01 · The situation
A covenant trips. Twenty-five days before anyone can price the options.
An auto-components borrower with an ₹85 Cr term loan files its half-year numbers. Debt service coverage comes in at 1.05 against a covenant of 1.25. The breach is real, the contract gives you a 30-day cure period, and the order book has actually recovered since the numbers were struck. The facility is F-3106, and the question is on the credit head's desk: waive, amend, secure, or exit — and what does each one cost the bank?
The financial spreads are automated. The early-warning dashboard flagged the borrower weeks ago. Portfolio exposure refreshes nightly. And the covenant breach that started a 30-day cure period has sat undecided for twenty-five of them, because credit, the relationship manager, legal and risk have each worked out a different answer, and none of them is the number the committee can vote on.
02 · Decision points
Four ways forward on F-3106 — each costs the bank something different.
On paper, a clean four-way choice. In practice you can't price it inside the cure period, because the facility agreement, the borrower's spreads, the security register and the consortium position are in four different places, and nobody has put a number on any of them yet.
03 · As-is
How it reaches committee today: 25 days, four people, no agreed number.
Half-year financials are uploaded and spread. Coverage tests at 1.05 against 1.25. The breach is real and the cure period starts running. The decision is now live, but nobody owns the number yet.
The credit analyst rebuilds the coverage ratio to check what drove it. The relationship manager canvasses the borrower and comes back with a recovered order book. Legal reads the facility agreement for the cure period and the cross-default. Risk pulls the sector and group exposure. Four people, four separate answers, no single agreed number.
The cure period runs down while the analysis is assembled. The borrower keeps drawing on the working capital line, so the uncovered exposure grows. The security valuation on file is fourteen months old, and nobody has modelled what an exit would actually realise.
It goes to committee on half-reconciled numbers, days before the cure period closes. The waiver carries, because it is the only option that needs no further work. Not because it was priced against the other three.
04 · Assisted
The analysis is built overnight. The committee votes this week.
Same facility, F-3106. The twenty-five days collapse, because the four separate manual jobs become one automatic one.
Catches the breach as the numbers land. The same financials are uploaded. The system is already reading the spreads, the covenant schedule in the facility agreement, the security register and the consortium position, so it catches the breach the hour it exists, instead of waiting for a monthly exception report.
Gives the decision an owner and a clock. It logs the breach as a live decision, starts the cure period counting, surfaces the cross-default clause and flags every other facility sitting on the same covenant terms, so the call has an owner and a deadline from hour one instead of drifting until the committee calendar comes round.
Does the four jobs as one. Instead of credit, the relationship manager, legal and risk each working a separate file, it pulls all four together by morning — what drove the ratio, what a reprice would earn, what the security would cover at current market, and what an enforcement would realistically recover — then shows what each of the four options costs the bank on one agreed set of numbers.
One screen, four costed options. The credit head opens one view, sees all four options costed on the same basis, and takes a recommendation to committee inside the first week of the cure period, not the last. The system did the assembly. The credit judgement is still the committee's.
05 · The options, costed
All four on one basis: exposure, time, and what has to hold true.
Illustrative figures · scenario data, not a live system or a real borrower
06 · The playbook
What the options were weighed against.
This is judgment your best people already carry. Written into the decision, it works on every breach — not just the ones they happen to see.
07 · The call
Amend & reprice — this week.
Reprice at 75 bps: coverage reset to 1.10, stepping back to 1.25 — + ₹1.4 Cr yield.
The order book has recovered — but two lenders quote 40 bps below the revised rate.
Committee note, amendment letter and covenant schedule drafted overnight.
Queued for the credit head and the committee — cure period closes in five days.
Amend & reprice
08 · After the call
Take a path and the whole workstream is already staged.
Credit committee note
The recommendation, all four options costed on the same basis, and the reasoning, laid out in the committee's own format.
Amendment & waiver letter
Drafted against the covenant clause in the facility agreement, dated, referencing the test period that failed.
Revised covenant schedule
Coverage reset to 1.10 for two test periods, stepping back to 1.25, priced at 75 bps over the existing margin.
Security & consortium position
Charges, guarantees and the consortium's own testing dates pulled together. The valuation on file is flagged as fourteen months old.
Routing & cure deadline
Queued for the credit head to settle and the committee to approve. Cure period closes in five days, flagged.
What happened next
The decision leaves the room.
Credit head resets coverage to 1.10 for three test periods, not two — the order book is confirmed through March; repricing unchanged at 75 bps, committee note updated with the order-book support.
Credit head adds the promoter guarantee to the amendment and orders a fresh valuation before it goes to committee.
Credit committee approves. The amendment and waiver letter goes to the borrower; F-3106 recommended as amend and reprice.
Saved to the bank knowledge base — what was decided, and why. Learned: coverage breach on a recovered order book; repriced 75 bps, covenant stepped back over three periods.
Illustrative follow-through · scenario data
Coverage breach on recovered order book; repriced 75 bps, covenant stepped back over three periods.
One-off breach from a receivables timing shift; cover intact, waived for a single test period.
Third consecutive breach, security eroded, promoter unresponsive; accelerated inside the cure period.
Every breach you close teaches the system. The next one starts from what your bank already knows, not from whoever happens to remember the last one.
09 · As-is vs assisted
Same breach, F-3106 — two very different credit functions.
| As-is manual · today | Assisted with Decisome | |
|---|---|---|
| Time to a recommendation | ~25 days of a 30-day cure period | By the next morning |
| Basis for the call | Four rival spreadsheets, no agreed number | One agreed set of numbers |
| Options priced | The waiver, because it needed no work | All four, costed on the same basis |
| Security position | Whatever valuation is on the file | Revalued at current market before you vote |
| Cure period | Runs down while the analysis is assembled | Counted from the hour the numbers land |
| Exposure while deciding | The borrower keeps drawing, unnoticed | Drawdowns flagged the moment the breach is live |
| Portfolio view | One facility at a time | Every facility on the same covenant terms, surfaced with it |
| Precedent used | Whoever remembers the last similar breach | Every breach the bank has resolved |
| Once the committee decides | Documentation built by hand, later | Staged, routed, ready to execute |
| Who decides | A committee, late, on partial facts | A committee, on time, fully informed |
F-3106 is an illustrative example built from how covenant breaches are commonly handled on mid-market term lending, not a client result. The twenty-five days, the four workstreams and the figures on this page describe the shape of the problem.
Is this your decision — or close to it?