Insights
Notes from the work.
What we have learned building AI that a company can stand behind.
A business is an engine, not an application.
A great deal of what is sold as AI for business today promises a company that runs itself. The agents read the mail, answer the customer, pay the invoice, and post the update, and the owner looks on.
That is not a business. It is an application. Let me explain what I mean.
A business is an economic engine. It produces money, and it produces things money only measures. It gives people work they are proud of. It leaves customers better off. It keeps alive the small societies that form around it: the staff, the families, the vendors, and the other small businesses that depend on it.
I have run one for most of my working life. At the Institute of Reading Development, the place we make the biggest difference is the classroom, where a live teacher works with a child who is learning to read. Everything else in the company exists to put more good teachers in more classrooms, and to let them teach without compromise.
That is why I use AI, and it is why I use it the way I do. My own mail runs through a console that drafts every reply and cannot send one. I send.
The agents do not run the company. They carry the volume, so our people can carry the quality.
I started Structural GrowthWorks as a lab for that idea, and the Institute became its first in-depth case study. Every company moves value through three links: what it brings in, what it makes of it, and how it reaches the customer. In each link a few steps decide the quality the customer receives. The best place for a person is the place where a person changes that quality the most. Find those places, put your best people there, and build the agents to carry everything else.
A stronger company can do more of the work that matters, and do it without compromise. That is what the agents are for.