01
Understand what Dots are
OpenAI Dots are always-on AI agents powered by GPT-6 Astra. Unlike a normal chatbot, a Dot can:
- Keep working between conversations
- Use its own cloud computer and browser
- Work across connected apps
- Remember relevant context
- Handle recurring tasks
- Research proactively
- Schedule future work
- Delegate certain work to other OpenAI tools and agents
- Ask for your judgment when needed
Think of a Dot as an AI operator with ongoing responsibilities, not just an AI you prompt occasionally.
02
Give your Dot a business role
Do not start with random tasks. Give it a clear responsibility such as:
- Growth Operator
- Research Assistant
- Sales Assistant
- Marketing Assistant
- Content Operator
- Customer Intelligence Analyst
- Competitive Researcher
- Operations Assistant
- Founder Assistant
Define exactly what success means for that role.
03
Connect your business tools
Connect only the apps your Dot genuinely needs. Useful categories include:
- Email
- Calendar
- Slack
- Business documents
- Project information
- Research sources
- Communication tools
- Other supported plugins
Review permissions carefully before granting access.
04
Build a business knowledge base
Teach your Dot the context behind your company. Give it:
- Business model
- Products and services
- Target customers
- Ideal customer profiles
- Brand positioning
- Competitors
- Pricing
- Marketing strategy
- Sales process
- Content guidelines
- SOPs
- KPIs
- Current priorities
- Examples of high-quality work
The better the context, the more useful its ongoing work can become.
05
Turn market research into a continuous process
Ask your Dot to regularly investigate:
- Competitor launches
- Industry developments
- Customer problems
- Emerging trends
- New technologies
- Market opportunities
- Content trends
- Changes relevant to your business
Instead of doing occasional research, build a recurring intelligence system.
06
Create a content engine
Use your Dot to help:
- Research topics
- Discover content opportunities
- Analyze competitors
- Generate content briefs
- Repurpose existing content
- Organize ideas
- Review performance
- Maintain a content calendar
- Identify successful themes
- Suggest future experiments
Keep final brand-sensitive publishing decisions under human review.
07
Improve sales processes
A Dot can support your sales workflow by helping:
- Research prospects
- Prepare account briefs
- Summarize customer information
- Prepare meeting notes
- Identify follow-up tasks
- Draft personalized outreach
- Organize sales research
- Track relevant developments
- Prepare proposals and materials
Require approval before sensitive external communication or commitments.
08
Build a customer intelligence loop
Ask your Dot to continuously organize customer signals from approved sources. Look for:
- Common questions
- Feature requests
- Objections
- Complaints
- Desired outcomes
- Buying motivations
- Repeated pain points
- Customer language
Turn those insights into better products, marketing and sales messaging.
09
Automate recurring operations
Schedule recurring work such as:
Daily
- Priority briefing
- Calendar review
- Important communication review
- Industry research
Weekly
- Competitor summary
- Content opportunities
- Customer insights
- Growth experiment ideas
Monthly
- Business performance review
- Strategic research
- Process improvement opportunities
10
Use Dots for growth experiments
Give your Dot a structured growth objective. Example:
- Goal
- Increase qualified inbound leads.
- Context
- Explain your product, audience and current acquisition channels.
- Responsibilities
- Research opportunities, identify bottlenecks, propose experiments and analyze results.
- Constraints
- Define budget, brand rules and actions requiring approval.
- Output
- A prioritized list of experiments with evidence, assumptions and measurable KPIs.
11
Create custom rules
Dots support controls governing how they handle actions. Depending on the action, you can configure behaviors such as:
- Take action without asking
- Take action when pre-approved
- Ask before taking action
- Hand the task back to you
Use stricter controls for
- Payments
- Purchases
- Sensitive information
- External communication
- Account changes
- Customer commitments
- Destructive actions
- High-impact business decisions
12
Use the Human + Dot model
Let your Dot handle
- Research
- Monitoring
- Organization
- Repetitive workflows
- Preparation
- First drafts
- Routine analysis
Keep humans responsible for
- Strategy
- Judgment
- Approvals
- Relationships
- High-stakes decisions
- Financial commitments
- Sensitive communications
13
Measure business impact
Track whether your Dot actually improves:
- Hours saved
- Leads generated
- Response time
- Content output
- Research speed
- Conversion rates
- Customer satisfaction
- Sales productivity
- Operational costs
- Revenue influenced
Automation without measurable business value is just activity.
14
Improve it continuously
Review its work regularly. Tell it:
- What was good
- What was wrong
- What information was missing
- What your preferences are
- What requires approval
- What great output looks like
Use this feedback and persistent context to make future work more aligned with your business.
15
The business growth system
Each layer adds to the last:
- Business goal
- + Relevant business context
- + Connected apps
- + Clear responsibilities
- + Scheduled work
- + Custom rules
- + Human review
- + Performance measurement
= An always-on AI business operator
Best principle
Do not ask: “What can my Dot do?” Ask: “What recurring responsibility in my business can I safely give my Dot?”
- Start with one
- Prove value
- Improve
- Expand
Start with one measurable workflow, prove its value, improve the process, then expand into additional responsibilities.
