Google opened AI Mode's information monitoring to all users on Sept. 28, having first offered it to Ultra subscribers on June 12 and then Pro subscribers. One day later, on Sept. 29, OpenAI launched dots, its always-on ChatGPT agents with their own cloud computer, browser, and access to connected apps.

The two products serve different roles. Google's feature tracks changing information within Search and sends updates; dots manage assigned tasks across connected apps and a cloud computer. What they share is persistence: the request stays active after the user leaves. Matt Southern, covering this for Search Engine Journal, proposes the term "standing instruction" for a request that remains live — his own coinage, not official language from either company.

For local businesses, retailers, and publishers, later updates could include information that appeared after the initial search. That is Southern's analysis, not a confirmed behavior. None of the documentation reviewed details when a page is fetched, evaluated, or selected.

Standing requests predate both products

Google Alerts has long converted topics into regular searches, emailing users "whenever we find matching search results." OpenAI's release notes date scheduled tasks in ChatGPT to January 2025. A July 9, 2026 entry introducing ChatGPT Work says Work can use scheduled tasks "that run once, repeat on a schedule or trigger, or monitor for changes." The current Scheduled tasks help page defines monitoring tasks as checks that "send a notification when a relevant update occurs" and can "use information from previous runs."

Google framed the gap in older tools at I/O in May. VP of Search Liz Reid, in an interview Roger Montti covered in July, described people who "have to keep checking and checking and checking," and cited a stock-tracking request with specific parameters where, in her words, "Nothing's going to do that."

How Google builds the monitoring task

Google's I/O post in May introduced information agents that "look across everything on the web, like blogs, news sites and social posts," plus real-time finance, shopping, and sports data. Reid wrote that an apartment hunter can "brain dump all of the exact requirements" and the agent "will continuously scan for you." Users create a task by adding "keep me updated" to a search, and updates arrive "along with helpful links to explore further on the web."

Robby Stein, VP of Product for Google Search, said in June that agents would send "detailed updates and links to the web." His Sept. 28 post shifts language: it refers to "info monitoring capabilities" rather than "information agents" and says Search will "continuously check across changing info on the web." It does not repeat the links language from May and June.

Reid's May post also says the agent sends a synthesized update "with the ability to take action" — without specifying what action, who takes it, or whether approval is required. The September post does not mention actions at all.

How OpenAI turns a prompt into ongoing work

OpenAI's launch post describes dots as "always-on agents" with their own cloud computer, browser, and connected apps. The getting-started page says a dot "can take on ongoing responsibility and keep making progress between conversations." Dots "decide when to pause and wake up to continue," and users can specify a time or supported event that triggers work.

Event monitoring depends on the connected service supporting it. "Connecting Slack or another source alone doesn't create a monitoring task," the Tasks page notes.

Proactive research runs alongside assigned work without a new request. Dots can gather updates by reading permitted sources and saving private notes before anyone asks — but their research tools cannot directly message others, change content through plugins, or control browsers or computers. That limit places a browser-based website watch outside proactive research. The documentation also notes that "some websites block cloud browsers or require further verification."

Dots are rolling out to Pro users outside the European Economic Area, Switzerland, and the UK, and to Business Premium users across supported regions. Enterprise beta access is off by default.

The shared layer: persistent intent

Both products let a user set a goal and apply stored criteria to information arriving later. Google's May post says a notification arrives "when listings meet your needs"; OpenAI's Tasks page suggests users define which changes deserve a notification. Google chooses the source categories Search checks; a dot's access depends on connected accounts and granted permissions.

Timing differs. Google's June post says updates happen "the moment new info is available," while a dot operates on a schedule, a supported event, or its own timing.

The safety documentation diverges sharply. OpenAI's safety page details Custom Rules and Auto-review, including user-defined limits and steps like password changes returned to the individual. Google's side offers only the May post's "ability to take action."

Google's leadership has signaled the direction. In an interview Montti covered in April, Sundar Pichai said "a lot of what are just information seeking queries will be agentic search" and that users "will do long-running tasks, can be asynchronous" — projections, not confirmed behaviors.

Why this matters for website visibility

Google's September examples for local businesses covered nearby new restaurants, pop-ups, and holiday activities, citing sites, forums, and social media as sources — without explaining how Search chooses among them for local topics. Retailer examples included back-in-stock alerts and price-drop notices, with the Shopping Graph named as a source; Google doesn't specify whether price updates come via merchant feeds, page crawls, or both. Its AI features page advises keeping Merchant Center and Business Profile data current.

For publishers, the May post included news sites among the sources agents scan, so an article published after a task was set could appear in a later update — though nothing establishes whether a given article would be selected, linked, or clicked.

Measurement splits into three distinct events that shouldn't be combined. Sites featured in AI Overviews and AI Mode appear in the Performance report under the Web search type. The Generative AI performance report — which Google says reached websites worldwide by Aug. 31 — counts impressions, meaning how many times links to a site appear in a generative AI feature. A retrieval alone is not a displayed-link impression, and a click-through is a separate event. Google counts external-link clicks in AI Mode in Search Console, while Analytics measures subsequent site activity. As of Sept. 30, neither the AI features page nor the report page mentions monitoring updates.

On the OpenAI side, an agency could hand a dot a recurring check across its connected platforms, but that's internal tooling — it reveals nothing about how customers find a site. Watching a public competitor page through a browser falls outside proactive research. As of Sept. 30, OpenAI's crawler page lists four user agents, none for dots.

What remains unexplained

Google's announcements don't clarify how Search selects sources for updates, whether every update includes links, or which user agents additional fetching might use. The September post says "globally" but omits languages and when Pro access arrived.

OpenAI's documentation doesn't specify which browser or fetch identity a dot uses on the public web, whether public-web updates include citations, or how a site owner can distinguish a dot's visit from other ChatGPT activity.

Google ended its September post with "Lots more to come!" Monitor for a dedicated monitoring help page, a Search Console update explaining how these updates are counted, and OpenAI's promised dots usage terms for each plan — the company said it would share them after the first month. A dot-specific crawler identifier and citation guidance would give site owners the clearest signals yet.