From strategy to sustained performance.

Most strategy work holds up in the boardroom. The analysis is sound, the business case adds up and the recommendation is approved. The program stalls later, once it reaches the people expected to run it.

We see three causes, in this order of frequency. First-line managers are left out of the design, so they are asked to deliver the new model while their own scorecard still measures the old one. The future-state model is designed without the people who will run it, and they resist a system they had no part in shaping. And the story moves from the current state to the future state without naming what is being lost; people grieve old structures even when those structures were broken.

Prosci’s ADKAR model (awareness, desire, knowledge, ability, reinforcement) describes the same discipline: skip a stage and the program lands as compliance. We build readiness into the design before the deck is written.

When results disappoint, leadership teams reach for a restructure. Boxes move and reporting lines change. Three quarters later performance looks the same, because the redesign changed structure and left accountability where it was.

An operating model has four layers. Structure sets who reports to whom. Accountability names who owns each outcome, which decisions they make and where issues escalate. Cadence is the set of governance routines that force the right conversation at the right level and interval. Capability is whether the people in each role can do the work the design assumes.

Accountability is where most programs underinvest. Named owners, named decisions and a regular cadence create it; a responsibility chart on the wall does not. AI programs follow the same pattern: a new tool changes the structure of the work, and adoption depends on the other three layers.

For decades, sales scorecards counted activity, weighted it by product and paid against a threshold. That worked because activity tracked output: the strongest sellers made more calls, opened more files and closed more business.

AI weakens that link. When preparation takes seconds and a model drafts part of each proposal, an average seller’s activity climbs with no matching gain in results. A plan that pays for activity then grows in cost without the revenue to fund it.

The plans that hold up move weight from activity to outcomes such as revenue, retention and the quality of advice. They measure productivity per hour, with the data to do it reliably. And they pay sellers who use AI to serve more customers well differently from sellers who use it to do the same work with less effort. Done well, the strongest sellers earn more and customers get more of their time.

Most organizations sell through some combination of a direct sales team, partners or intermediaries, and digital self-service. In many, the balance between them is the product of years of drift.

The channels differ in what each sale costs to win: partner fees, direct compensation and overhead, and digital marketing spend produce different unit economics for the same sale. In many markets they also differ in what happens next. Customers who arrive through intermediaries tend to shop harder at renewal, and direct teams more often build relationships that span several products.

Pricing, capacity, technology investment and compensation all follow from the mix. Organizations that set it on purpose make those decisions from one strategy. Those that let it drift pay for it in margin.