
Google is absorbing more informational queries into AI answers, while the systems beneath SEO, paid media, and analytics are becoming more dependent on trusted inputs. For B2B teams, the practical shift is clear; visibility and efficiency increasingly depend less on producing more and more on whether these systems can identify, trust, and use the signals behind your brand, conversions, and business data. This edition looks at where that shift is already changing decisions.
Week in Brief
- Google is continuing to move informational search toward AI-generated answers. People Also Ask answers hit near 100% AI Overviews in September.
- Google updated its Preferred Sources guidance for AI Mode and AI Overviews on 10 September, the user-controlled signal that badges reader-chosen sites inside AI answers.
- Google launched first-party data and measurement upgrades on 10 September, including a Data Strength Uplift Metric that quantifies what an advertiser’s data setup is worth.
- OpenAI launched a Data Agent in ChatGPT Work on 10 September that connects to company data and builds dashboards from a plain-language question.
- The shift in Search is clear; AI systems increasingly reward signals they can trust and act on, not raw content or data volume.
AI Overviews Are Reshaping How Informational Questions Are Answered in Search
Google is continuing to move informational search toward AI-generated answers. In its official guidance, Google says AI Overviews are designed to help users quickly understand complex topics or questions, while AI Mode allows people to explore those questions conversationally and in greater depth. Google has also confirmed that AI Overviews now reach more than 2.5 billion monthly active users, while AI Mode has surpassed one billion monthly users globally.
Google’s direction is therefore clear even without relying on third-party estimates about individual SERP features; more informational discovery is happening inside AI-generated Search experiences, and website visibility increasingly depends on whether content is selected as a supporting source within those answers.
What this advancement in AI overview means for B2B, ITES, and SaaS
Companies that rely heavily on broad informational queries such as “what is GEO”, “how does LinkedIn Ads attribution work”, or “what is entity SEO” should no longer judge those pages only by organic clicks. A user may now get enough information directly from an AI Overview to continue researching without immediately visiting the source.
That does not make informational content less important. Google says AI Overviews and AI Mode use content from its Search index and rely on its existing ranking and quality systems to surface relevant supporting links. Strong informational content therefore still helps establish what a company knows and gives Google material it can retrieve when generating AI-powered responses.
For B2B companies, the practical shift is from measuring only clicks to also measuring AI visibility. Mid-funnel, comparison, implementation, and vendor-evaluation content have also become increasingly important because these pages provide the specificity AI systems need when users move from learning about a topic to evaluating a solution.
Use Google Search Console’s dedicated Generative AI performance reporting to identify which pages are appearing in AI Overviews and AI Mode. Compare those pages with your traditional Search impressions and clicks. Pay particular attention to informational pages where conventional clicks are weakening, but AI visibility remains strong, which may indicate the page is still influencing discovery even when the traditional click path is changing.
Google is automatically migrating Campaign-level Broad Match and Automatically Created Assets to AI Max
Google is automatically upgrading Search campaigns that use the campaign-level broad match setting or Automatically Created Assets to AI Max, in a migration running from 1 to 30 September, per the Google Ads Developer Blog published 12 August and advertiser emails sent 5 August. Google switched off the creation of new legacy broad match and ACA configurations on 3 August, so the affected set is fixed. The migration happens in place, campaign by campaign, with equivalent AI Max settings and brand controls carried over. Dynamic Search Ads were split out of this window and delayed to February 2027. Search Engine Land notes there is no rollback once a campaign is migrated.
What this migration of ads to AI Max means for B2B, ITES, and SaaS
B2B advertisers running broad match Search campaigns for lead generation are being moved onto AI Max whether or not they opted in, with search term matching enabled by default. AI Max widens where your ads can match, which for a B2B account with a narrow ICP can mean spend leaking to loosely related queries the old setup filtered out. The migration also changes the reporting layer: new search-term-matching segments appear, and final URL expansion can break landing-page joins in any warehouse or BI pipeline keyed on the URL. Companies with tight negative keyword lists and clean conversion tracking will see the least disruption. Companies running broad match with thin negatives are most exposed to wasted spend.
Before your campaigns migrate, tighten negative keyword lists and confirm brand exclusions, then check whether any reporting pipeline joins on landing-page URL.
Google updated its Preferred Sources guidance
Google updated its Search Central documentation for Preferred Sources on 10 September. Preferred Sources is the user-controlled feature that lets a reader mark sites they trust; Google extended it into AI Overviews and AI Mode on 27 May, badging a chosen source inside the AI answer. Google added an embeddable one-click button on 20 August and has reported more than 600,000 selected sources since the May rollout, up from 345,000. Google says a reader is roughly twice as likely to click through to a source once they mark it preferred.
What this feature update means for B2B, ITES, and SaaS
This is one of the few AI-search levers that runs in a B2B company’s favor, and it depends on audience loyalty rather than an algorithm’s judgment. For a SaaS or ITES brand whose readers, customers, and newsletter subscribers already trust them, the Preferred Sources button converts that trust into a badged placement inside AI answers for those readers. It does not lift you in everyone’s results; rather, it is personalized to opted-in readers, so it rewards brands with a real audience, not brands chasing reach. Companies with an engaged email list or active community can act on this now. Companies with no direct audience relationship get little from it, which is itself a signal about where to invest.
If not previously, add Google’s embeddable Preferred Sources button to your highest-traffic pages and your newsletter, and ask your existing subscribers to select you.
Google launched first-party data and measurement upgrades
Google announced a set of first-party data and measurement upgrades on 10 September, per the Google Ads and Analytics blog. Data Manager is expanding beyond Google Ads into Google Analytics and Display & Video 360, and its API is adopting the IAB Tech Lab’s ECAPI standard for sending conversion data across platforms. A new Data Strength Uplift Metric in Google Ads estimates how many conversions an advertiser’s first-party data setup recovers. Google is also adding agentic features to Meridian, its open-source marketing mix model, and its Meridian GeoX experimentation library reached general availability. Google leads the announcement with two of its own figures: an 11% lift in Search conversions from connected data and a 26% incremental return on ad spend.
What this data means for B2B, ITES, and SaaS
As Google’s automated bidding leans harder on first-party signals, the quality of your conversion data is becoming the difference between AI bidding that works and AI bidding that burns budget. For B2B, that means feeding Google real pipeline signals including qualified leads, opportunities, and closed revenue from your CRM, rather than just raw form fills. The Data Strength Uplift Metric gives a marketing leader a number to take to finance when arguing for CRM-to-Google conversion integration. Companies already importing offline conversions will benefit fastest. Companies bidding on top-of-funnel form fills alone are the ones the metric will expose.
OpenAI launched a Data Agent in ChatGPT Work
OpenAI introduced a Data Agent in ChatGPT Work on 10 September. The agent connects to approved company data sources including Redshift, BigQuery, Snowflake, Databricks, and MongoDB, plus files from Google Drive and SharePoint, and investigates what changed in a business metric and builds interactive dashboards from a natural-language question. It enforces existing table, row, and column permissions, and OpenAI named NTT Data, Thermo Fisher, and ServiceTitan among early users. The framing is everyday business questions: why sales slowed, where spending is rising, which accounts are at renewal risk.
What this Data Agent in ChatGPT means for B2B, ITES, and SaaS
This lowers the technical bar for a marketing or RevOps team to interrogate its own data without waiting on an analyst. A B2B marketing leader could point it at pipeline data and ask which channels are driving qualified opportunities this quarter, then share the dashboard. The catch is the same as everywhere else this week; the answer is only as good as the data the agent can reach, and it inherits whatever gaps and permission limits already exist. Companies with clean, connected data will get useful answers. Companies with fragmented data will get confident-looking answers built on incomplete inputs, which is worse than no answer.
Apple said Siri now handles 2.5 billion requests a day
At Apple’s 9 September event, Lilian Rincon, Apple’s VP of Product Management for AI, said Siri handles 2.5 billion requests every day, reported by Search Engine Roundtable on 10 September. Apple positioned a new version of Siri as going further than before, with AppleInsider noting the AI Siri launch arrives in beta with daily usage caps.
Siri is a consumer voice-and-assistant surface, and there is no B2B query or citation mechanism to optimize for today. The number matters as an early opportunity and a marker of how much assistant-mediated search volume now sits outside Google.
On Our Radar for Next Week
Once a broad match or ACA campaign is on AI Max, there is no rollback. Watch your migrated B2B campaigns’ search terms report through late September for loosely matched queries, and confirm your landing-page reporting still joins correctly.
Questions We Were Asked This Week
People also ask feature almost stopped sending us clicks. Do we stop writing informational content?
No, but change what you measure. People Also Ask answers are now 97% AI Overviews per AlsoAsked’s 19.2 million query sample, so the click that box used to pass to your page is answered inside the result. Informational content still does the work of teaching Google and the LLMs what your brand knows. Keep publishing depth on the questions your buyers actually ask before a purchase, and start tracking whether your page is the source cited in the AI answer, not only whether it earns the click.
Google migrated our broad match campaigns to AI Max without asking. What do we do now?
Check the migrated campaigns before you judge the results. The migration runs through September, and there is no rollback once a campaign moves. AI Max turns on search term matching by default, which widens where your ads match, a real risk for a B2B account with a narrow ICP and thin negative keyword lists. Pull the search terms report on each migrated campaign and add negatives for the loosely related queries you do not want. Then confirm your conversion tracking and brand exclusions carried over, and check that any reporting pipeline keyed on landing-page URL still joins, because final URL expansion can quietly break it.
Everyone is talking about first-party data again. Is this just Google selling us something?
There is a sales motive, but the underlying shift is real. As Google’s automated bidding leans harder on first-party signals, the quality of your conversion data increasingly decides whether AI bidding finds pipeline or wastes budget. Google’s new Data Strength Uplift Metric puts a number on what your data setup recovers, which is useful mainly as a business case for connecting your CRM to Google Ads. If you already import offline conversions, you are ahead. If you bid on form fills alone, this is the quarter to fix it.
Should we be worried that AI tools like ChatGPT’s Data Agent can now analyze company data directly?
OpenAI’s Data Agent in ChatGPT Work connects to approved data sources and builds dashboards from a plain-language question, and it enforces your existing permissions. The value for a marketing or RevOps team is asking questions of pipeline data without waiting on an analyst. The risk is that its answers are only as good as the data it can reach. Test it on a question you already know the answer to before you rely on it for one you do not. The broader signal is that clean, connected data is becoming the difference between an AI tool that helps and one that misleads.
White Bunnie Observations
Three of this week’s launches look unrelated, but they point to the same shift. Preferred Sources is a trust signal for AI search. The Data Strength Uplift Metric is a trust signal for Google Ads. The ChatGPT Work Data Agent runs on the trust and completeness of a company’s own data. In each case, the advantage is moving away from producing more content or more data, and toward giving AI systems higher-quality signals they can act on.
For SEO and GEO, that means recognizable brands, audience preference, structured entities, and authoritative content. For paid media, it means better first-party conversion and customer signals. For analytics, it means connecting marketing data to actual business outcomes. The question a B2B company asked in early 2026 was “Are we visible on Google and ChatGPT?” The more useful question now is “are we giving AI systems enough trustworthy data to select our brand when they generate an answer, a recommendation, or an action?”

