Guide 4 min read

Limitation-period math is a date problem, not a law problem

Knowing the limitation rule is the easy part. Applying it to a client's account requires a trigger date, a counting convention and a set of tolling conditions. Here is how Interview does the arithmetic, and where it stops.

2025 · 05 · 22·admin

Every lawyer knows the limitation period for a straightforward contract claim in their own jurisdiction. Very few can tell you, from memory and without a calendar, the last day on which proceedings may be issued for a breach that occurred on 29 February 2024, in a jurisdiction that counts from the day after the breach, excludes the last day if it falls on a public holiday, and tolls the period for any interval in which the defendant was outside the jurisdiction.

That is the point of this guide. The law is known. The arithmetic is where things go wrong, and the arithmetic is what AI Interview does.

Four inputs, one output

Interview’s limitation calculation takes four inputs and produces a candidate date with the working shown.

1. The rule. Period length and the event that starts it, for the matter type and jurisdiction. This is a lookup in a table we maintain per jurisdiction, with a last-verified date on each entry. The table is reviewed by qualified lawyers in each jurisdiction on a schedule, and the brief shows the verification date so that a lawyer can decide how much to trust it.

2. The trigger date. The date of the event that starts the period. In a simple contract claim, the breach. In a negligence claim, the damage, or the date the client knew or ought to have known of it, depending on the jurisdiction. Interview extracts candidate trigger dates from the client’s account, normalises them (“the week after Easter last year” becomes a date range) and shows the lawyer which date it used and why.

3. The counting convention. Does the period start on the day of the event or the day after? Does the last day count? What happens if the last day falls on a weekend or public holiday in the court’s location? These rules differ between jurisdictions and sometimes between courts in the same jurisdiction. Interview applies the convention for the court the matter would most likely be brought in, and says which one it assumed.

4. Tolling and extension conditions. Minority, incapacity, the defendant’s absence, acknowledgement of the debt, fraud or concealment, standstill agreements. Interview asks about the conditions that apply in the identified jurisdiction and, where the client’s answer engages one, shows the adjusted date alongside the unadjusted one.

The output is a date, or a small set of dates where the jurisdiction or the trigger is uncertain, each with the chain of inputs that produced it.

Where the errors come from

Across the preview period and the first weeks of general availability, we have reviewed every case where a lawyer corrected the computed date. The corrections fall into a short list of categories, in descending order of frequency.

  • The trigger date was wrong, because the client gave an approximate date and Interview resolved it to the wrong end of the range. We now default to the earlier end, which produces the more conservative deadline, and show the range.
  • The wrong matter type was inferred. A client describing a defective product may have a contract claim, a tort claim and a statutory claim, each with its own period. Interview now computes all three where more than one is plausible and lists them rather than picking.
  • The jurisdiction was ambiguous. A cross-border contract, a client who has moved, an employer in one country and a workplace in another. Interview lists candidates and computes for each.
  • A tolling condition applied and was not asked about, because the client’s account did not surface it. This is the hardest category. Interview’s question set for each jurisdiction includes the common tolling conditions, but a client who does not know that the defendant spent eight months abroad cannot tell us.
  • The table was out of date. This has happened twice. Both entries were corrected within a day and the affected briefs were re-issued with a notice. The verification date on each entry exists so that lawyers can apply their own judgement about staleness.

What has not appeared in the corrections, so far, is a counting error. The arithmetic, once the inputs are fixed, is the reliable part.

A checklist for the lawyer reading the brief

The brief is a starting point, not a conclusion. We suggest the following, in order, and it takes less than five minutes.

  1. Check the trigger date against the client’s account. It is the first thing shown and the most common source of error. If the client gave a range, note which end was used.
  2. Check the matter type. If the brief lists more than one, decide which claims you would actually bring and discard the rest.
  3. Check the jurisdiction and the court. The counting convention follows from this.
  4. Read the tolling questions and answers. If a condition is marked “not asked” because the client’s account did not reach it, ask it yourself.
  5. Look at the table verification date. If it is older than you are comfortable with for that jurisdiction, verify the rule independently.
  6. Diarise the earliest candidate date, not the most likely one, until the uncertainty is resolved.

What Interview will not do

It will not tell the client their deadline. The client is told that the firm will advise on timing. A computed date shown to a client without a lawyer’s review is a liability, not a service.

It will not pick a jurisdiction, a matter type or a trigger date where the facts leave more than one open. It lists them. Choosing is legal judgement.

It will not treat a date as certain because the client was certain. Clients are confident about dates and frequently wrong, and the brief says “client stated” rather than “on”.

The rule is in the book. The deadline is in the calendar. Interview gets the calendar right so that you can spend your five minutes on the book.

See it on a contract you have already reviewed.

Send us a draft your team has already redlined and we will show you what ZAAN catches, and what it misses.