Establish what actually left
Separating code, documents, data and knowledge, because they are different problems with different answers.
Technology
Technology disputes.
Technology disputes usually begin as a suspicion rather than a discovery. A former engineer surfaces at a competitor with something familiar, a partner starts selling around you, or a product appears that resembles yours more than coincidence allows. What decides these is rarely outrage. It is what was written down before anybody fell out, and what can be established without guessing.
What this looks like when it goes wrong
The version that arrives most often involves someone leaving. An engineer, a founder or a whole team departs, and a product appears afterwards that is uncomfortably close. What matters is what actually went with them, because code, documents, customer information and knowledge carried in a person's head are four different problems with four different answers, and they are routinely treated as one.
The next is a partner going around you. A reseller starts contracting with your customers directly, an integration partner builds the thing it was integrating with, or a customer decides that rebuilding is cheaper than renewing and hires the people who know how. Each of those may be a breach and each may simply be commercial life, and the difference sits in the agreement rather than in how it feels.
Then there is a competitor that has copied the visible layer. The interface, the structure, the naming, the documentation, the onboarding. It is the most infuriating version and frequently the weakest, because the parts of software that are easiest to see are not the parts that are easiest to protect.
The last is failure inside a live relationship. An implementation that never worked, an outage that cost a customer money, a migration that lost something. Both sides have a contract, both sides have a long email trail, and the dispute escalates because somebody sent a letter in the tone they were feeling rather than the tone that would have ended it.
What actually decides it
What was signed before the fallout does most of the work. Employment terms, assignments of what people create, confidentiality obligations, and any restrictions on what a leaver may do next. Restrictions of that kind are only enforceable within limits, and whether a particular one holds turns on how it was drafted and on what the person has actually done. A restriction drafted to cover every eventuality is one of the commonest reasons a restriction achieves nothing at all, and that is a drafting question rather than a general rule.
The next question is what the right actually reaches. Copyright in software protects the way it was written rather than what it does, so a competitor achieving the same result by different means is on much safer ground than one holding your files. Confidence and trade secret protection often carry more of the weight in these disputes than copyright does, and they depend on the information having been treated as confidential in practice, not merely described that way in a policy.
Then there is what can be established rather than believed. Disputes of this kind are decided on records that existed before anybody was looking, and preserving what exists is the first practical step in any of them. The gap between what a business is confident happened and what it can demonstrate is where most of these matters are actually lost, and it is worth measuring that gap honestly before a position is taken publicly.
Finally there is what you want and what pursuing it costs beyond fees. Stopping something, being paid, keeping a customer and deterring the next person are different objectives and they point to different routes. A formal process is public, it obliges you to disclose your own material, and it consumes the attention of exactly the people the business needs building. A dispute with a partner or a customer can also cost more in the market than it recovers. The quiet routes remain available for a while and stop being available once a demand has gone out.
What we do
Preserve the position early
Making sure the records that decide it still exist before anybody is told they are being looked for.
Test what the rights reach
An honest assessment of what is protectable here and what is simply competition you dislike.
Covenants and confidentiality
What the documents say, whether they hold, and what can realistically be asked of a former employee.
Resolve it quietly first
Most of these end without a public fight, and the quiet route closes the moment a demand goes out.
Escalate properly when needed
Formal proceedings run with regulated partners where the matter genuinely requires them.
When to spend nothing
A former employee doing the same job somewhere else, having taken nothing, is not a dispute. Neither is a competitor whose product resembles yours because you are both solving the same problem for the same buyers. Acting on either is expensive, unlikely to reach anything, and hands the other side a story. The honest answer in a good number of these situations is to compete rather than to litigate.
There is also a reputational calculation people skip. A public fight against a smaller competitor draws attention to them, and a dispute with a customer is watched by every other customer. Where the sums are modest and the relationship still has value, the cheaper resolution is usually the commercial one, and it stays available only while nothing formal has been sent.
Before anything is sent
Positions harden the moment the other side takes advice, and the quiet routes stop being available once a demand has gone out. While nothing has been sent, everything is still open.