A customer comes back to the helpdesk for the third time this week. And for the third time, the robot greets her like a stranger. She explains her problem all over again.
The robot is not being rude. When a conversation ends, it keeps nothing. Each new chat starts from a blank page.
The Notebook
So the team gives it a notebook. After each chat, it writes down what matters. Before the next one, it reads that note back.
That notebook is called agent memory: notes saved outside the model and read back in at the start of the next conversation. Now the customer returns, and the robot picks up where they left off.
Notice where the memory lives. It is not inside the model. The notes are saved outside it, and what the model is handed at the start of the next conversation is what lets it pick up where they left off.
Three Questions a Notebook Raises
A notebook solves the problem of forgetting, and it raises three questions. What gets written down? Who is allowed to read it? And what happens when a note is wrong?
The first decides what the assistant holds about a customer. The second decides who else can see it. The third matters because the notes are read back in at the start of the next conversation, whatever they say.
Why It Matters
The video ends on one question to ask: what does our assistant remember, and who checks it?
The model does not remember you. The notebook does. So the quality of the memory depends on the notebook: what it records, who can read it, and who checks it.
Agent Memory is an example of Context Engineering. Agent memory reads saved notes back before the next chat, deciding what the model sees about this customer.
Agent Memory contrasts with Parameters. The model keeps nothing between chats; agent memory is a notebook outside the model, not knowledge stored in its parameters.
Data Flywheel(lesson unlocks soon) contrasts with Agent Memory. Agent memory keeps notes on one customer for the next chat; a flywheel feeds everyone’s corrections back to improve the product.
Preference Gap(lesson unlocks soon) builds on Agent Memory. Only a better picture of the person closes the gap, and agent memory is where such notes are kept between chats.
The Business Engineer framework maps where each part of the AI stack sits and who holds the leverage. These pills are short lessons from the same library.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
