One adjustment.
A different covenant result.

The fictional reporting group is measured on 30 September 2026, using four discrete quarters from October 2025 to September 2026. All amounts are US dollars. The $1.2 million restructuring expense has already been deducted from reported earnings.

The review began with the adjustment included. Challenging it stopped the determination. A conditional preview tested inclusion without changing or approving the record. The reviewer then excluded the adjustment, approved the new version and calculated the 4.00× breach.

A correctly identified breach is a valid result. The outcome, verification, approval and version used remain separate. Resolving a question does not approve the calculation; a preview never grants approval.

Debt of $18 million less $2 million of nettable cash gives $16 million net debt. Base EBITDA of $4 million plus the capped $1 million adjustment gives $5 million adjusted EBITDA. Dividing $16 million by $5 million gives 3.20×.

A displayed ratio of 3.50× can be within or outside the limit. With $5 million adjusted EBITDA and $2 million nettable cash, debt of $19,500,000 meets the limit; one cent more breaches it. The exact amounts decide the result.

Additional checks cover restricted cash, the cash cap, negative earnings, duplicate expenses, missing versus zero amounts, and large monetary values. Amounts remain exact to the cent from source to result.

The demonstration kept review, approval and calculation separate. Changing the treatment required fresh approval; uncertainty stopped a determination.

The completed review provides evidence for your existing sign-off process. Approval here authorizes a calculation; it does not issue a lender certificate or replace the signatory’s review.

The assistant helps inspect and challenge the evidence. Reproducing the arithmetic does not establish source authenticity, agreement interpretation or the approver’s authority.

The demonstration rejected outdated approvals, incomplete review packages, attempts to approve previews and final calculations without approval. Earlier source records remained available after a later version was approved.

On 25 October 2026, the September calculation and review workflow passed in a development environment. The recorded calculations and saved evidence passed their checks. Reviewer actions were simulated to test approval controls; they were not independent financial or legal approval.

The public extract contains synthetic financial facts, recorded results and worked examples at the covenant limit. The extract supports inspection of the financial workings. Private account details are excluded. It is not the complete original decision record and cannot establish the original approvals or reproduce the full application run.

The demonstrated scope is a fictional covenant calculation and review process. Source amounts are fixed; supported edits change adjustment treatment. A customer agreement and source mapping need their own qualification. The December planning example adds a forecast and permitted choices; it does not change the September result.

Covenant verification comes first. The optional planning example below builds on reviewed covenant constraints, with its own forecast, permissions and approval. ABL is a complementary review, separate from planning. The separate working-capital study uses a different fictional company. Their benefits cannot be combined.

This December projection builds on the reviewed covenant rules, adding its own forecast, payment and funding permissions, and separate approval. Planning is a separately scoped evaluation; this example does not establish a customer’s position or deployment readiness.

A $3 million invoice can be settled for $2.9 million on 1 October, $2.95 million on 30 November, or $3 million on 31 December. A $3 million collection arrives on 30 November. The choice is when to pay and whether to use the quoted $2 million loan.

The $50,000 is a modeled discount, not total profit or realized savings. Existing interest is paid in either plan. The selected plan gives up $10,000 of net benefit relative to the October choice to meet the covenant. Borrowing principal is never counted as a benefit.

Paying in October without borrowing fails the cash floor. Waiting until December without borrowing also meets the limits, but loses the $50,000 November discount. All three borrowing alternatives breach the covenant. The comparison covers all six permitted payment/borrowing combinations.

The public extract contains only synthetic financial assumptions, schedules and results. Recorded outcomes and independently worked reference cases are distinguished. It cannot establish the original approvals or reproduce the complete application run.

The completed local demonstration reviewed and approved the assumptions, compared the baseline with the selected plan, then challenged the Q4 forecast. With that fact unresolved, the answer became cannot determine. A conditional preview did not change or approve the planning record.

Resolving the forecast required a fresh review and approval. Changing the underlying covenant treatment made the earlier planning approval stale. With the adjustment excluded, none of the permitted plans met the covenant; the result was no feasible plan within the agreed choices, not permission to relax the limit.

The saved comparison and its original reviewed assumptions remained available after later changes. Approval authorized analysis only; the demonstration made no payments and drew no loan.

Demonstrated locally on 25 October 2026. The scripted workflow exercised assumption review, approval, plan comparison and later evidence inspection. Reviewer actions were simulated for the demonstration. This is evidence for the fictional example, not independent agreement review, a customer deployment or a turnaround guarantee.