OpenAI Dots for Professional Services
OpenAI's dots can sign into your apps and work around the clock. Here is where they help a professional services firm, and what to check before they touch client work.
OpenAI's dots are a capable always-on assistant for a person's own work at a professional services firm: research, inbox and calendar, follow-ups, and one-off tasks the person checks. For repeated client work like document collection, engagement setup, and data entry, they plan each run fresh and keep records per user, so most professional services firms will want a verifiable agent for that work and keep dots for the rest.
This fall, three AI companies shipped the same idea within weeks: an agent with its own computer that works for you around the clock. xAI launched Grok Bot in August, Meta launched Muse in September, and OpenAI answered with dots at the end of the month. Operations teams at professional services firms are already asking whether one of them can take work off their plate.
This guide is for that question. It covers what OpenAI's dots are, where they fit at a professional services firm, and the five things the firm has to answer before an always-on agent touches repeated client work. Every product fact comes from OpenAI's own pages as of October 1, 2026, and these products are changing fast.
What is OpenAI Dots?
dots are OpenAI's always-on agents, announced at DevDay on September 29, 2026. Each dot runs on GPT-6 Astra with its own cloud computer and browser, connects to over 4,000 apps through ChatGPT, and works toward your goals in the background. You can reach your dot in ChatGPT, Slack, and Microsoft Teams.
dots learn from feedback over time, can delegate to other agents, and OpenAI says it envisions teams of dots. Specialist dots with their own identity are in an enterprise pilot.
dots are good at the open-ended work of a busy professional: following up on threads, researching ahead of a meeting, drafting an update, and keeping track of loose ends across thousands of apps.
- Plans. Your first dot is included with ChatGPT Pro (from $100 a month, not yet in the EEA, Switzerland, or the UK) and Business Premium. Enterprise, Edu, and Healthcare get a beta an admin has to turn on.
- Model. GPT-6 Astra.
- Custom Rules. Each rule is set to take action without asking, take action if pre-approved, ask before acting, or hand off to you.
- Auto-review. A separate system checks planned actions, like emails and file changes, against your instructions and rules before they run.
- Hard stops. Deleting data and granting security-sensitive access always need confirmation. Password changes and transfers between financial accounts are handed back to you.
- Records. Activity View shows tasks in progress, scheduled, and completed. Business, Enterprise, and Edu content is not used for training by default.
Product details as of October 1, 2026, from the vendor's own announcement, help, and security pages.
Where professional services firms want to use dots
Nobody at a professional services firm needs an agent to do the professional judgment. They want back the hours their teams spend on repeated, system-to-system work:
- Client document collection. Requesting, chasing, and filing the documents each engagement needs.
- Engagement setup. Sending the engagement letter, opening the client and project in the PSA, and setting up the folder structure.
- Data entry. Moving figures from client PDFs and spreadsheets into QuickBooks, the tax system, or the CRM.
- Time and invoicing. Chasing timesheets, building invoices, and reconciling billed against worked.
- Recurring reporting. Pulling the same numbers into the same client report every month.
Picture a consulting firm's operations team using dots to run engagement admin: chase timesheets on Fridays, draft invoices, and nudge clients about unpaid balances, with a Slack summary each morning. With access to over 4,000 apps, a dot can reach most of the tools a firm uses.
Reach is not the hard part; consistency is. Each Friday's timesheet chase and each invoice is planned fresh, so the firm gets a capable assistant, not a procedure. When a client disputes an invoice, the firm needs to show how it was built, from which time entries, under which rate card, and that answer should not depend on what the model decided that week.
Five questions before dots touch client work
None of these are reasons to avoid dots. They are the questions any professional services firm has to answer for any agent that acts on a client's behalf.
Will dots do the work the same way every time?
You give a dot a goal and define what it can do on its own; it works out what needs to happen next. Scheduled tasks and Custom Rules shape that, but OpenAI does not describe a saved, step-by-step workflow, and its help center notes that a dot "can make mistakes, including when following your rules."
For document collection, engagement setup, and data entry, that matters. You cannot test a procedure on last month's cases and rely on the result if the next run is planned again from scratch.
Who supervises dots at a professional services firm?
OpenAI's controls are layered: Custom Rules per action type, an Auto-review system that sits outside the dot's environment, hard stops for destructive and security-sensitive actions, and stricter recipient checks as data gets more sensitive. Those are real safeguards. Like the others, they decide which individual actions need a person, not whether the overall procedure is right.
For accounting firms, the AICPA's quality management standard (SQMS No. 1, effective December 15, 2025) asks firms to design a system of quality management that covers the technology they rely on. Consulting and advisory firms answer to engagement terms and client audits that ask the same question: who did the work, and how was it checked?
Can the firm show what dots did?
Activity View shows each user what their dot is doing, has scheduled, and has finished. OpenAI's dots materials do not describe an exportable audit log or an admin-level log of dot activity.
Engagement files need to show what was received, what was entered, and who reviewed it. If an agent moved the numbers, the file needs to show what it did, with which document, on which day.
What happens to client data?
Business, Enterprise, and Edu content is not used for training by default; on personal plans it depends on the Improve the model setting, which can include actions your dot takes. Disconnecting an app does not delete what your dot already pulled from it; removing that data means resetting the dot.
Tax preparers are limited by Internal Revenue Code Section 7216 in how they use and disclose return information, and the AICPA Code treats confidential client information as a core duty. An agent with a staff member's email and file access sits inside both.
OpenAI says plainly: "Dots can still make mistakes, so always review consequential work."
Do dots reach the systems professional services firms run on?
Over 4,000 apps through the ChatGPT ecosystem, plus Slack and Teams as places to talk to your dot. OpenAI's dots materials do not single out specific practice-management, document-management, or CRM connectors.
Professional services operations run through QuickBooks or the GL, a practice-management or PSA tool, the CRM, Microsoft 365 or Google Workspace, and Box, SharePoint, or Drive for files. Reaching a system is only half of it; the other half is doing the same thing in it every time.
| OpenAI Dots | Caddi | |
|---|---|---|
| How the work is defined | A goal plus Custom Rules | Tested code you can read and change |
| Same input, same path? | Not claimed by the vendor | Yes, for the coded steps |
| Role of AI in a run | Plans and performs each step | Named judgment steps, scoped and logged |
| Human control | Custom Rules, Auto-review, and hard stops | Review the procedure; exceptions go to a named person |
| Record of each run | Per-user Activity View; no stated audit export | Every run logged, tied to its source documents |
| Built for | A person's own varied work | Professional services firms' repeated back-office workflows |
What a verifiable agent does differently
The fix is not a better prompt or a stricter approval setting. It is moving the procedure out of the model. A hybrid agent runs the workflow as code, which can be read, tested, and versioned, and calls AI only for the steps that need judgment, each one scoped and logged.
For client document collection, that means the code sends the request, checks what arrived against the engagement checklist, files each document, and enters the figures the same way every time, while AI handles the judgment inside it: which document is the K-1, which number is the ending balance on a messy statement. Anything it is unsure of goes to a staff member, logged.
That split is what makes an agent verifiable: a reviewer can read the procedure before it runs, every judgment call is on the record, and changes are made on purpose.
More on the idea in what are verifiable AI agents and what are hybrid agents.
Connect Quickbooks to Caddi
Allow1Sign in to your tools
One click each for Quickbooks, Gmail and Box. No API keys.
Automate client document collection.
ScreenshareType it2Show or tell it
Screenshare the job, or type it in chat. Change it any time.
- Running
14 done1 for review
3Watch it run
Caddi does the work and flags anything unsure for review.
How professional services firms can use both
OpenAI's dots and a verifiable agent are not competing for the same job. Let people use dots for the work around them, where every task is different and the person who asked checks the result.
Put the repeated client work, document collection, engagement setup, and data entry, on an agent the firm can verify. That is what Caddi builds: an ops person shows Caddi the workflow on a screen share or describes it in chat, Caddi runs it as code across the firm's tools with the judgment steps logged, and the team changes it in plain English as the work changes. Across Caddi customers, 99% of agent runs complete successfully.
Keep reading
- What are verifiable AI agents?
- Grok Bot vs. Muse vs. Dots for professional services firms
- Grok Bot for professional services
- Meta Muse for professional services
- OpenAI Dots for law firms
- OpenAI Dots for RIAs
- Caddi for accounting firms
- Data entry automation
- Document automation
- Caddi + QuickBooks
- AI use cases for accounting firms
- What are hybrid agents?
Caddi
See how Caddi AI Agents can run repeated client work the same way every time and log every run for review
Frequently asked questions
Is OpenAI Dots safe for professional services firms?
OpenAI's dots have real safeguards: custom Rules, Auto-review, and hard stops. For a person's own work, those are sensible. For repeated client work, a professional services firm also has to answer for supervision, records, and confidentiality, and dots plan each run fresh with records kept per user. Most firms will keep it for personal work and use a verifiable agent for client workflows.
Can OpenAI Dots run document collection, engagement setup, and data entry unattended?
They can attempt it: dots work in the background and can reach many apps. But each dot decides how to handle each case on that run, and the vendor does not claim repeatable runs, so the firm cannot test the procedure once and rely on it. For unattended client work, a verifiable agent that runs the steps as code is the safer fit.
Does OpenAI Dots keep an audit trail?
Activity View shows each user what their dot is doing, has scheduled, and has finished. OpenAI's dots materials do not describe an exportable audit log or an admin-level log of dot activity. For professional services firms, check whether that record can be exported, retained, and reviewed at the firm level before relying on it.
Does OpenAI Dots train on our data?
Business, Enterprise, and Edu content is not used for training by default; on personal plans it depends on the Improve the model setting, which can include actions your dot takes. Disconnecting an app does not delete what your dot already pulled from it; removing that data means resetting the dot.
What should a professional services firm use for repeated client work?
A verifiable agent: one whose procedure you can read, whose judgment steps are logged, and whose every run leaves a record. Caddi builds these for professional services firms. Your team shows the workflow on a screen share or describes it in chat, and Caddi runs it as code across your tools, with exceptions routed to a named person.