Point of view

Good technology decisions start before the technology.

Short perspectives on the operating choices underneath AI adoption, enterprise knowledge, and delivery.

AI adoption must give time back to the team

If training AI feels like extra work, adoption stalls.

Most AI rollouts ask people to do something new on top of their existing work: write prompts, maintain documentation, report usage. Teams reasonably resist. Some also worry that teaching AI what they know is the first step toward replacing them.

The better model makes context a byproduct of good delivery. Approved stories, review corrections, tests, and retrospective decisions become reusable guidance only after a person validates them. The benefit returns to the team: less repeated explanation, fewer recurring mistakes, faster onboarding. Measure the result against the team’s own baseline—and never redefine the measure to flatter the tool.

Team memory and enterprise knowledge have different jobs

One giant knowledge base serves no one well.

Team context belongs close to the code and the people who use it every day: implementation patterns, local domain rules, and lessons from review. It is living and specific, and it helps people and agents make a correct change in that environment.

Enterprise knowledge answers different questions: which platform already supports a capability, what a regulatory change would touch, what onboarding a client requires. Those answers span systems, policies, and business rules that source code cannot explain. Keep the two layers distinct, connect them deliberately, and agents can work from the right context for the decision at hand.

Distributed delivery needs accountable leadership

Capacity without ownership is just more people.

Adding experienced people can move a modernization or data program forward quickly. It can also create a second organization nobody fully owns. The difference is senior accountability: someone responsible for onboarding, quality expectations, escalation, and fit with the client’s operating model—not only at placement, but throughout the work.

In healthcare, domain knowledge matters as much as technical skill. A team that already understands claims, pricing, and regulated data starts closer to useful work. Access boundaries, PHI requirements, and vendor approval should be settled before the team arrives, not discovered after.