Limitation period checks: why you should not rely on a chatbot
By the ChamberOne team 5 minute read
Few errors in legal practice are as final as a missed limitation period. The claim may have been strong. The client may have been blameless. If the time has run, the case is gone, and the firm may face a claim of its own. So when a chatbot offers a confident answer to 'how long do I have to sue?', the right reaction is suspicion, however neat the reply.
Key points
- Limitation depends on claim, place, start date and exceptions, so one-line chatbot answers are unsafe.
- Take each period from the primary source for the right place and diary the most cautious date.
- Require a second lawyer to calculate the date independently from the source.
- Check the other time limits too: notices, contract bars, tribunals and appeals.
Why limitation is hard to get right
Limitation is not one number. The time allowed depends on the type of claim, the law of the place, who the claimant is and when the claim is treated as arising. Different causes of action have different periods. Some claims have a separate long-stop. Claims against public bodies, claims involving children or people lacking capacity, claims for personal injury and claims by or against estates may follow special rules.
On top of that come questions about when time starts, whether it can be paused or extended, what counts as starting proceedings and whether an acknowledgment or a standstill agreement changes things. A one-line answer almost always hides several of these questions.
How a chatbot goes wrong
A general AI tool predicts plausible text. For limitation, plausible is not good enough. Common failure patterns include the following.
- Wrong place: giving the period from another country or state because the claim type sounds familiar.
- Stale rule: describing a period that has since been amended.
- Missing exceptions: stating the general rule but not the special rules for minors, disability, fraud, concealment or latent damage.
- Wrong start date: counting from the date of the event when the rule counts from knowledge, or the reverse.
- Confidence without a source: giving a number with no statutory reference, or one that does not exist.
- Ignoring other time limits: such as pre-action notices, ombudsman deadlines or contractual time bars, which may be much shorter.
Other deadlines hiding behind limitation
Statutory limitation is only one of the time limits that can end a claim. Contracts may set their own notice requirements. Public bodies may require notice of a claim within a short period. Regulators and ombudsman schemes often have short windows. Employment and tribunal claims commonly have much shorter periods than ordinary civil claims. Appeals have their own deadlines.
Check each of these for every matter, in the place where the claim will be brought. A tool that answers only about the general civil period misses most of them.
Build a rules-based diary instead
The safer approach is to treat limitation as a diary problem with human-verified inputs. The principle is that the rule comes from the primary source, the date is calculated by a person, and the diary is a system that cannot be skipped.
- 1
At intake, identify every possible cause of action and the forum for each.
- 2
For each, find the relevant period in the primary source for that place, and write down the reference.
- 3
Identify the start date, with your reasoning, and the date on which time would expire on the most cautious view.
- 4
Enter the cautious date in the firm's diary with at least two reminders well in advance, and a named responsible lawyer.
- 5
Diary a separate date by which proceedings must be issued internally, safely before expiry.
- 6
Record any facts that might extend or interrupt time, and revisit the calculation when new facts appear.
The second-lawyer check
Every limitation calculation should be checked by a second lawyer who works from the source, not from the first lawyer's note. They should reach their own date, and the two should compare. Disagreement is useful: it exposes a question that needs further thought. Record the check on the file with the date and the initials of both.
Make this check routine, not optional. Firms that reserve it for 'difficult' cases tend to find that the case which caused a problem looked simple at the time.
Where AI can help without being the authority
- Reading a client's documents and listing the dates of events, so that you can see when time may have started.
- Producing an issues list: the possible causes of action and parties, which you then research.
- Drafting a letter to the client explaining why a date matters, using your confirmed facts.
- Prompting you with questions you might have missed, such as whether the claimant was a minor at the time.
In each case the tool helps you think. The date and the rule come from the primary source and your own judgment.
- Lists the dates of events
- Produces an issues list
- Drafts a client letter from confirmed facts
- Prompts you with questions you might have missed
- The rule comes from the primary source
- The date comes from your own calculation
- A second lawyer checks it from the source
When time is short
If a client arrives close to an expiry date, treat the matter as urgent from the first minute. Issue protectively where appropriate rather than waiting for a full assessment. Consider seeking agreement from the other side, in writing, to extend time, but do not rely on an informal assurance. Tell the client plainly about the risk and record that you did. Check the guidance from your regulator or bar, and your insurer's requirements, on handling time-critical matters and on notifying problems.
A culture that protects clients
The best defence is cultural. Make it normal to ask 'what is the limitation position?' at the start of every dispute matter. Make it easy to ask a colleague to check. Make sure that no one is afraid to say 'I am not sure'. A firm in which people double-check each other's dates, and in which the diary is trusted, will rarely meet the disaster that a quick chatbot answer invites.
Training helps. Run a short session each year using anonymised near misses from the firm or the wider profession, and ask each lawyer to work out the limitation position from scratch on a sample claim, then compare answers. The differences are the lesson, and they show how easily two careful people can reach different dates without a second check.
This article is general information about the law, legal practice and technology, current when it was published. It is not legal advice. Laws and guidance change: check the sources linked above and take advice on your own circumstances from a lawyer qualified where you are.
Read next
AI in Witness Bundle and Proof of Evidence Preparation: What to Automate and What Not To
A practical guide for litigators on which parts of witness bundle and proof of evidence preparation AI can handle, and which parts it must not touch.
6 minute readHallucinated citations: what court sanctions teach about checking AI output
The patterns behind fake authorities reaching courts, and the supervision and checking controls that would have caught them.
5 minute readBuilding a litigation chronology with AI
Extract dated events, link each to its source page and check for conflicts before the chronology is used.
5 minute read