New RewardGet found. Get trusted. Get chosen.
Book a strategy call
Checked across

Surface marks show where New Reward checks visibility. They are not partner badges or outcome guarantees.

Back to Field guide
Field guide
Published |7 min read

OpenAI just shipped always-on Dots. Proactive research is read-only — your pages still have to be fetchable.

On September 29, 2026, OpenAI launched Dots — always-on agents in ChatGPT. Help Center pages opened this fire say a dot can look for useful updates before you ask (“proactive research”) by reading permitted connected sources and saving private notes. Those research tools cannot send messages, change content through plugins, or control a browser or computer. Cloud browser use is a separate capability for assigned work. OpenAI did not publish ranking formulas, citation rates, or a promise that background research will open your first-party URL. Conditions and proof still beat another dashboard seat.

James Brady portrait

By James Brady

Chief AI Officer, Product & AI Operations

ChatGPTGEOAI visibilityOpenAI

Google gave Searchers free AI Mode monitors on Monday. OpenAI gave Pro and Business Premium users always-on Dots on Tuesday. Neither product replaces first-party pages agents can actually fetch.

On September 29, 2026, OpenAI published Introducing dots. The launch post says dots are powered by GPT‑6 Astra, have their own cloud computer, connect through ChatGPT’s plugin ecosystem, and can work toward goals around the clock. It names “proactive research”: when you aren’t actively working with a dot, it looks for ways to help in the background using apps you’ve already connected, with tools restricted to be read-only — they can’t send messages, change app content, or control your browser or computer.

Sources opened this fire: Introducing dots (openai.com, dated September 29, 2026; HTTP 200); Dots privacy, security, and safety FAQs (Help Center; HTTP 200; Updated ~13 hours before this fire); Getting started with your dot (Help Center; HTTP 200); Manage dots in ChatGPT workspaces (Help Center; HTTP 200 — Enterprise cloud browser / network / computer toggles). Adjacent trade clock: OpenAI Dots Run Read-Only Research When Nobody Is Asking, Search Engine Journal / Matt G. Southern, published 2026-09-30 (HTTP 200 — points at the same Help pages; not the primary).

The short version

  • Sep 29, 2026: OpenAI launches Dots — always-on agents in ChatGPT with their own cloud computer and plugin connections.
  • Availability (OpenAI’s words): rolling out to Pro users in markets excluding the European Economic Area, Switzerland, and the UK; Business Premium across all supported ChatGPT regions; Enterprise / Edu / Healthcare beta when a workspace admin enables it (off by default). Access may take several days. First dot included at no extra cost on Pro / Business Premium.
  • Proactive research (Help Center): a dot can look for useful updates before you make a request. It can read information from permitted connected sources and save private notes. Research tools cannot directly: send messages to other people; change content through plugins; control a browser or computer. Custom Rules cannot remove those research restrictions.
  • Assigned work is different: Enterprise workspace docs list separate Cloud browser use, Cloud network access, and Cloud computer use toggles. A browser that can open websites during a task is not the same mode as read-only proactive research on connected apps.
  • OpenAI did not publish how often a public first-party URL appears inside a Dot’s work, a ChatGPT Search citation lift from Dots, or ranking guidance for SEO/GEO.
  • Operator move: keep first-party facts consistent with the apps buyers actually connect (calendar, email, Slack, docs, CRM); keep pages crawlable for agents that do get browser access on a task; run one proof loop if you have access. Monitoring ≠ execution. Tools ≠ citations.

What OpenAI actually said

Launch post (Sep 29). OpenAI describes dots as frontier intelligence with their own cloud computer, learning from feedback, working 24/7, and connecting through an ecosystem of plugins (“over 4,000 apps” is OpenAI’s figure). On proactive research:

When you aren’t actively working with it, your dot looks for ways to help in the background. We call this “proactive research”. It does this by using the apps you’ve already connected with tools that are restricted to be read-only, which means that they can’t send messages, change app content, or control your browser or computer.

Privacy / safety FAQ. Same product rule in Help Center form:

Your dot can also look for useful updates before you make a request. This is called proactive research. It can read information from permitted connected sources and save private notes, but its research tools cannot directly: Send messages to other people. Change content through plugins. Control a browser or computer.

The FAQ also states proactive research restrictions cannot be removed by the user or the dot, and that OpenAI doesn’t train directly on proactive background research or its notes (information may still enter eligible conversations depending on settings — see the FAQ; not a New Reward invent).

Getting started. Availability carve-outs (EEA / Switzerland / UK for Pro), Business Premium coverage, Enterprise beta off by default, gradual rollout. Scheduled runs and proactive updates appear in the conversation; scheduled work is managed from the dot’s profile.

Enterprise manage page. Workspace owners get separate cloud controls for browser use, network access, and computer use. Enabling dots does not grant access to every app or website; plugin permissions and each service’s authorization still limit what a dot can access.

That is the product message. No ranking guarantee. No “brands will be cited X% more because Dots exist.”

Four different “watching” words (do not mash them)

  1. Google AI Mode info monitoring (Sep 28 Stein; packed yesterday for New Reward as ai-mode-info-monitoring-global — still 404 this fire, do not link). Consumer Search watches topics across sites, forums, social, real-time data, Shopping Graph.
  2. OpenAI Dots proactive research (this pack). Background read of permitted connected sources; no browser/computer control in that mode.
  3. Dots / Work Cloud browser during assigned work. Separate capability; Enterprise can disable cloud browser use. This is where a public page might be opened — still not a published citation rate.
  4. ChatGPT Search citations / Search Console generative AI impressions. Different products. A Dot watching Slack is not a Search Console AI Mode impression and not a ChatGPT Search inline citation.

Mashing those four in a client deck is how you invent outcomes OpenAI and Google did not publish.

What this means for a New Reward operator

  1. Treat Dots as buyer-side agents, not a ranking badge. Someone on Pro or Business Premium can park always-on work that reads connected apps and saves notes overnight. Your job is still to be a clear, fetchable first-party source for the facts those people and agents use — offers, hours, inventory, policies, proof — not to claim OpenAI will push you into every Dot’s notes.
  2. Ship the pages and the connected facts. If buyers’ dots read email, calendar, Slack, or CRM plugins, inconsistency between your site and those surfaces is a visibility problem. Forums and social are Google’s monitoring list; connected apps are OpenAI’s proactive list. Same operator discipline: one truth, dated pages, NAP/hours/stock that match.
  3. Separate proactive research from browser tasks. Read-only connected-source research cannot control a browser. Cloud browser use is a different switch. If you care whether an agent can open your checkout, docs, or service pages, that is eligibility for assigned browser work (and for WebMCP-style tools where they exist) — not a claim about proactive research.
  4. Run one proof loop if you have access. When Dots reach an account you control, (a) connect one low-risk app and note whether proactive updates appear without a new prompt; (b) assign one research task that needs a public URL in your category and screenshot whether the cloud browser opens your domain and whether any citation UI appears. That is evidence. It is not a market-wide rate.
  5. Keep WebMCP and citations honest. New Reward’s live Field Guide already says site tools are not citations. Dots sit beside that stack: an agent that can browse or call tools still needs a page worth fetching. OpenAI’s launch and the three Help pages opened this fire do not link Dots to search rankings or ask publishers to modify sites for Dots.

Honesty block

  • OpenAI did not publish that proactive research crawls the open web for SEO/GEO citations. The Help Center says permitted connected sources and private notes.
  • OpenAI did not publish ranking formulas, citation rates, or a guarantee that a Dot’s assigned browser task will open or cite a given domain.
  • OpenAI did not publish a country-by-country Pro rollout calendar beyond excluding EEA, Switzerland, and the UK at launch, or an exact Enterprise beta invite list.
  • “Over 4,000 apps” is OpenAI’s plugin-ecosystem figure from the launch post — not a New Reward count of apps that matter for your category.
  • Cloud browser use, cloud network access, and cloud computer use are separate Enterprise controls from proactive research’s read-only limits. Do not describe proactive research as “Dots browse your site overnight” unless you are describing an assigned task with browser capability enabled — and even then, OpenAI published no visit rates.
  • Google AI Mode info monitoring, Search Console generative AI impressions, ChatGPT Search citations, and Dots proactive research are different products. Do not mash them.
  • Google, Microsoft, Perplexity, and Cloudflare did not confirm anything about this OpenAI launch in pages we opened as primary for this thesis.
  • No invented caps on how often New Reward customers appear after a scan or after a Dot task.

Related New Reward pages (live, HEAD/GET 200 this fire)

James Brady portrait

James Brady

Chief AI Officer, Product & AI Operations

James Brady writes from the operating-system side of New Reward: product direction, AI workflows, verification, and the proof boundaries that keep visibility work honest.

Start here

Ready to turn this into a next step?

Book the appointment if you want the live walkthrough. Run the score if you want the fast read first.

Book the AI-visibility demoGet your AI-visibility score

Where this fits

This is part of how New Reward improves search and AI visibility. See the SEO, AEO, and GEO offering or read how search and AI visibility work together.

Prefer New Reward in Google

Keep reading

More from the Field Guide

Field guide

7 min read

Google just called manual fact-check of AI content “critical.” Titles and schema count too.

On October 1, 2026, Google Search Central updated its guidance on using generative AI content on your website. The docs now say generative models predict a likely sequence of words from training data, so outputs may contain inaccuracies (hallucinations), and that it is critical to manually fact-check and review all AI-generated content for accuracy and trustworthiness before publishing. That review also applies to metadata that can appear in Search results — title elements, meta descriptions, structured data, and image alt text. Google did not publish a new ranking formula, a citation rate for AI Overviews or AI Mode, or an end date for the September 2026 spam update. Conditions and proof still beat another dashboard seat.

GoogleGEO

Read the guide
Field guide

7 min read

Google just made AI Mode monitoring free for everyone. Your pages still have to be the source.

On September 28, 2026, Google Search VP Robby Stein said AI Mode’s info monitoring — previously for Ultra and Pro subscribers — is rolling out to everyone globally. People can ask Search to watch topics across sites, forums, social posts, real-time data, and the Shopping Graph. Google did not say how sources are chosen or whether every update still carries web links. First-party pages that stay fresh are the condition. A dashboard watching your brand is not the same job as being the page a buyer’s monitor can ground.

AI ModeGEO

Read the guide
Field guide

6 min read

Google’s September 2026 spam update is rolling. Here’s the 14-day operator plan.

On September 24, 2026 at 09:15 PDT, Google’s Search Status Dashboard released the September 2026 spam update. It applies globally and to all languages. Google says the rollout may take up to two weeks. This is the fourth spam update of 2026 after March, June, and August. Google did not name a new spam tactic in the incident note. Annotate Search Console during the window, review spam policies if rankings move, and do not invent a targeting story Google did not publish.

SEOspam update

Read the guide
← Back to Field guide
DemoGet your score