5 live days · Your work, in practice
Keep recurring project work moving.
Choose Intensive to move beyond one-off prompts into proactive agents that run defined work over time. The five days cover curriculum days 3–7.
Build an AI working method around a project you recognize. Start with a useful brief, give recurring preparation and follow-up to proactive agents, then coordinate several proactive AI agent teams to carry a client or internal project through to a reviewed handover.
A professional often owns the space between an unclear request and a result someone else can use. The course follows a sample service-improvement project, with room to substitute a comparable assignment from your role while keeping scope, commitments and review responsibility clear.
What you learn
Learn to build and supervise proactive agents for request intake, research updates, action tracking, deliverable preparation and reporting, with clear permissions and exception review.
Create and test proactive AI agents for project preparation. Name a role such as Project Follow-through, then hand suitable work to OpenAI Dots to follow accepted requests and dependencies between your reviews. Choose one run or a continuing responsibility, with approved connections, agreed reports and a clear way to pause or retire the work.
Proactive Agents · OpenAI Dots
Create it. Name it. Put it to work.
Create and test a proactive AI agent, give its role a stable name, and define what it may do. Then hand supported work to OpenAI Dots for one run or a continuing responsibility. Your Dot can keep working between conversations, using approved context and connected tools, and return results or ask for a decision without a fresh prompt for every step.
Give the role a name and a brief
Name the role Project Follow-through. Provide the approved request, scope, evidence, owners, deadline and acceptance criteria. Limit connections to the permitted project queue and folder; record the accepted brief so later requests can be checked against it.
Choose a run or an ongoing responsibility
A Dot has its own cloud computer and browser. When the assignment, files and tools are available there, it can work while your phone or computer is off. Always-on means available to continue the responsibility; it may wait for a schedule, input or approval and remains subject to access and usage limits. Work using local files or applications still needs that connected computer online with the ChatGPT app open.
Set the return and the review rhythm
Return current actions, draft deliverables and decisions holding up progress under the same named responsibility, with source versions and completed, waiting and blocked work clearly separated. Agree whether you need a scheduled report, an update on a material change, or a decision request. The project owner approves scope, commitments and stakeholder communication.
Agree when results are due, which changes deserve an alert and where to review them. Confirm recurring schedules, time zones and end dates; connecting an app alone does not enable monitoring. Test a changed input and a blocked step. Before retiring the responsibility, stop its delegated work and cancel its schedules separately. Keep the reviewed instructions so you can recheck access and resume when needed.
Connect only what the role needs
Use approved project material and limit personal or customer data to what the task needs. Keep client-confidential records and employee information out of unapproved connections. If your role involves hiring or other decisions about people, use fictional examples and keep those decisions with the authorized human reviewer.
A project connection does not authorize access to every client workspace or permission to contact stakeholders. Define the permitted project, recipients and actions before continuing work.
This practice is part of the existing proactive agent assignments and review time. It does not add class days or extend the hours shown in your booking.
The learning map
Work units by day
5 days of practical work for Consultants, project leads, managers and professionals who turn information into plans, decisions and client work. Every unit below has an output and a review point.
Your learning plan: these are the tasks you will practice. The time suggested for each task is a guide; your booking shows the confirmed class dates and start and end times.
This is a five-day Intensive: curriculum days 3–7. Intro and the MasterCourse Capstone are separate stages, not included in this purchase.
Proactive Intake agent · Request triage
Organize requests as they arrive
- 3.1Define the request proactive intake agent
- 3.2Test repeat and incomplete requests
Proactive Research agent · Changing evidence
Keep the project evidence current
- 4.1Create an evidence-update proactive agent
- 4.2Review weak and conflicting findings
Action proactive agent · Dependencies
Keep actions and dependencies visible
- 5.1Maintain the action register
- 5.2Handle delays and blocked work
Delivery proactive agent · Controlled revisions
Prepare the next useful deliverable
- 6.1Build a deliverable-preparation proactive agent
- 6.2Respond to reviewed changes
Project reporting · Ongoing operation
Test the proactive agents over a project cycle
- 7.1Produce a recurring project review
- 7.2Test interruption and resumption
The work behind the map
Daily detail
Follow the work from its starting material to a result you can inspect. Each checkpoint gives you a reason to approve, revise or stop.
Intensive
Days 3–7 · Proactive Agents
Keep recurring work moving.
Build proactive agents that run defined tasks over time. Give each one a trigger, working record and limits, then review its exceptions and results.
Build the proactive agent and test one complete run. Name its role, approve its inputs and connections, then hand suitable work to a Dot. Choose a single assignment or a continuing responsibility with clear reports, review points and a way to stop.
Day 03Organize requests as they arrive
Build a proactive agent that processes new sample requests into a consistent queue. Give it a narrow job and preserve what it has already handled.
- 3.1
Define the request proactive intake agent
Set a new-request trigger, required fields, allowed source locations and triage prompt. Save request identifiers and route missing or out-of-scope material to a person.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A request proactive intake agent and a structured work queue.
- 3.2
Test repeat and incomplete requests
Run successive batches containing a duplicate and a request without an owner. Verify saved state, one bounded retry for temporary failure and an explicit unresolved-items list.
Proactive agent practice. Create and test the project preparation proactive agent on one sample run. Name the role Project Follow-through and hand supported work to a Dot using an approved request queue, project brief and evidence folder. Define permitted actions and the return you expect; inspect the first result before giving it an ongoing responsibility.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A tested intake run and exception record.
By the end of the day
A recurring proactive intake agent that prepares work for a person’s prioritization.
Review checkpoint
The proactive agent may classify and draft questions; it must not accept new commitments, change priorities or contact a client without permission.
Day 04Keep the project evidence current
Give a proactive research agent a defined source set and update cycle. Useful monitoring reports what changed and why it matters to the assignment.
- 4.1
Create an evidence-update proactive agent
Configure a scheduled check of approved sample sources. Record the last accepted version, extract relevant changes and connect each finding to the project question.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A research-update proactive agent and source register.
- 4.2
Review weak and conflicting findings
Introduce an outdated source and a contradictory update. Have the proactive agent distinguish observed changes from interpretation and queue unsupported claims instead of merging them into the brief.
Proactive agent practice. Connect the approved evidence location or provide a reviewed export. Set when the proactive research role should check for updates. Return dated findings and conflicts with the accepted brief, keeping source selection within scope.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A change digest with confidence limits and review questions.
By the end of the day
A recurring evidence digest that preserves references and flags what needs judgment.
Review checkpoint
Open a sample of the cited sources and confirm that the reported change is real, relevant and not simply a duplicate.
Day 05Keep actions and dependencies visible
Build a proactive agent around approved project actions. Its work is to maintain a reliable picture of commitments, not to invent owners or deadlines.
- 5.1
Maintain the action register
Trigger an update when approved notes or decisions arrive. Extract actions, owners and due dates, preserve earlier decisions and mark unassigned items for review.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- An action-tracking proactive agent and versioned action register.
- 5.2
Handle delays and blocked work
Simulate a late dependency and a changed deadline. Produce an impact summary and a draft follow-up; require approval before any external message or commitment change.
Proactive agent practice. Confirm the action-register review schedule and escalation rules. Introduce a missed dependency; let the proactive role identify affected work and ask for the owner’s decision without changing commitments.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A dependency report and approval-ready follow-up queue.
By the end of the day
An action record that shows what changed, who owns it and what cannot move yet.
Review checkpoint
Check that inferred dates and owners are not presented as agreed commitments and that blocked work remains visible.
Day 06Prepare the next useful deliverable
Give a proactive agent a repeatable drafting assignment using the accepted brief and evidence. A changed requirement should result in a traceable revision.
- 6.1
Build a deliverable-preparation proactive agent
Define the accepted-brief trigger, template, source snapshot and review standard for a sample implementation plan. Save the draft version and stop when mandatory inputs are missing.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A delivery proactive agent and a first implementation-plan draft.
- 6.2
Respond to reviewed changes
Return a draft with specific feedback and introduce a changed requirement. Have the proactive agent revise only the affected sections, record the changes and escalate conflicting instructions.
Proactive agent practice. Keep deliverable preparation active against the approved brief and review notes. Apply accepted changes, preserve the previous version and return work against the same acceptance criteria at the agreed review point.
Suggested time for this task: 2 hours in Full Course; 1 hour in Crash Course. Part of your class day, not extra time.- Practice result
- A revised deliverable and change log.
By the end of the day
A delivery proactive agent that prepares and revises work against an agreed brief.
Review checkpoint
Review whether the revision satisfies the request without quietly expanding scope or losing a prior accepted requirement.
Day 07Test the proactive agents over a project cycle
Connect the proactive agents through a shared project record. Run a compressed sequence of updates to see what a manager would need to review over time.
- 7.1
Produce a recurring project review
Schedule a proactive reporting agent to collect accepted evidence, actions and deliverable status. Highlight changes, blocked work and decisions needed without sending the report automatically.
Suggested time for this task: 1 hour 40 minutes in Full Course; 50 minutes in Crash Course. Part of your class day, not extra time.- Practice result
- A project-review proactive agent and a concise status pack.
- 7.2
Test interruption and resumption
Interrupt a run, repeat an input and withhold a decision. Check bounded retries, completed-work records and the rule for resuming only when the missing decision is supplied.
Proactive agent practice. Let Project Follow-through continue a cloud-capable assignment while you leave the chat. Inspect the result, block one input and test recovery. Then retire the responsibility: stop active delegates, cancel future schedules and retain the reviewed instructions. Recheck access and inputs before calling it back into service.
Suggested time for this task: 2 hours 20 minutes in Full Course; 1 hour 10 minutes in Crash Course. Part of your class day, not extra time.- Practice result
- A failure-test log and a project-proactive agent operating guide.
By the end of the day
A tested group of proactive agents for recurring preparation, follow-up and reporting.
Review checkpoint
Confirm that each exception has an owner and that repeated inputs do not create repeated tasks or communications.
Your time commitment
How long is each class day?
You attend 5 live classes, one on each scheduled day. Choose Full Course or Crash Course for all 5 days. The times below include breaks.
Full Course
6 hours per day
More time to practice, ask questions and review your work.
Suggested plan for one day
- Lessons and hands-on practice
- 4 hours
- Review and questions
- 30 minutes
- Lunch break
- 1 hour
- Two short breaks
- 15 minutes each
5 class days × 6 hours = 30 hours totalIncludes all scheduled breaks.
Crash Course
2 hours 30 minutes per day
The same topics in shorter classes, with prepared examples and less practice time.
Suggested plan for one day
- Lessons and hands-on practice
- 2 hours
- Review and questions
- 20 minutes
- One short break
- 10 minutes
5 class days × 2 hours 30 minutes = 12 hours 30 minutes totalIncludes all scheduled breaks.
Lunch, breaks and review time are already included in each day. The task times in Daily detail fit inside the practice time above. Your trainer may adjust the plan; the start and end times shown with your selected dates are the ones to follow.
Choose your starting point
One progression. Three choices.
Intro teaches individually directed prompts. Intensive puts recurring work into proactive agents. MasterCourse connects the whole pathway and adds the Factory Capstone.
Intro
Learn to turn a work request into a clear prompt, check a project brief and compare options against the evidence.
Choose Intro for individual summarizing, planning and drafting tasks. Bring a role-relevant example or use the supplied service-improvement scenario.
Intensive
Learn to build and supervise proactive agents for request intake, research updates, action tracking, deliverable preparation and reporting, with clear permissions and exception review.
Choose Intensive to move beyond one-off prompts into proactive agents that run defined work over time. The five days cover curriculum days 3–7.
Create and test proactive AI agents for project preparation. Name a role such as Project Follow-through, then hand suitable work to OpenAI Dots to follow accepted requests and dependencies between your reviews. Choose one run or a continuing responsibility, with approved connections, agreed reports and a clear way to pause or retire the work.
MasterCourse
Learn to connect role-specific prompts and proactive project agents into a Factory of proactive AI agent teams that research, plan, prepare and review work together.
Choose MasterCourse for all ten days and the exclusive Capstone. Test a bounded project from initial request to accepted handover using simulated updates.
Build a Factory of proactive AI agent teams for project preparation. Use a Dot to coordinate continuing work across their handoffs and return current actions, draft deliverables and decisions holding up progress. Test changing inputs, blocked work and a later restart. The project owner approves scope, commitments and stakeholder communication.
Intensive is a five-day purchase, not a seven-day bundle. The three-day Capstone is available only inside the ten-day MasterCourse.
Live guidance · Continued reference
Come prepared. Keep what you learn.
Bring a computer, a reliable connection and a work example you are authorized to use. Sample or sanitized material lets you practice without introducing confidential source files.
Practice with guidance
Work through the assignments live, compare results and use trainer feedback to refine your prompts, proactive agent instructions and review standards.
Revisit your classes
Students have access to recordings of their classes in the members portal.
Specialist tools · Separate subscriptions
Tools serve the assignment.
Two specialist tools chosen for practical fit and an established learning ecosystem. Claude, OpenAI, Gemini and Open Source options support the prompt and proactive agent work. Specialist examples connect it to your field.
Microsoft 365 Copilot
Research, analysis and document work in Microsoft 365; supported agent creation and workflows are listed.
Plan and access
Copilot Business $25.20/user/month on monthly subscription.A separate qualifying Microsoft 365 license is required. Advanced agent usage may be metered; do not imply every workflow is bundled.
Notion AI
Workspace documents, search and Notion Agent support; Custom Agents can handle recurring work.
Plan and access
Business $20/member/month on the annual USD view; Free and Plus have limited AI trials.Custom Agents use Notion credits separately; current listed rate is $10 per 1,000 credits. Confirm workspace permissions.
Do I need my own software access?
Yes. Arrange full access to the required features before class, including suitable paid plans, usage credits and organization permissions. Software costs are separate from tuition. Free access may not cover every exercise.
How these tools support your course
Intensive · Proactive Agents
Use supported Microsoft 365 Copilot and Notion AI features, permitted exports or controlled simulations to process requests, track actions and prepare deliverables. Match proactive agent permissions to the project record.
Vendor information checked September 29, 2026. Plans and limits can change. Proactive Agents use supported connections or approved exported inputs; specialist subscriptions do not imply universal proactive agent integration. Human approval remains part of publishing, releasing work and professional decisions.
OpenAI Dots · Proactive agents at work
Hands-on proactive agent and Dots work starts in Intensive. Check Dots availability, supported connections and your organization’s permissions before class. You authorize each app or browser session; Dots does not inherit all your logins. Subscriptions and usage charges are separate from tuition. Some specialist work needs a reviewed export or an online local application.
A project connection does not authorize access to every client workspace or permission to contact stakeholders. Define the permitted project, recipients and actions before continuing work.
Before you choose
FAQ
Clear scope makes it easier to choose a course and arrive ready to work.
Which level should I choose?
Choose Intro for individual summarizing, planning and drafting tasks. Bring a role-relevant example or use the supplied service-improvement scenario. Choose Intensive to move beyond one-off prompts into proactive agents that run defined work over time. The five days cover curriculum days 3–7. Choose MasterCourse for all ten days and the exclusive Capstone. Test a bounded project from initial request to accepted handover using simulated updates.
What does Factory mean?
A Factory coordinates supervised teams of proactive AI agents around a shared outcome. Each AI agent team has a defined responsibility, a deliverable and an acceptance check for the next handoff. You control permissions, review exceptions and approve release. The MasterCourse-only Capstone tests continuing cycles as well as individual runs.
Are recordings available?
Yes. Students can access recordings of their classes in the members portal.
How long is each class, and how many days do I attend?
This Intensive has 5 class days. Full Course lasts 6 hours each day, including lunch and two short breaks. Crash Course lasts 2 hours 30 minutes each day, including one short break. Choose one format for the whole course. Both cover the same topics; Full Course gives you more practice time. Check your selected dates for the exact start and end times in your time zone.
Which time zone will I see?
The class chooser detects your time zone automatically. Use Change time zone to select another location; the dates and hours update together.
Can Intro or Intensive tuition count toward MasterCourse?
The amount actually paid for Intro and/or Intensive can be credited once toward one MasterCourse enrollment within twelve months of each purchase. Intro and Intensive do not credit each other. No credit toward other programs, software or cash. Credit requests are reviewed against purchase records and eligibility; they are not an automatic checkout discount.
Full refund, transfer and credit policyWhat happens if I miss a day?
A missed day has a guaranteed place on the same curriculum day in a later cohort, free of charge, within twelve months of the original cohort's last scheduled day. After this window, new enrollment uses the then-published fee; day credits have no cash value and cannot buy a different curriculum day. An attended day may be repeated without charge within the same twelve-month window, subject to space. One free transfer of the whole seat or its unused days to a later cohort, in either published format, without a reason or a price increase, within twelve months of the original cohort’s last scheduled day. A day already consumed on one format is consumed on both; unused curriculum days are portable across dates and formats.
Full refund, transfer and credit policyWhat should I know before enrolling?
Full refund for a cancelled live program before its first scheduled session begins, including individual and group purchases and an approved MasterMind season. No refund after the purchased program starts, including dissatisfaction or absence, except for a session the provider fails to deliver. Statutory rights are unaffected. Partial attendance consumes the day under the published attendance rules. Read the full terms before choosing your class.
Full refund, transfer and credit policyWhen do I start using Dots?
Intro teaches individual tasks, clear prompts and review. Intensive teaches you to create proactive agents, test their work and hand supported responsibilities to OpenAI Dots. MasterCourse adds a Factory of proactive AI agent teams with shared records and coordinated handoffs.
Can the work continue after I close my phone or computer?
A Dot has its own cloud computer and browser. When the assignment, files and tools are available there, it can work while your phone or computer is off. Always-on means available to continue the responsibility; it may wait for a schedule, input or approval and remains subject to access and usage limits. Work using local files or applications still needs that connected computer online with the ChatGPT app open.
How do I stay in control over time?
Give the Dot and its assignments recognizable names, keep approved context current, and agree what deserves a report. It can surface missing input, access or decisions. To pause or retire work, check active delegates and saved schedules separately; pausing the main conversation does not stop them all. Review access and inputs before resuming.
What material can I use for the cloud exercises?
Use approved project material and limit personal or customer data to what the task needs. Keep client-confidential records and employee information out of unapproved connections. If your role involves hiring or other decisions about people, use fictional examples and keep those decisions with the authorized human reviewer.