What may an agent do alone?
Four settings, from never to act alone. Every product ships them under a different name. Where you set each one, and what each one risks.
The most immersive way to learn AI agents in half a day. Two live faculty sessions, and a case personalized to your own organization and role.
Two live sessions with Dirk Hofmann. Between them, two hours inside a case built from your organization, where you decide, defend and get challenged.
Cannot make a live session? Both are recorded, and your case stays open afterward.
Paski is the AI tutor built for Harvard Data Science Review courses. Your prework answers give it the shape of your week and your context, and it pulls them into the simulation. Then it stays beside you for the two hours: walking you through each decision, playing the boss or stakeholder who pushes back, and giving feedback on the position you write. Dirk sets the frame. Paski personalizes it. The decisions stay yours.
Each one plays out in your case. Dirk opens on them, Paski works them through with you, and the debrief shows how the room handled them.
Four settings, from never to act alone. Every product ships them under a different name. Where you set each one, and what each one risks.
An agent acts in someone's name, under someone's logins. Whose, who can switch it off, and what the log has to show.
What the agent takes, what it hands back, and who now reads the log instead of doing the job.
A three-week course teaches everyone the same thing. This teaches you what agents mean for your role, in your organization, with faculty in the room.
Dirk Hofmann, Harvard Data Science Review author, opens and closes the morning in person. Not a recording of a lecture.
Your prework answers shape the simulation before you log on. No time spent on someone else's company.
Leaders working through the same decisions at the same time. The showcase opens three or four of their cases, and the conversation continues afterward.
%20(7).png?width=1400&height=2000&name=HDSI%20AI%20Agents%20AHlf%20Day%20Certificate%20(1400%20x%202000%20px)%20(7).png)
Issued on completion of the morning and your one-page position. Shareable on LinkedIn.
Past participants in Harvard Data Science Review courses include leaders from Nike, Apple, PwC, AWS and Google.

Harvard Data Science Review author · DAIN Studios
Author of the A.G.E.N.T framework, published in the Harvard Data Science Review. He opens the morning and leads the debrief with André Grobler.
.gif?width=52&height=52&name=Faculty%20AI%20GIFS%20%20(10).gif)
Dirk's faculty avatar
Briefs you the evening before on the dynamics, and sets the scene for your case on the morning. Trained to teach the way Dirk does.
.png?width=52&height=52&name=Paski%20(4).png)
Your AI tutor
Personalizes the case, the practice and the feedback to you. Walks you through each decision, plays the colleague who pushes back and the board you defend to, and reviews your one page.
The morning is two live sessions with a Harvard Data Science Review author and two hours inside a simulation built from your own context, where you decide, get pushed back on by your boss or a stakeholder, and write a position you can stand behind. It is built to cover in one morning what a lecture course spreads over weeks, because none of it is about someone else's company.
Both live sessions are recorded and your case stays open afterward. The morning is best live, and still complete if you miss a session.
No. This is for leaders and managers deciding what agents may do in their organization. No coding, no setup. If you can describe your role and your organization, your case can be written.
A Certificate of Completion from Harvard Data Science Initiative, your one-page position on agents for your organization, recordings of both live sessions, and your case to return to afterward.
Full refund up to 24 hours before the live morning. Email enrollments@hdsrcourses.org.
Tuesday, November 10, 9am–12:35pm ET. Registration closes Monday, November 9, 5pm ET, when prework is due.