AOM 2026

My slides from This AOM PDW presentation, Aug 1, 2026

Link to google slides

Custom GPT (ChatGPT) for running the Effectuation Discovery Lab simulation

I created a Custom GPT available to anyone with the URL for you to use / run later. Click on the link to start the Chat session.

Here is the markdown file for the Custom GPT for those who want to unpack it to learn. 

The markdown file (Effectuation_DIscovery_Lab.md) is the build specification — the whole recipe for the GPT. It's written about the GPT in third person ("The GPT guides students…") and it covers every field you fill in when you create a Custom GPT: the Name, the Description, the Knowledge-file handling, the tone/"mild surprise" opening, the four principles, the behavioral rules, the full six-stage arc, and the suggested conversation starters. A professor could rebuild the entire tool from this file alone. 

Build your own custom copy (5 minutes) - 

1. In ChatGPT, go to Explore GPTs → Create → Configure.

2. Name and Description fields: copy those sections from the markdown file I've linked above.

3. Instructions box: paste everything from Behavioral Rules through the end of Stage 6: Commit. (This is the engine — it's the one section that makes the simulation run.)

4. Conversation starters: copy the four listed at the bottom of the file.

5. Knowledge: upload the supporting PDFs.

6. Save, then test by typing "Start the Effectuation Discovery Lab."

(Build your own custom copy detailed instructions)

More Takehome files (In addition to Prompt Files Above)

1. Toward end of session: The Experiential Prompt Arc: Design Worksheet a clean worksheet document for you and your students to complete.

2. Custom GPT for Todd Grave's Raising Cane's effectuation simulation (interactive)

3. Custom GPT for Todd Grave's Raising Cane's effectuation simulation (non-interactive, hypothetical session transcript)

To use GPTs in your classroom

Method 1: The "One-Click Clone" (No Uploads Required)

I have built this as a Custom GPT inside ChatGPT (using a Plus/Team/Enterprise account) to handle all the heavy lifting on my end.

1. My Action: I created the Custom GPT. I pasted my 6-Stage Prompt Arc into the instructions, and I physically uploaded the core PDFs (Sarasvathy, 2001; Nicol & Macfarlane-Dick, 2006) directly into the Custom GPT’s "Knowledge" section.

2. I Published It: I set the visibility of your Custom GPT to "Public" or "Anyone with the link."

How the files "travel": In this scenario, the files are completely integrated into the hosted AI tool. The other professor does not upload anything. The prompt instructions, the behavior boundaries, and the background PDF files are all bundled together in the cloud under your link. When their students chat with that specific link, the AI references the PDFs already uploaded to Craig Armstrong's account and incorporated into the published custom GPT.

Method 2: The "Open-Source Kit" (Manual Configuration)

If another professor wants to completely customize the tool—for example, if they want to change my entrepreneurship scenario to an Organizational Behavior focus, or if their university forces them to use an isolated, secure campus AI playground (like a private Azure OpenAI environment)—they cannot use my public link. They must build their own version from scratch.

This is where your physical Takeaway Handouts come into play. To clone your pedagogy manually, they will execute a two-step setup:

Key words: Academy of Management, effectuation discovery lab, effectuation, experiential prompt, worksheet, design, slides, Custom GPT, markdown code, adaptable six-step learning arc, GPT code, open source