publicly available via the Google for Developers YouTube playlist. Announced updates include developments to AI Mode in Search, AI Overviews, Gemini API capabilities, and Google AI Studio. The playlist has accumulated over 9 million views on the main keynote alone, reflecting the scale of developer and practitioner attention.
- When
- Google I/O 2026 was streamed live approximately two weeks before this issue, in late May 2026. Recap videos published to the Google for Developers YouTube channel from 26 May 2026 onward.
- How it shifts discovery
- If you have not watched the keynote or at least the five-minute developer recap, do that before the week is out. I/O sets the product roadmap that your SEO and GEO decisions will be running against for the next twelve months. The updates to AI Mode are particularly relevant: AI Mode is Google's most direct structural challenge to the ten-blue-links format, and changes announced there affect how content gets surfaced, cited, or bypassed entirely. The Gemini API updates matter for teams building internal tools on top of Google's models, and for understanding the infrastructure that AI Overviews runs on. What I'd do: assign someone on your team to watch the Developer Keynote and produce a one-page summary of every search-adjacent announcement, with a column for 'what this changes about our current approach.' If you already did this when I/O aired, now is the time to turn those notes into Q3 priorities.
- Questions to ask
- Which specific I/O announcements change how AI Overviews or AI Mode selects and cites sources, and have we updated our GEO content brief template to reflect them?
- Are there new Gemini API capabilities we should be testing internally for content production, search analysis, or SERP monitoring?
- What did Google say, explicitly or implicitly, about the future of traditional organic results alongside AI Mode?
- Sources
04 · Google AI / International Search
Gemini 3.5 Live Translate launches with near real-time speech translation across 70+ languages
- What
- Google has released Gemini 3.5 Live Translate, a new audio model that provides near real-time speech-to-speech translation across more than 70 languages. Unlike sequential turn-by-turn translation systems that wait for a speaker to finish, this model processes speech continuously as it streams, staying a few seconds behind the speaker throughout a session. It automatically detects languages without manual configuration and preserves intonation, pacing, and pitch. The model is rolling out across Google AI Studio and the Gemini Live API for developers in public preview, Google Meet for enterprises in private preview, and Google Translate on Android and iOS for general users.
- When
- Announced and rollout began 9 June 2026. Developer public preview via Gemini Live API and Google AI Studio starts immediately. Enterprise private preview in Google Meet begins June 2026. Consumer rollout via Google Translate on Android and iOS starts the same day.
- How it shifts discovery
- The direct search-relevance here is indirect but worth flagging for teams with international audiences. Gemini 3.5 is the model generation feeding Google's broader AI infrastructure, including AI Overviews. Capability jumps at the model layer tend to ripple into search product quality over the following quarters, particularly for multilingual query understanding and cross-language content surfacing. For teams running multilingual SEO, this is a signal that Google's ability to understand and serve content across languages is getting meaningfully better, which raises the bar for localised content quality. Thin translated pages that relied on the gap between language quality in search ranking and human content are more exposed than ever. What I'd do: if you have a multilingual content programme, audit your lowest-quality translated pages now. A model that handles multilingual inputs with greater fluency and context will be less forgiving of machine-translated content that lacks depth.
- Questions to ask
- Do our multilingual pages meet the same depth and accuracy standard as our primary-language content, or have we been relying on lower quality thresholds for secondary markets?
- How might improved multilingual model capability affect which language-version of a page Google chooses to surface in AI Overviews for a given query?
- Are there new developer use cases in the Gemini Live API that our engineering team should be evaluating for internal search or customer-facing tools?
- Sources
05 · Paid Search / SERP Format
Google tests dotted underline styling and 'Top pages' label on Search ad sitelinks
- What
- Google is running a visual experiment on mobile Search ads in which sitelink assets are displayed with a dotted underline rather than the standard solid underline. A second variation of the same test adds a 'Top pages' label above the sitelink row. Both variants were spotted by search practitioners on mobile and reported on 9 June 2026. No official confirmation from Google has been issued. This is a UI test only and does not change sitelink eligibility or bidding mechanics.
- When
- Spotted and reported 9 June 2026. Status is an unconfirmed live experiment. No rollout timeline published.
- How it shifts discovery
- SERP format tests like this rarely stay as small visual tweaks. The 'Top pages' label in particular is worth watching: labelling sitelinks as 'Top pages' frames them as editorially significant rather than just additional navigation options, which could increase click-through on the sitelink row. If this rolls out, it may shift click distribution within an ad from the headline to the sitelinks, which has implications for landing page strategy and conversion tracking. Make sure each sitelink points somewhere genuinely useful with its own conversion path, not just a category page included to fill the asset slot. What I'd do: check that your active sitelink assets are pointing to high-quality, distinct destination pages. If this test rolls out and sitelinks get more prominence, weak sitelink destinations will waste the visibility.
- Questions to ask
- Are our current sitelink assets pointing to distinct, high-value pages, or are some just padding to meet the minimum asset count?
- Do we have conversion tracking set up at the sitelink destination level so we can measure if click distribution shifts when this format rolls out?
- How does increased sitelink prominence affect the visual competition between our paid ads and adjacent organic results on mobile?
- Sources
Key takeaways
What to walk away with this week
Bing's AI opt-out feature means GEO teams can no longer treat AI-answer impressions on Bing as a stable metric. Establish a baseline now and monitor for shifts over the next six weeks.
OpenAI's confidential S-1 filing signals that ChatGPT Search will face growing commercial pressure post-IPO. Teams relying on organic ChatGPT citations should be stress-testing that strategy against likely monetisation changes.
Google I/O 2026 set the search product roadmap for the next twelve months. If your team has not reviewed the keynote announcements and translated them into updated GEO and SEO priorities, that is the most urgent action item this week.
Gemini 3.5's multilingual capability improvements are a long-term signal that thin or machine-translated content is increasingly exposed. Multilingual teams should audit their lowest-quality pages now.
Google's sitelink format test, particularly the 'Top pages' label variant, could shift click distribution within paid ads toward sitelinks. Audit sitelink destination quality before this potentially rolls out broadly.