By
Gigabit Systems
November 5, 2026
•
20 min read
Meet Dots: The AI Assistant That Keeps Working After You Close ChatGPT
Meet Dots: The AI Assistant That Keeps Working After You Close ChatGPT
Your Next Employee Might Never Need a Lunch Break.
Imagine having a personal assistant who remembers your priorities, checks your calendar, researches important decisions, organizes information, and follows up on unfinished projects.
Now imagine that assistant continuing to work while you’re in meetings, driving home, or sleeping.
You don’t need to know how to program. You don’t need to understand complicated AI terminology. And you don’t need to sit in front of a computer issuing instructions all day.
That’s the idea behind Dots, a new artificial intelligence agent introduced by OpenAI, the company behind ChatGPT.
Launched in September 2026 and now rolling out to eligible users, Dots represents a major shift in how artificial intelligence can be used.
Until recently, most people used AI like a sophisticated search engine: ask a question, receive an answer, and move on.
Dots is designed to do something fundamentally different.
Instead of simply answering your questions, it can take responsibility for getting things done.
For individuals, that might mean organizing a family vacation, monitoring important deadlines, or helping manage a busy schedule.
For businesses, it could mean tracking sales opportunities, preparing reports, researching vendors, monitoring projects, and identifying work that needs attention.
But giving artificial intelligence access to your email, documents, calendar, and business systems also introduces serious cybersecurity and privacy considerations.
Understanding both the opportunities and the risks is essential.
First, What Exactly Is a Dot?
Think of a Dot as a digital assistant that lives inside ChatGPT.
Unlike an ordinary chatbot, your Dot is designed to remember what you’re working on and continue making progress between conversations.
It has access to its own cloud-based computer and browser.
That means it can research information, analyze documents, work with connected applications, and prepare results for you to review.
You can communicate with it through ChatGPT, and supported users can also interact with it through platforms such as Slack or Microsoft Teams.
You can even have a voice conversation with your Dot.
Here’s a simple example.
With regular ChatGPT, you might ask:
“Help me organize a vacation to Florida.”
ChatGPT could recommend hotels, suggest activities, and prepare an itinerary.
But once the conversation ends, you would typically need to return and tell it what to do next.
With Dots, you could assign an ongoing responsibility:
“Help me plan our family vacation. Research hotels, compare options, keep track of our budget, and remind me when important decisions need to be made. Don’t book or purchase anything without asking me.”
Your Dot can continue working on that project, organize the information, and return when it has something useful to report.
You remain responsible for important decisions.
The difference is simple: ChatGPT helps you think. Dots can help you follow through.
What Makes Dots Different From Other AI Assistants?
There are five particularly important capabilities.
1. It Remembers Context
Dots can start with relevant information from your existing ChatGPT memory and learn from your feedback.
For example, if you regularly prepare professional reports using a particular format, you can teach your Dot that preference.
Over time, you may need to repeat fewer instructions.
However, memory can also include sensitive context. Users should review what information they are comfortable allowing an ongoing AI agent to use.
2. It Can Work While You’re Away
Dots operates using its own cloud computer.
You don’t necessarily need to leave your laptop running.
Depending on the task, permissions, and available tools, it can continue researching, preparing documents, and making progress between conversations.
3. It Can Connect to Applications You Already Use
Dots can work with supported services such as calendars, email, document storage, and business collaboration platforms.
For example, a connected calendar can help it understand upcoming appointments.
Connected documents can provide the information needed to prepare a report.
You choose which applications to connect and authorize.
4. It Can Monitor Ongoing Responsibilities
Instead of asking an AI assistant to perform the same task repeatedly, you can give your Dot a continuing responsibility.
For example:
“Every weekday morning, review my upcoming appointments and tell me what requires preparation.”
Your Dot can create a recurring task and provide updates according to the schedule you establish.
5. It Can Ask for Decisions
A useful assistant shouldn’t blindly make every decision.
Dots is designed to bring important questions back to you.
You can establish boundaries requiring approval before sending communications, changing information, making purchases, or taking other consequential actions.
These capabilities are powerful, but they are not unlimited. What a Dot can actually do depends on your plan, connected applications, permissions, and the tools available to it.
How to Set Up Your First Dot
You do not need technical experience to get started.
As of November 2026, Dots is rolling out to eligible ChatGPT Pro and Business Premium users. Enterprise, Education, and Healthcare workspaces have additional eligibility and administrative requirements.
Dots is not currently included with every ChatGPT subscription, and availability varies by region and rollout status.
Step 1: Open Dots
Visit:
You can also access Dots through the supported ChatGPT desktop or mobile applications.
Make sure your application is updated.
If you don’t see the option, your account may not yet be eligible or the rollout may not have reached you.
Step 2: Create Your Assistant
Follow the onboarding instructions.
You can give your Dot a name, choose its appearance, and personalize how you interact with it.
The name is entirely up to you.
You might call it Assistant, Office Manager, Researcher, or something more personal.
Step 3: Explain What You Want Help With
This is the most important part.
Don’t begin by giving your Dot access to everything.
Start with one specific responsibility.
For example:
“You’re my personal assistant. Help me organize my weekly schedule, identify upcoming deadlines, and prepare a short morning summary. Do not contact anyone or change any appointments without my permission.”
That establishes a clear objective and initial boundaries.
Step 4: Connect Only the Applications You Need
If you want your Dot to review your calendar, connect your supported calendar application.
If you want it to analyze documents, connect the appropriate document source.
You do not need to connect your email, cloud storage, and every other application immediately.
Start small.
Step 5: Establish Rules
Explain what your Dot may do independently and what requires approval.
For example:
“You may research information, read the calendar I’ve authorized, and prepare draft reports. You may not send emails, modify files, make purchases, or share information externally without my explicit approval.”
Use the available permission and Custom Rules controls to reinforce those instructions.
A written instruction is useful, but it is not a substitute for restricting actual access.
Step 6: Give It a Test Assignment
Start with a simple task.
Ask it to prepare tomorrow’s schedule, research a purchase, or organize a small project.
Review the results carefully.
Correct mistakes and explain your preferences.
As you become comfortable, gradually assign more responsibilities.
Five Practical Ways Individuals Can Use Dots
The best way to understand this technology is through real examples.
Example 1: Your Personal Morning Briefing
The problem: You have appointments, responsibilities, and deadlines scattered across your calendar and notes.
The solution: Have your Dot prepare a personalized morning briefing.
How to set it up:
Create your Dot and connect your supported calendar.
Ask it to review upcoming appointments.
Tell it what information matters most.
Schedule the briefing for a convenient time.
Review the first few briefings and refine the format.
Instructions to give your Dot:
“Every weekday at 7:00 AM, prepare a morning briefing. Show today’s appointments in chronological order, identify scheduling conflicts, remind me about deadlines within the next seven days, and highlight anything requiring preparation. Keep it concise. Do not change appointments or contact anyone.”
Your briefing might include an early meeting, an upcoming payment deadline, and a reminder that you need to prepare materials for tomorrow’s appointment.
Instead of checking multiple sources, you begin the day with an organized summary.
Example 2: Planning a Family Vacation
The problem: Researching hotels, attractions, transportation, and pricing takes hours.
The solution: Assign the research and comparison work to your Dot.
How to set it up:
Tell your Dot the destination and travel dates.
Provide your approximate budget and preferences.
Ask it to research accommodations and activities.
Have it organize options into a comparison.
Require your approval before any reservation or payment.
Instructions to give your Dot:
“Help me plan a five-day family vacation to Orlando. Our total budget is $5,000. Research hotels, transportation, and family-friendly activities. Compare at least five hotel options by price, location, amenities, and cancellation policy. Prepare an itinerary and identify anything that needs to be reserved. Do not make bookings or spend money without my approval.”
Your Dot can organize the research and bring decisions back to you.
You still need to verify prices and availability before committing.
Example 3: Researching an Expensive Purchase
Suppose you’re shopping for a new refrigerator, laptop, or vehicle.
Instead of spending hours reading specifications and reviews, ask your Dot to research the options.
Instructions to give your Dot:
“I’m shopping for a new laptop with a maximum budget of $1,500. I need Windows 11 Pro, at least 32GB RAM, a 14-inch display, and strong battery life. Compare five suitable models, identify their advantages and disadvantages, and provide links to reputable sellers. Alert me to significant price reductions over the next two weeks. Don’t purchase anything.”
This is particularly useful when you have clear specifications but don’t have time to compare dozens of products.
Your Dot can prepare the research and, where supported, perform scheduled price checks.
Example 4: Managing a Home Renovation
Imagine renovating a kitchen.
You have contractor estimates, appliance choices, measurements, deadlines, and dozens of decisions.
Upload the relevant documents or connect an authorized project folder.
Instructions to give your Dot:
“Help me manage my kitchen renovation. Organize the contractor estimates, compare pricing, track decisions, and maintain a list of outstanding questions. Every Friday, prepare a project summary showing completed items, upcoming milestones, and potential delays. Never approve expenses or contact contractors without asking me.”
The benefit isn’t simply getting one answer.
It’s maintaining continuity throughout a project that might last several months.
Example 5: Keeping Track of Important Responsibilities
Dots can also help with recurring personal administration.
For example, you could ask it to maintain a checklist of upcoming renewals, insurance deadlines, household maintenance, or school-related responsibilities.
Instructions to give your Dot:
“Help me keep track of household responsibilities. Maintain a list of the renewal dates and maintenance schedules I provide. Remind me 30 days before major renewals and one week before routine maintenance. Do not access financial accounts or authorize payments.”
This is a useful beginner project because it delivers practical value without requiring extensive access to sensitive information.
Six Powerful Business Use Cases
For businesses, Dots may become significantly more valuable.
Many companies employ people whose time is consumed by repetitive administrative work.
Preparing reports, tracking deadlines, researching vendors, following up on projects, and organizing information are necessary activities.
But they don’t always require continuous human involvement.
An AI agent can potentially handle portions of that work while employees focus on customer relationships, judgment, and higher-value responsibilities.
Business Example 1: The Executive Assistant
Best for: Business owners, executives, managers, and busy professionals.
The problem: Meetings, deadlines, and administrative responsibilities compete for attention.
Setup:
Connect an approved business calendar and, if appropriate, limited email access.
Ask your Dot to prepare a daily executive briefing.
Instructions:
“Act as my executive assistant. Every weekday at 7:30 AM, review my authorized calendar and relevant emails from the previous business day. Identify today’s meetings, urgent deadlines, unanswered messages requiring my attention, and scheduling conflicts. Prepare a briefing organized by priority. You may draft responses, but you must not send them or modify appointments without my approval.”
The result: A daily overview of what matters, without manually reviewing every calendar entry and email thread.
Start with calendar-only access if you’re concerned about granting broad mailbox permissions.
Business Example 2: Sales Follow-Up and Opportunity Tracking
Best for: Sales teams, professional services firms, and small business owners.
The problem: Sales opportunities are frequently lost because someone forgets to follow up.
Setup:
Connect an approved CRM or provide a regularly updated sales spreadsheet.
Explain your sales process and follow-up expectations.
Instructions:
“Every weekday, review the sales opportunities in the authorized CRM. Identify prospects with no activity for seven days, proposals awaiting responses, and opportunities approaching their expected closing dates. Prepare a prioritized follow-up list with suggested email drafts. Do not contact prospects or modify CRM records without approval.”
The result: Salespeople receive a prioritized list of opportunities that require attention.
The Dot can reduce administrative effort without replacing the judgment required to manage customer relationships.
Business Example 3: Weekly Business Performance Reports
Best for: Business owners, operations managers, and department heads.
The problem: Important information is scattered across spreadsheets, reports, and departmental updates.
Setup:
Provide access to approved reports or upload the relevant files.
Define the metrics you want tracked.
Instructions:
“Every Friday at 3:00 PM, review the approved weekly sales and operating reports. Prepare a one-page executive summary showing revenue, expenses, outstanding receivables, major changes from the previous week, and issues requiring management attention. Identify missing or inconsistent data. Do not estimate figures unless clearly labeled.”
The result: A consistent management report that highlights trends and exceptions.
Importantly, the Dot should identify missing data rather than invent numbers.
Financial results must still be reviewed against authoritative accounting records.
Business Example 4: Managing Customer Projects
Best for: IT companies, contractors, consultants, marketing agencies, and professional services firms.
The problem: Projects involve multiple people, changing requirements, and deadlines that can easily be overlooked.
Setup:
Share the project scope, timeline, approved documentation, and relevant task sources.
Instructions:
“Help manage this customer project. Track outstanding deliverables, deadlines, dependencies, and decisions requiring customer approval. Every Monday, prepare a status report showing completed work, work in progress, upcoming milestones, and potential risks. Flag scope changes separately. Do not commit to new pricing, deadlines, or contract terms.”
The result: Better project visibility and fewer overlooked responsibilities.
This is especially valuable for MSPs managing migrations, hardware deployments, office relocations, or cybersecurity improvement projects.
Business Example 5: Vendor Research and Purchasing
Best for: Office managers, procurement teams, and business owners.
The problem: Comparing vendors can involve hours of research, pricing analysis, and contract review.
Setup:
Provide product specifications, purchasing requirements, and a budget.
Instructions:
“Research three qualified vendors for 25 business laptops. Each must include Windows 11 Pro, 32GB RAM, and a three-year warranty. Compare pricing, warranty coverage, delivery estimates, and vendor reputation. Prepare a recommendation with source links and clearly identify any information that could not be verified. Do not place orders or share our company information without approval.”
The result: A structured comparison that management can evaluate.
Business Example 6: Cybersecurity and IT Operations Reporting
Best for: MSPs, internal IT departments, and security-conscious organizations.
The problem: IT teams often receive large volumes of alerts, reports, and technical information.
Setup:
Begin with exported, sanitized reports or an approved read-only integration.
Instructions:
“Every Monday, analyze the approved endpoint security and patch compliance reports. Identify devices missing critical updates, recurring security alerts, endpoints without recent check-ins, and trends requiring investigation. Produce a prioritized report with supporting evidence. Do not modify security settings, disable accounts, isolate devices, or execute remediation without explicit authorization.”
The result: An organized operational summary that helps technicians prioritize their work.
A Dot should not replace an EDR platform, SIEM, SOC, or incident response team.
It can assist with analysis and reporting, but security decisions and remediation require appropriate technical oversight.
The Most Important Skill: Learning How to Give AI Good Instructions
You don’t need to become a programmer.
But you do need to communicate clearly.
Compare these two requests.
Vague instruction:
“Manage my business.”
That leaves far too much room for interpretation.
Useful instruction:
“Every Monday, review the weekly sales spreadsheet, compare results with the previous week, identify accounts with declining sales, and prepare a one-page report. Use only the supplied data. Do not contact customers or change records without approval.”
The second instruction establishes the objective, information source, schedule, output, and boundaries.
A useful formula is:
Goal + Information + Schedule + Output + Permissions
For example:
Goal: Track outstanding customer proposals.
Information: Approved CRM records.
Schedule: Every weekday at 8:00 AM.
Output: A prioritized summary with recommended next steps.
Permissions: Read and draft only; no external communications without approval.
The more clearly you define these elements, the more useful and predictable your Dot is likely to become.
The Security Risks: What Happens When AI Has Access to Your Business?
This is where organizations need to be especially careful.
A conventional chatbot can provide an incorrect answer.
An AI agent with access to business systems may be capable of taking an incorrect action.
That difference is enormous.
Imagine an AI assistant connected to your email, calendar, customer database, and cloud storage.
If it misinterprets instructions, processes malicious content, or receives excessive permissions, the consequences could extend beyond an inaccurate response.
It could potentially expose confidential information, modify records, send unintended communications, or disrupt business processes.
OpenAI has introduced safeguards for Dots, including permissions, action rules, automated reviews, and monitoring.
Those protections reduce risk.
They do not eliminate it.
Risk 1: Giving AI Too Much Access
A business owner might connect an entire Microsoft 365 environment when the Dot only needs access to a single calendar.
That creates unnecessary exposure.
The solution: Apply least privilege.
Grant only the access required for the assigned task.
Start with read-only permissions whenever possible.
Avoid giving an AI agent global administrator privileges or unrestricted access to confidential repositories.
Risk 2: Prompt Injection
Prompt injection is one of the most important emerging AI security threats.
It occurs when malicious instructions are embedded in content that an AI system reads.
Imagine your Dot is reviewing incoming vendor emails.
An attacker sends an email containing hidden instructions telling the AI to ignore its previous directions and forward confidential documents to an outside address.
The email is not an authorized instruction from you.
It is untrusted content attempting to manipulate the AI.
OpenAI has built protections intended to distinguish user instructions from malicious content, but prompt injection remains a real security concern.
The solution: Restrict what the agent can do, require approval for sensitive actions, and treat external emails, websites, and documents as untrusted information.
Risk 3: Confidential Information Exposure
A Dot connected to business applications may encounter customer information, employee records, contracts, financial documents, or other confidential material.
That information may be subject to contractual, privacy, or regulatory obligations.
For healthcare organizations, protected health information requires particular care.
Law firms must consider client confidentiality and privilege.
Schools must consider student privacy requirements.
The solution: Use an appropriately governed business environment, confirm the service and features are approved for the information involved, and avoid granting access to sensitive data without a documented need.
Risk 4: Incorrect Decisions and Hallucinations
AI can produce convincing answers that are factually wrong.
An agent might misunderstand a contract clause, misread a spreadsheet, or recommend an unsuitable vendor.
If it is allowed to act independently, those mistakes can become operational problems.
The solution: Require human review of financial, legal, medical, security, and other consequential decisions.
For reports, require source references and clear identification of uncertain information.
Risk 5: Unintended Actions
Imagine asking your Dot to help clean up old documents.
You intended for it to identify unnecessary files.
It interprets the instruction as permission to delete them.
Even if the system allows you to stop the task, stopping it does not necessarily undo changes already made.
The solution: Explicitly distinguish between reviewing, drafting, recommending, modifying, and deleting.
Require approval before destructive or externally visible actions.
Risk 6: Persistent Access and Memory
Dots can retain useful context and proactively review connected information.
That means access and information exposure are not necessarily limited to the moment you are actively chatting.
Disconnecting an application also does not automatically erase information the Dot previously obtained.
The solution: Periodically review connected applications, saved context, scheduled tasks, and permissions.
When employees change roles or leave an organization, include AI agent access in your offboarding process.
Ten Rules for Using AI Agents Safely
Whether you’re an individual or a business, these are the precautions I recommend.
Start with one low-risk task. Don’t give an unfamiliar AI agent responsibility for critical business operations on day one.
Connect only necessary applications. A calendar assistant doesn’t need access to your accounting system.
Prefer read-only access. Allow the AI to observe and recommend before granting permission to modify information.
Require approval for external communications. Emails, messages, and customer-facing actions should be reviewed before sending, especially during initial deployment.
Never give unrestricted financial authority. Purchases, transfers, banking changes, and payments should require independent human authorization.
Protect administrator credentials. Do not give agents broad administrative access merely for convenience.
Review activity regularly. Check what the Dot has researched, accessed, prepared, or changed.
Use organizational accounts for business work. Avoid placing confidential company information in an employee’s personal AI environment.
Verify important information. AI-generated reports, pricing, legal interpretations, and technical recommendations can contain mistakes.
Establish an emergency stop procedure. Know how to pause the Dot, revoke application permissions, disable local computer access, and investigate unintended actions.
The guiding principle is straightforward.
Give AI responsibility for preparing work before giving it authority to execute work.
What Businesses Should Do Before Deploying Dots
For organizations considering Dots, the first step should be establishing a basic AI governance policy.
It doesn’t have to be a complicated 50-page document.
But it should answer several important questions.
Which employees may use AI agents?
Which applications may they connect?
What categories of company information are permitted?
What actions require management approval?
Who reviews agent activity?
How are incidents reported?
How are permissions removed when an employee leaves?
For larger organizations, IT administrators should review the available workspace controls before enabling Dots.
Eligible Enterprise workspaces can manage permissions for Dots, connected communication platforms, local computer access, and Custom Rules.
During the November 2026 administrative preview, eligible Enterprise, Education, and Healthcare workspaces can enable Dots for selected users. Administrators should review the November 18 default-availability transition and their organization’s policies.
Businesses should also evaluate usage limits, billing, data handling, and any regulatory requirements before deployment.
For an MSP, this creates an important opportunity to help clients introduce AI safely through application assessments, permission reviews, employee training, and ongoing security governance.
The objective should be productive adoption, not unrestricted automation.
How Much Does Dots Cost?
As of November 2026, Dots is rolling out to eligible ChatGPT Pro and Business Premium users, with Enterprise options also available.
The first Dot is included with eligible Pro and Business Premium plans, subject to usage allowances.
More substantial work delegated to ChatGPT Work or Codex may consume separate usage allowances.
Enterprise organizations have additional pricing and billing arrangements.
Because availability and pricing are evolving, users should check the current plan details before subscribing or deploying Dots across an organization.
You can learn more at:
The Bigger Picture: AI Is Moving From Conversation to Responsibility
The arrival of Dots represents an important change in artificial intelligence.
For years, AI assistants have primarily responded to individual requests.
They helped write emails, summarize documents, answer questions, and generate ideas.
Those capabilities are valuable.
But most of the responsibility for coordinating the work remained with the human user.
You still had to remember what needed to be done, gather the information, issue instructions, review the output, and initiate the next step.
Dots attempts to reduce that burden.
Instead of constantly telling AI what to do next, you can give it an ongoing responsibility and ask it to keep track of progress.
For individuals, that could mean fewer forgotten deadlines and less time spent organizing information.
For businesses, it could mean faster research, better project visibility, more consistent reporting, and fewer administrative bottlenecks.
But this technology also changes the cybersecurity equation.
When artificial intelligence moves from providing information to taking action, security becomes even more important.
Organizations must carefully control permissions, monitor activity, protect sensitive data, and preserve human accountability.
The Bottom Line: Hire Your AI Assistant Carefully
Dots has the potential to make artificial intelligence useful to people who have never considered themselves particularly technical.
You don’t need to understand machine learning.
You don’t need to write code.
You simply need to identify work that consumes your time, explain what you want accomplished, and establish clear boundaries.
Start with a daily briefing.
Then perhaps a research project.
Then an ongoing administrative responsibility.
Review the results, refine the instructions, and expand gradually.
For businesses, treat AI agents the way you would treat a new employee with access to important company systems.
Give them the information they need.
Explain their responsibilities.
Limit their permissions.
Review their work.
And don’t give them the keys to everything on their first day.
The future of AI isn’t just a chatbot that answers questions. It’s an assistant that remembers what needs to get done—and helps make sure it actually happens.
The opportunity is enormous.
So is the responsibility to deploy it securely.
70% of all cyber attacks target small businesses, I can help protect yours.
#ArtificialIntelligence #AIAgents #Cybersecurity #ManagedIT #BusinessAutomation
Imagine an employee who works 24/7, remembers everything, and never needs a vacation. OpenAI just made that possible. Here’s how to put it to work—and what you should NEVER let it access.
Official Sources and Setup Guides
I verified the setup process, capabilities, security safeguards, and rollout requirements against OpenAI’s current documentation:
See some more of our most recent posts...