What would convince every team that a workflow improved?
Compare completion, team effort, rework and operating burden with a before-and-after worksheet that makes missing evidence visible.
The revised workflow is running. Sales completes its part sooner. Operations still has a queue. Finance is asking whether the new records can be trusted.
An operating leader needs a shared basis for deciding whether to expand the change. Each team’s experience belongs in that decision, alongside evidence about the customer outcome and the work required to produce it.
We recommend agreeing on a small before-and-after worksheet before the next trial. Use it to make the comparison inspectable: which cases were included, what happened to them, who did the work, and what remains unknown.
Define the comparison before collecting results
Choose one case type and one start and finish. For customer onboarding, the interval might run from a complete signed agreement to verified readiness for the agreed service. Keep the same boundaries before and after the change. Moving the starting line past a difficult intake step would hide that work.
Specify which cases qualify. Separate standard agreements from unusual commercial terms or missing information. Record every eligible case entering each observation period, including cases that fail or remain open. Report their status and age separately; a completed-case average alone can conceal the backlog.
Use comparable periods with the same operating-day pattern and enough follow-up for cases to mature. In a hypothetical trial, a team might compare two four-week intake periods and inspect each case fourteen days after entry. That is an example to adapt to the process, not a universal minimum. Record holidays, staffing changes, volume and case mix rather than assuming equal-length periods are equivalent.
Fill one worksheet together
The Institute for Healthcare Improvement distinguishes outcome, process and balancing measures. Balancing measures examine whether a change in one part of a system creates problems elsewhere. That distinction is useful when departments share a workflow. IHI: Establishing Measures.
Our worksheet below applies that idea to a business operating review. Fill the before and after columns from the same definitions and retain links to the underlying records. Leave missing evidence marked unavailable.
| Measure and recording rule | Before | After |
|---|---|---|
| Eligible cases: total entering the period, split by agreed case type | Count and mix: ___ | Count and mix: ___ |
| Completion: cases meeting the same finish criteria by the follow-up cutoff | Completed / eligible: ___ | Completed / eligible: ___ |
| Elapsed time: median and slowest cases, with open-case ages shown separately | Time and open cases: ___ | Time and open cases: ___ |
| Team effort: hands-on minutes for sales, operations and finance, including review and correction | Each team’s minutes: ___ | Each team’s minutes: ___ |
| Rework: cases reopened or corrected, with reason and owning team | Cases / eligible: ___ | Cases / eligible: ___ |
| Operating burden: support, exceptions and supervision beyond routine case work | Minutes and incidents: ___ | Minutes and incidents: ___ |
| Cost and coverage: actual service charges, setup recorded separately, missing records identified | Cost and gaps: ___ | Cost and gaps: ___ |
Name who gathers each row, who checks it and when the group will review it. Record total team minutes as well as minutes per eligible case so a change in volume is visible. If effort is sampled, disclose how cases were selected and avoid presenting the sample as a complete time record.
Follow the work that moved
Consider a hypothetical onboarding change that lets sales submit a shorter form while software prepares the operating record. Sales may spend less time on entry. Operations may spend more time resolving missing requirements, and finance may repeat checks because the supporting evidence is harder to find.
Use the worksheet to investigate that possibility. Ask each team to identify a routine case and an exception, then trace the minutes, corrections and waits across the same records. Check whether the customer reached readiness sooner and whether the receiving teams accepted the result. This example describes a review method; it claims no measured benefit.
Agree in advance what trade-offs are acceptable. A small increase in finance review might be justified by better completion quality. That choice needs an owner, evidence and explicit agreement from the affected team. A lower total should not hide an unmanageable burden concentrated in one department.
Make the next decision reproducible
A before-and-after comparison can guide the next trial without proving that the change caused every observed difference. Explain altered staffing, case mix and measurement coverage. If important records are missing, choose a narrower test or gather more evidence instead of filling the gap with a favorable assumption.
Close the review with a decision to expand, revise or hold, its supporting evidence and a date to check again. Leave the definitions and collection steps with the operating team so it can repeat the review without the original project team. That is capability transfer applied to measurement.
If the scope or evidence needed for that decision is unclear, explore Northbeam Solutions’ Rapid Assessment.
When the engineering rigor matters more than the demo.
If your AI program is stalling between prototype and production, a two-day Rapid Assessment is the fastest way to find out whether you have an engineering problem, a design problem, or a governance problem.
Bill Tennant is Founder & Principal of Northbeam Solutions. Northbeam Solutions embeds alongside business and technical teams, builds production-grade AI work with their people, and leaves behind the engineering rigor, governance, and capability that separates AI operational efficiency from AI experimentation.