Interview. Litigation requires you to preserve and produce every message relating to one counterparty over the last four years. Most of those conversations are in Slack and in two other SaaS tools you do not operate. Walk me through what you would show counsel in week one, and what you would have had to build beforehand.

What the interviewer is testing

Whether you understand that a hold over vendor-held data is a product-capability question before it is a legal one, and whether you will say out loud which parts you cannot do.

The clarifying questions that change the answer

• Which plan tier? Slack's Discovery APIs and native legal hold are enterprise-tier capabilities. On a lower plan the capability does not exist, and the honest answer in week one is an upgrade negotiation, not an export.
• What does retention keep? Slack's own guide is explicit that the Discovery API can return content that was edited and deleted only if retention policies or legal holds preserved it. A workspace set to delete after 90 days has already destroyed year one.
• What is a hold scoped to? Native holds are scoped by custodian. A channel full of relevant discussion where no participant is a named custodian can still be aged out by retention, so custodian lists and channel lists are different questions.
• What is in scope beyond channels? Direct messages, files, canvases, and shared channels with external organisations, where the other party is a separate controller with its own retention.
• The other two tools. Each needs the same three answers, and at least one will not have an export that includes deletions.

A strong answer's arc

1. Suspend deletion within hours, wider than you think you need. A date-bounded preservation across the organisation is reversible; data aged out during the scoping argument is not.
2. Then narrow to custodians and channels as counsel identifies them.
3. Export into an evidence store you control, with per-item hashes and a manifest recording who exported what and when. Search your copy; never search the live tool, because the live tool keeps changing underneath the result.
4. Write the gaps down on day one: what the API cannot return, which period predates the retention setting, which tool has no deletion history. A gap disclosed at the start is a fact, and the same gap found by the other side in month six is an argument about spoliation.

Common weak answers

"We will export the channels" skips edits, deletions and direct messages. "We have a retention policy" describes deletion, which is the opposite of a hold. "We told people not to delete anything" relies on the custodian being both diligent and uninvolved. And "legal will handle it" misses that the only person who knows what the export excludes is you.

What a strong answer adds

The pre-built capability: a continuous archive into storage you own turns a hold from a negotiation with a vendor into a query against your own data, and it decouples your obligations from a plan tier somebody may downgrade on cost grounds. The metric to carry into the review is time from hold notice to preservation in force, measured in hours, with a rehearsal once a year. And the cost: an archive of every message is itself a liability, discoverable in the next matter, so it needs its own retention schedule. The counterweight is channel design. If a conversation must be preserved, it belongs in a channel that is captured; if it does not need preserving, the cheapest control is for the record never to exist.