What Google announced, and when
The feature has had two names and three launch dates in five months, which is worth setting out before anything else.
At Google I/O on 19 May 2026, Liz Reid, who leads Google Search, described “the era of Search agents” and said Google was “starting with information agents”. These would operate “in the background, 24/7”, look “across everything on the web, like blogs, news sites and social posts” plus Google's real-time data on finance, shopping and sports, “monitor for changes related to your specific question”, and send “an intelligent, synthesized update, with the ability to take action”. Her examples were an apartment search and a sneaker release. Information agents, she said, would launch first for paying Google AI Pro and Ultra subscribers that summer.
On 12 June Robby Stein posted on X that information agents were “now available in all AI Mode languages & markets for Google AI Ultra subscribers”, that an agent would “send detailed updates and links to the web the moment new info is available”, and that Google would “expand to more people this summer”.
On 28 September Stein posted again. This time the wording was “info monitoring capabilities in AI Mode in Search”, which Ultra and Pro subscribers “have been using” and which was now “rolling out to everyone globally” in the Google app. The post says: “Just tell AI Mode exactly what to look for & Search will continuously check across changing info on the web like sites, forums and social posts, our real-time data sources, and Shopping Graph of 60B+ products, so you can stay up-to-date with fresh info. And it'll even suggest tasks when you're searching.” It lists three uses: new restaurants or pop-ups nearby, local holiday activities with children, and back-in-stock and price-drop updates.
Two things changed between June and September besides the audience. The word “agent” dropped out, and so did the promise of links. Search Engine Journal, whose report is the headline that led us here, noticed both. We do not read much into the wording of a social post, but the June post is the only Google statement that updates carry links, and it described the paid version.
How it differs from Google Alerts
Google has offered a way to be told when something new appears about a topic for more than two decades: Google Alerts. Its help page describes it as a way to “get emails when new results for a topic show up in Google Search”, and gives “mentions of your name” as an example. You type a query and can set how often you are notified, the types of sites you see, the language, the region, how many results you want and which account receives them.
Set side by side on what Google itself has published, the two tools differ in five ways that matter to anyone watching a company's coverage.
Input. Alerts takes a search query. AI Mode monitoring takes a description in ordinary language (“exactly what to look for”), which can carry conditions a keyword cannot, such as only new coverage in trade press, or only reviews that mention delivery problems. Whether it honours conditions like those reliably, we do not know.
Output. Alerts sends matching results. AI Mode sends what Google calls a “synthesized update”: a written summary in the Google app. A list of results can be checked line by line. A summary has already made choices about what to include and how to describe it.
Sources. Alerts draws on Google Search results, which you can narrow by type of site. AI Mode monitoring, by Google's account, draws on sites, forums and social posts, Google's real-time data and its Shopping Graph. Google has not said how it weighs them, or which social platforms it can read.
Delivery. Alerts arrives by email. AI Mode monitoring is in the Google app. Google has not said whether updates can be sent anywhere else.
Documentation. Alerts has a help page with settings and troubleshooting steps. For AI Mode monitoring we could find a conference blog post and two posts on X. There is no published description of how often it checks, how many tasks one person can set, how long a task runs, or how to stop one.
What it is good for in monitoring, and what it is not
For a company or its agency, the obvious question is whether this replaces, or adds to, the tools already used to watch coverage. The honest answer is that Google has not built or described it for that job, and we have not been able to test it, so what follows is reasoning from Google's own descriptions.
What it could add. A plain-language brief, reaching across news, forums and social posts at once, is closer to what a communications team actually wants than a keyword alert. Coverage that does not use a company's exact name, a forum thread about a product problem, or a change on a competitor's page is the sort of thing keyword tools miss. If AI Mode monitoring can do that, it is a useful extra pair of eyes, and free.
What it cannot yet be. A monitoring record has to be complete enough to count, specific enough to act on and stable enough to report. Nothing Google has published describes exporting results, keeping a history, sharing a task with a colleague or counting mentions. A synthesised update is, by design, a selection. And AI summaries do get attributions wrong: when the Tow Center tested eight AI search tools, they named the wrong source more than 60 per cent of the time, as set out in how often AI search tools cite the wrong source. A team relying on summaries alone would not know what it had not been told.
Share of 200 queries per tool answered incorrectly, Tow Center test of 1,600 queries across eight tools, published March 2025
Source: Tow Center for Digital Journalism, Columbia Journalism Review, AI Search Has A Citation Problem, Klaudia Jaźwińska and Aisvarya Chandrasekar, 6 March 2025.
ChatGPT's 67% is calculated from the article's statement that it “incorrectly identified 134 articles” in “its two hundred responses”. The article gives per-tool rates only for the tools shown; collectively, the eight tools “provided incorrect answers to more than 60 percent of queries”. Google's AI Mode was not among the tools tested.
Where it fits. The ongoing-monitoring routine set out in watching what search and AI say over time already includes checking what AI answers say about a company on a schedule. AI Mode monitoring can automate part of that check. It does not replace a source-by-source feed, which matters more now that some of those feeds are closing: Reddit's RSS feeds, the last logged-out route for scripts, stop on 13 November, as set out in can you still monitor Reddit for mentions. Google says its monitoring reads forums; it has not said whether that includes Reddit.
The other side: other people monitoring you
The more consequential change may be the one a company cannot see. Once monitoring is available to everyone, customers, investors, job candidates, competitors and journalists can each ask AI Mode to keep them updated on a company. Google is also going to suggest such tasks unprompted, according to Stein, while people search. Each of them will receive Google's summary of whatever it has found, at the moment it finds it, in an app the company has no view into.
That raises the stakes on a question that already matters for AI answers generally: which pages about a company are there to be found. Google has not said how monitoring picks its sources. The nearest published rule is in its documentation on AI features: to be shown as a supporting link in AI Overviews or AI Mode, “a page must be indexed and eligible to be shown in Google Search with a snippet”. Google also says there are “no additional requirements” and no special markup for AI features. If monitoring updates draw on the same index, which is our inference and not Google's statement, then the practical points are the familiar ones: coverage that is indexed quickly, readable as text and allowed to show a snippet is coverage an update can use. How long indexing takes is in how long Google takes to index a new page.
It also changes the timing of a bad day. A synthesised update delivered “the moment new info is available”, in Stein's June phrase, means a damaging forum thread or an inaccurate report can reach the people watching a company before a correction exists. The first version of events an update describes is the one those people get.
The question Google has not answered
For publishers and for companies alike, the open question is the same one Search Engine Journal raised: how Search picks the sources behind an update, and whether the update links to them. Google's June wording promised links; the September wording does not mention them. A notification that summarises a news story without a link replaces a visit to that story. One that links to it can send traffic. Search Console's report on AI features counts impressions but shows no clicks, as set out in what Search Console shows for AI answers, and Google has not said whether monitoring updates are counted in it at all, so a publisher has no direct way to see what monitoring sends it either.
Until Google documents the feature, the practical way for a company to learn how it behaves is to use it: set a task for its own name, its executives and its category, and compare what the updates say, and link to, against what its existing monitoring finds over the same weeks.
Methodology and limitations
Google's statements are quoted from the Google Search I/O 2026 post on blog.google, dated 19 May 2026, and from Robby Stein's posts on X of 12 June and 28 September 2026, which we read in full through a public mirror of X's post data because X does not serve posts to logged-out readers. Google Alerts is described from Google's help page “Create an alert”, and AI Mode's link eligibility from Google Search Central's “AI features and your website”. All were read on 1 October 2026. We searched blog.google, Google Search Central and Google's help centre for a dedicated page on AI Mode monitoring or information agents and found none. Search Engine Journal's reports of 12 June and 28 September and Search Engine Roundtable's of 28 September were read as secondhand accounts and are used only where they agree with the primary posts; one other report we read gave the 12 June Ultra announcement as the date of the global rollout, and we did not rely on it. We did not test the feature: it runs in the Google app on a signed-in account, and we have not been able to confirm that the rollout has reached an account we can use. Everything said here about how it might behave for a company is reasoning from Google's own descriptions, and is marked as such. Google may document or change the feature at any time.
Conclusion
Google AI Mode can now be asked, by anyone, to keep watching a topic and report back, and a company is a topic like any other. As a monitoring tool for the company itself it is, for now, an undocumented extra: free, broad and possibly better than keyword alerts at catching coverage that does not use the company's name, but without the exports, history or source lists that make a monitoring record usable. Google Alerts still does the narrower job of listing new results more transparently. The bigger change is outside the company's view. Customers, investors and journalists can now receive Google's running summary of what is being said about a business, built from sources Google has not said how it chooses. The only lever a company has over that summary is the one it has over AI answers generally: making sure the pages about it that are indexed, current and quotable are accurate ones.
Frequently Asked Questions
What is information monitoring in AI Mode?
A feature of AI Mode in Google Search that keeps checking for new information on a topic you describe and sends you an update when it finds something. Google introduced it at I/O on 19 May 2026 as “information agents”, launched it for AI Ultra subscribers on 12 June, and began rolling it out to all users on 28 September.
Can AI Mode monitor news about my company?
Probably, in the sense that you can ask it to. Google says you can tell it “exactly what to look for” and that it checks sites, forums, social posts and Google's real-time data. But every example Google has given is consumer-facing, such as restaurants, events and price drops, and Google has published no guidance on using it to track a company. We have not tested it.
How is it different from Google Alerts?
Google Alerts emails you new search results that match a query, with settings for frequency, source type, language, region and volume. AI Mode monitoring takes a description in plain language and sends a written, synthesised update in the Google app. Alerts gives you a list of links you can check; AI Mode gives you a summary whose sources Google has not said how it chooses.
Does an AI Mode update link to the sources it used?
Google said in June that updates would include “links to the web”. The 28 September announcement does not mention links. Google's documentation says a page must be indexed and eligible to show a snippet to appear as a supporting link in AI Mode, but it does not describe monitoring updates specifically.
Should a communications team replace its monitoring tools with it?
No. Nothing Google has published describes exporting results, keeping a history, sharing a task with colleagues or counting mentions. A synthesised summary is also harder to audit than a list of links. It is better treated as a way to see what an AI summary of your company might look like to someone else.
Can customers or journalists set AI Mode to watch my company?
Anyone with the feature can presumably ask it to watch any topic, including a company. What they receive is Google's summary of whatever sources it picks, so the pages about you that are indexed, current and quotable are what that summary can draw on.
Sources
- official Google, Liz Reid, Google Search's I/O 2026 updates: AI agents and more, 19 May 2026. Information agents: sources, “synthesized update”, launch first for AI Pro and Ultra subscribers. Read 1 October 2026.
- official Robby Stein, vice president of product, Google Search, post on X, 12 June 2026. Information agents available to AI Ultra subscribers in all AI Mode languages and markets; updates with “links to the web”. Read 1 October 2026 via a public mirror.
- official Robby Stein, post on X, 28 September 2026. Info monitoring rolling out to everyone globally; sources checked; suggested tasks; three example uses. Read 1 October 2026 via a public mirror.
- platform documentation Google Search Help, Create an alert. Google Alerts settings and delivery. Read 1 October 2026.
- documentation Google Search Central, AI features and your website, last updated 10 December 2025. Supporting links must be indexed and snippet-eligible; no additional requirements. Read 1 October 2026.
- trade press Matt G. Southern, Search Engine Journal, Google Rolls Out AI Mode Info Monitoring To All Users Globally, September 2026, and its June report on the Ultra launch. Secondhand; used for the comparison of the two posts. Read 1 October 2026.
- trade press Barry Schwartz, Search Engine Roundtable, Google Rolls Out Monitoring Capabilities In AI Mode To All, September 2026. Secondhand. Read 1 October 2026.
- research Klaudia Jaźwińska and Aisvarya Chandrasekar, Tow Center for Digital Journalism, AI Search Has A Citation Problem, Columbia Journalism Review, 6 March 2025. Incorrect-answer rates by tool. Read October 10, 2026.