The useful habit, when a restrictive covenant lands on your desk with a signature already on it, is to stop reading it as an instruction and start reading it as a document that has to survive a challenge. Canadian courts approach a post-employment covenant as presumptively unenforceable, a restraint of trade that the employer must justify. That reversal matters. The question is not whether you can find a reading that makes the clause sensible, but whether the employer can show the clause it actually wrote is unambiguous, reasonable between the parties, and no wider than the interest it protects.
Start with the words that define the restricted activity
Ambiguity is the first and quietest failure point, because a clause can look perfectly firm and still be unenforceable for the simple reason that nobody can say what it covers. In Shafron v KRG Insurance Brokers, decided by the Supreme Court of Canada in 2009, the restricted area was the "Metropolitan City of Vancouver," a phrase with no legal or settled commercial meaning. The Court held that an ambiguous covenant is, for that reason, unreasonable and unenforceable. A careful reader looks for the same defect elsewhere: undefined terms like "the business," "similar employment," "any capacity," or a customer list that includes people you never met.
Check the geography against the territory you actually worked
Geographic scope is tested against the employer's real protectable interest, not its ambitions. A clause covering the province when you sold into three counties is vulnerable, and a clause with no geographic limit at all has to justify that absence through something else, usually a narrow definition of prohibited activity. Ontario's Court of Appeal struck down a covenant in H.L. Staebler Company v Allan in 2008 where the restriction on doing business with clients carried no territorial boundary and no limit on the kind of work restrained. Read the map against the words. If the two do not match, note the gap.
Ask what the duration is measured against
Two years is not automatically too long, and six months is not automatically safe. Duration is judged against how long the employer's advantage actually lasts: the renewal cycle of the accounts, the shelf life of pricing information, the time it takes a replacement to build the relationship. A two-year restriction on an insurance producer whose policies renew annually reads differently from a two-year restriction on someone who sold a commodity on thirty-day terms. The other question is comparative. If a non-solicitation clause would fully protect the interest, a court following Elsley v J.G. Collins Insurance Agencies will ask why the broader non-competition clause was needed at all.
Understand what severance can and cannot do
This is where most expectations break. Blue-pencil severance permits a court to strike out a discrete, trivial part of a clause, the sort of deletion that leaves the remaining words standing on their own without alteration of meaning. It does not permit a court to substitute a smaller radius, shorten a term, or read down a definition. Notional severance, which reads a provision down to a legal maximum, was expressly rejected for employment restrictive covenants in Shafron, on the ground that it would hand employers a free option: draft wide, and let the court supply the reasonable version later. So the clause you signed is the only clause on offer. Courts will not improve it for the party that wrote it.
What to do with what you find
Write down, in your own words, the three answers: what activity is barred, where, and for how long. Where you cannot answer without guessing, mark it, because that guess is the ambiguity argument. Then collect the facts that bound the employer's interest: your territory, your account list, the renewal cycle, whether you held pricing or only relationships. An hour with an employment attorney spent on that summary produces a usable read on enforceability; the same hour spent reading the contract cold mostly produces a slower read of the contract. Enforcement policy for non-competes is a live subject at the Federal Trade Commission in the United States, and the drafting problems it examines are the same ones Canadian courts have been resolving for decades.
The covenant that survives is usually the narrow one, drafted by someone who knew exactly which customers and which months mattered. That is worth knowing before you negotiate a departure, because the strength of your position is set by words already on the page, and those words can be read carefully in an afternoon.