OpenAI has introduced dots, an always‑on AI personal assistant that lives inside the ChatGPT desktop and mobile apps for Pro‑tier subscribers. The service runs continuously, linking to a user’s calendar, email and other tools, and attempts to anticipate work that needs attention. For owners of small and mid‑sized businesses, the promise is less manual scheduling and more proactive support, but the trade‑off is increased exposure of personal and corporate data to an external AI platform.
What happened
According to TechCrunch, OpenAI unveiled dots at its developer day in late September and demonstrated it at TechCrunch Disrupt in October. Alexander Embiricos, who leads OpenAI’s enterprise product team, explained that the assistant is designed as a single, persistent conversation rather than a series of discrete queries. The product is available to ChatGPT Pro users, with plans starting at $100 a month. Once connected, dots can read calendar entries, scan email, and interact with other computer resources to flag conflicts, suggest actions and even conduct “proactive research” on behalf of the user.
Embirocis noted that similar functionality could be assembled using OpenAI’s Codex coding tool, but that would require more manual setup. Dots, by contrast, is meant to feel intuitive, with the assistant deciding when to intervene. In early testing, the assistant warned of a calendar conflict only three minutes before the event, a timing the team expects to improve. The product also includes privacy controls mirroring those of ChatGPT: users can disable model training and limit resource access, though the controls are not yet highly granular.
Why it matters for businesses
The core value proposition for businesses is the reduction of routine coordination work. An assistant that watches calendars and inboxes can surface scheduling conflicts, upcoming deadlines and even preliminary analysis of operational incidents without a direct prompt. For sales or operations teams that already use CRM or workflow tools, such a capability could free up time for higher‑value activities.
However, the always‑on nature of dots raises data‑privacy concerns. The assistant needs ongoing access to email, calendar and potentially cloud drives to be effective, and OpenAI’s model may retain conversation data unless training is turned off. While OpenAI claims the same privacy settings as ChatGPT, the lack of fine‑grained permissions means businesses must decide how much personal or client information they are comfortable sharing. The assistant’s ability to perform “proactive research” also implies it will periodically query external sources on the user’s behalf, which could have compliance implications in regulated markets.
Practical steps for owners and managers
- Start with a pilot – Enable dots for a single team member on a non‑critical account to evaluate its usefulness and observe how often it intervenes.
- Review permission settings – Before connecting email or cloud drives, audit which folders and accounts the assistant can access and disable any read‑only access that is not required.
- Turn off model training – Use the privacy control to prevent OpenAI from using your conversations to improve its models, reducing data exposure.
- Set clear escalation rules – Define situations where the assistant should seek explicit approval before taking action, such as sending emails or scheduling meetings.
- Monitor usage and costs – Track the number of tokens consumed by the assistant and compare it against the $100‑per‑month subscription to ensure the ROI justifies the expense.
The bottom line
Dots represents a step toward truly proactive AI assistants that operate without a prompt, offering potential productivity gains for busy business owners. The technology is still early; timing of alerts and the granularity of privacy controls need refinement. Companies that adopt dots should begin with a controlled trial, tighten permission settings and keep a close eye on data‑privacy implications. If the assistant can reliably surface the right information at the right time, it could become a valuable complement to existing CRM and automation workflows, but the decision to grant it continuous access must be weighed against the risk of exposing sensitive business data.