Legal IntelligenceCareersReview capacity

Careers / Review capacity

Faster drafts still need
somebody’s review time.

Estimate the work arriving at review, including exceptions. Then compare it with the hours actually protected for that work.

An editorial planning model with invented defaults. It does not establish safe staffing, adequate professional review, productivity gains or return on investment.

01 / Make the assumptions visible

One workflow. One week.

Use non-sensitive aggregate estimates. Fractional items are allowed for an expected weekly average. Nothing is submitted or stored.

Count outputs that actually reach this review step.

Applied to every item, including items with an exception.

Between 0 and 100. This is an assumption unless observed.

Extra effort beyond the first review; do not count it twice.

Time available for this workflow, after other commitments.

02 / Compare effort with time

Illustrative weekly workload

Estimated review load800 min

13 h 20 min

Review load800 min
Protected time600 min

Planning gap

200 min

3 h 20 min beyond the stated capacity.

First reviews600 min

Additional exception work200 min

Expected exception items10 per week

Bars share a scale in minutes. No utilization or staffing ratio is calculated. Display values are rounded; very small positive values are labeled rather than shown as zero.

The arithmetic

Items × (first-review minutes + exception rate × additional minutes)

40 × (15 + 0.25 × 20) = 800 minutes

Protected hours × 60 = available minutes.

10 × 60 = 600 minutes

A gap invites a review of scope, inputs, exception patterns and protected time. Reducing an input until the model balances does not prove that less review is adequate.

03 / Keep an assumption record

Copy the scenario, not a claim of efficiency.

The record preserves inputs, the calculation and any zero-review warning. It does not collect names, matter details or evidence files.

If clipboard access is unavailable, the record will be selected for manual copying. Reloading restores the illustrative defaults.

Model boundaries

An average cannot approve a workflow.

The model applies one average first-review time to every item and one average additional effort to the stated share of exception items. It does not model bursts of arrivals, queues, urgency, review quality, qualifications, multiple rounds of correction, training or other work unless their time is included in the inputs. Expected exception counts may be fractional; they are averages, not a predicted count for a particular week.

Zero items means no modeled demand. Zero capacity does not create an infinite ratio; it means no protected time was supplied. Zero first-review minutes with a positive workload triggers an incomplete-model warning. None of these states approves release or staffing.

The ABA’s Formal Opinion 512 (29 July 2024) discusses competence and supervision in generative AI use. The SRA’s Misuse of AI warning notice (17 August 2026) addresses accountability, supervision and controls for those it regulates. Their jurisdictions and authority differ. Neither supplies the defaults, validates this calculator or prescribes a universal review ratio.

Working on a team process?

If you need help connecting review capacity, ownership and handoffs, describe the team workflow in the AI Vortex operating brief. Paid systems work, subject to fit and scope.