ChatGPT Dots: an AI assistant that keeps work moving.

What is ChatGPT Dots?

An individual assistant is called a dot; dots is the product family. OpenAI describes a persistent agent that keeps context and supports ongoing work. Beyond an individual chat request, it can pursue a goal across conversations and delegate tasks to other agents.

Our assessment: what matters to a business is turning a broad request into a verifiable result. Define the process, required material and acceptance criteria. A precise brief such as “Prepare outstanding issues from our project documents every Monday” is easier to evaluate than “Organise our company”.

Is ChatGPT Dots available in Germany?

OpenAI’s current getting-started guide distinguishes plans: Pro is rolling out outside the European Economic Area, Switzerland and the UK. Germany is in the EEA. Business Premium is listed across supported ChatGPT regions. Enterprise, including Edu and Healthcare, gets a beta that a workspace administrator must enable; it is off by default. Access rolls out gradually.

Check actual access in your account before scheduling a pilot or choosing a different plan. An arbitrary Business subscription is not evidence of Dots access: the current help page explicitly specifies Business Premium. It does not list Dots access for Free or Plus.

What does Dots cost and which limits apply?

OpenAI’s billing guide lists Business Premium at US$100 per user per month billed annually, or US$125 billed monthly. Business requires at least two Standard or Premium seats in total, and a mix is possible. These are plan prices, not a separate Dots price. Check actual checkout terms; we do not convert them to euros here.

Current Dots documentation says the first dot is included without an extra charge in Pro or Business Premium. Conversations with it do not count towards ChatGPT usage limits. Deeper work has an allowance, with extended limits in the first month after launch. Delegated Work or Codex tasks use their usual allowances. This does not promise permanently unlimited background work.

Compare the plan, tasks, ongoing supervision and rework together. Measure cost per reviewed result. The documentation describes future additional dots and capacity options; those announcements do not establish a reliable extra price today.

Cloud computer, apps and optional access to your device

A dot has its own cloud computer and browser. Cloud work can continue with your laptop off. Connected apps follow their permissions. Local access is additional, initially disabled and revocable. It requires your computer to be online with the ChatGPT app open. Cloud browsers do not automatically inherit local logins; your company VPN or internal applications are not automatically available there.

For a first pilot, connect only the required source. An approved project folder is easier to assess than several entire mailboxes. Record which systems may be read and where results should be stored. A working cloud session does not demonstrate access to every internal application.

Setup and communication: ChatGPT, Slack and mobile app

Create your dot in the desktop app or desktop browser. Name it and connect the apps and channels you need. Afterwards, use the mobile ChatGPT app where your access is enabled. Slack and voice calls you initiate are supported; a dot cannot initiate calls at launch. The administrator guide still describes Microsoft Teams as an invite-only alpha. Text messaging is not promised for Germany; the help guide describes only a limited US Pro beta.

Start with a defined output: an outstanding-issues table, a source briefing or an internal draft. Specify separately where any message should go. Test a channel first with non-sensitive material.

What do proactive and always-on mean?

Dots can support ongoing goals, reminders and recurring checks. The documentation distinguishes fixed schedules from monitoring supported events. Set timings such as “every weekday at 8 am” explicitly. Proactive research can read permitted sources and keep private notes. Its research tools cannot directly send messages, change content through plugins or control computers.

For businesses, a focused reporting rhythm helps: report a meaningful change, its source and the proposed next step. Use the pilot to check whether the required source actually supports events. Preparing a result and sending it are separate steps.

Approvals, Custom Rules and safety checks

Plugin connections and permissions are shared across Dots, ChatGPT, Work and Codex. Review accounts you have already connected, too.

Custom Rules define boundaries for supported actions. Auto-review separately checks certain planned actions; core safeguards cannot be switched off through those rules. OpenAI describes human takeover for highly sensitive steps such as money transfers or password changes. External content may carry malicious instructions. The safeguards reduce this prompt-injection risk without eliminating it.

Follow work in Activity View. When stopping it, the controls documentation says Pause halts the main task, without automatically stopping every delegated task or future schedule. Stop those separately. Define who approves drafts and how the team recognises a failed task.

Privacy and memory: what remains stored?

Dots share context with ChatGPT Memory and can retain information from apps. Individual dot memories currently cannot be selectively viewed, edited or deleted. Disconnecting an app stops new access without removing context already learned. Deleting the dot removes its context; separately stored files, chats and ChatGPT memories need separate deletion steps.

OpenAI says Business, Enterprise and Edu data are not used for model training by default. On personal plans, training depends on data controls. That alone does not establish whether your particular use meets privacy requirements. Our recommendation: agree data types, access, contracts, retention and a complete deletion process with your company’s responsible people before the pilot.

Three useful pilot tasks for SMEs

Project preparation: collect outstanding questions for the next meeting from approved documents. The output lists the document, source location, question and responsible role. The project lead decides priorities. Assess completeness and correct attribution rather than simply the length of the summary.

Market monitoring: watch a limited list of public vendor sources. Request only changes relevant to your products or processes. A useful briefing separates verified facts, possible implications and unanswered questions. Your specialist checks whether the proposed change actually affects the business.

Proposal preparation: structure an approved enquiry against your service catalogue and identify missing details. Sales reviews the draft, prices and commitments. Measure the whole process through an approved template, including corrections. A quick AI draft is not yet a dependable proposal.

A sample brief for your first Dots pilot

“Every Monday at 8 am Europe/Berlin, review only the approved project folder. Prepare an internal draft of outstanding issues, each with a source and date. Flag missing information. Do not message other people or edit source files. Report missing access or ambiguous issues. The project lead reviews the result. Confirm the saved schedule, time zone and output destination.”

Adapt the brief to your process and configure matching rules. Test prepared examples, including a missing file and a conflicting statement. Measure handling time including review, errors, rework and ongoing maintenance. Decide on additional sources or tasks only afterwards.

When is Dots worth evaluating for your business?

Dots is particularly interesting when a task needs following over several days and information comes from different sources. A one-off text revision usually needs less structure. For fixed, rule-based processes, we also consider conventional automation. With confidential data, access and deletion boundaries matter alongside output quality.

lol.marketing helps you make this decision independently of vendors. The free AI System Check helps assess your workflow, systems and key constraints. That provides a starting point: test Dots, evaluate another AI solution or improve the process first. We do not sell OpenAI licences or claim a vendor partnership.