Implement is where the redesign installs alongside the running workflow, not in place of it. Two paths: a Workflow Leverage Sprint installs one workflow in two to six weeks. A 100-Day Operating Leverage Program installs two to four workflows as one operating system. The Audit chooses which.
Choice is set by the roadmap, by your principal's bandwidth, and by which workflows are coupled in cash impact. Not a menu the buyer picks from cold.
One workflow. Two to six weeks. A redesign installed, not just documented. Best for contained workflows where the cash impact and the owner are clear: bid-to-award handoff, client onboarding cycle, renewal motion, inventory reconciliation, or billing exception cycle. Pays back inside the engagement window when the workflow is the right one. If it is the wrong one, the Sprint does not pay back and we say so in the scoping conversation before contracting.
Two to four workflows. One hundred days. Executive sponsorship from the principal. Not a Sprint times three. The cross-workflow discipline is different. Best for operators whose Audit roadmap names more than one workflow and whose principal is committed to running the redesign as a portfolio.
The Program picks up the Audit's recommendation memo and turns it into a working operating change across two to four workflows. The first phase looks heavy because Program coordinates the design across multiple workflows. Where a Sprint freezes design on one workflow in two weeks, a Program freezes design across two to four workflows in a month. One Vaival senior is named and accountable from day one. No deck-and-disappear.
Lock scope across the two to four workflows in view, name owners on both sides, and freeze the design before any system or process work begins. Establish what "good" looks like per workflow in measurable terms: cycle time, error rate, cash trapped, or whatever the Audit roadmap named. By day 30 you have a written design memo, signed off by your executive: current state, target state, what changes, what does not, and how the workflows sequence against each other. This is the moment we tell you what we will not do, before we start spending your budget.
Process changes go live in stages. System changes are scoped, configured, integrated, and exercised against the workflow before they touch production. AI-assisted tasks pass the AI Go-Live Gate before they make a decision your customers, employees, or regulators will feel. Your team is in the room throughout. By day 70 the new operating workflow runs on a defined slice of the business with old-state and new-state operating in parallel.
The last thirty days cut the workflow over fully to the new operating state, measure it against the day-30 design memo, and hand the workflow back to a named owner on your side. At day 100 you own the workflow. Vaival closes with a written report: what changed, what the numbers say, what we got wrong, and what to do next. We are available for a Managed Operating Pod follow-on if you want it, and out of the way if you do not.
The redesign installs in parallel with the running workflow until one full cycle has matched the measurement baseline. Rollback to the prior cadence is documented before the first new-cadence run.
Same workflow, two tracks, for one full cycle. Measurement baseline gets compared cycle-over-cycle. No cutover while a live job is mid-milestone.
Cutover only after the new cadence matches or beats the baseline. Approval rule documented in writing. AI-assisted tasks have human owners with audit trails before any model is deployed.
If the new cadence holds under live use, the old one retires with documented sunset. If it does not, rollback runs against the documented prior state. No "ship it anyway" pressure.
No exceptions. Tasks that cannot meet the gate are deferred or descoped per the engagement letter change-control clause, not shipped anyway.
Named human on your side with decision authority over the workflow.
Documented edges of the workflow the AI-assisted task sits inside.
Pre-AI measurement of the workflow so the change can be evaluated.
Authoritative system for the data the model uses. No ambiguity.
Who approves outputs, when, and against what threshold.
What happens when the model is wrong or uncertain. Documented escalation.
Sampling rate, review cadence, and acceptance criteria for AI-assisted output.
Weekly or biweekly review where AI-assisted work is inspected against baseline.
The number the AI-assisted task is supposed to move. Tied to the baseline.
Each of these is the right answer for some buyer. We name when, and we name what they cannot give you that Implement can.
Strong on methodology and the partner who flies in monthly. The operator-grade install is not theirs to run. The methodology will be sound. The cutover-without-cutting-the-live-job is unlikely to be inside the scope.
Forty hours a week, one head. Enough to walk one workflow at a slow pace. Not enough to install a cross-workflow redesign on a hundred-day clock or to carry the AI Go-Live Gate discipline.
Sometimes the right call. Stalls most often on sequencing, not capability: the operator who would redesign the workflow is the same operator already running the work the workflow is hurting.
If the fit is there, we name the scope and kickoff inside forty-eight hours. If your workflow is too small for a Sprint to pay back, or if the right next step is the Audit before an Implement engagement, we say so on the call.