AI belongs inside the business, not beside it.

I first learned a lot about that discipline early in my career at EDS, based in New Jersey and Pennsylvania and working in a classic mainframe environment supporting AT&T.

Great foundations were established for me there.

The Systems Life Cycle was taken seriously: information planning, definition, analysis, business and technical design, construction, testing, implementation and production support.

Releases were formal. Batch cycles ran overnight. Requirements were gathered carefully. We looked for patterns, worked out which features belonged together, and considered what should be developed and when.

There was unit testing and system testing. Quality assurance was not something bolted on at the end.

The technology looked very different. Green screens. Reports that often had to make sense on paper as well as on screen. GUI wasn’t yet part of the vocabulary.

But there was a real appreciation for the data and for the user.

And there was true teamwork.

As client-server and distributed processing arrived, the technology, pace and possibilities changed.

Some fundamentals did not.

I have watched that pattern repeat through successive generations of technology and, for more than 25 years, from another vantage point: working with leadership and management teams.

Starting things is rarely the problem.

Knowing what to stop can be.

So can deciding what to pause, what to accelerate and what needs reshaping.

Common to say. Uncommon to do.

The tendency is to keep adding more to the agenda and wait for other things to fall off accidentally.

Something that once required months can now become tangible remarkably quickly. A team can demonstrate an idea before the management conversation around it has really begun.

That is one reason I think AI belongs inside the business rather than beside it.

I would start with the work.

Look at the workflows. Bring the people doing the work into the conversation.

Where is the friction? Where are the hand-offs? Where is information repeatedly assembled or reworked? Where does human judgement matter? Where would better insight change an outcome?

Then look across those workflows.

There will be nominees for attention. There will also be patterns among the nominees, and patterns among the obstacles.

I sometimes describe what management needs as a heads-up display. Another way of saying it is: see the whole landscape.

Enough to know what is moving, what is stuck, what is competing for attention and where something interesting may be happening.

Take it element by element, workflow by workflow.

The Fundamental Pyramid® is useful to me for this reason. It enables a leadership and management team to see the business as an enterprise and agree or align on where attention is needed.

The Pyramid is not the point.

The management habit is.

See enough of the landscape to make deliberate choices.

Then move.

The disciplines I first encountered at EDS have not become obsolete because the technology has become faster.

The requirement still matters. The data still matters. The user still matters. Testing and readiness still matter. Teamwork still matters.

They do not need all of the old machinery around them.

But the responsibility remains.

AI belongs inside the business because that is where the work is.

That is where the opportunities are found, the trade-offs are made, and management decides what moves forward.