Hypothetical Mid-Case eDiscovery Vendor Migration That Preserved Privilege

Hypothetical scenario: a Tampa-area solo employment-defense attorney inherited a wage-and-hour matter two months into discovery. Eighty-seven thousand documents had already been uploaded to a familiar cloud eDiscovery platform with an AI-assisted review engine. The matter was on a tight production schedule, and the platform was producing usable results.

Then opposing counsel served a discovery request asking which third-party AI vendors had touched privileged materials, citing Morgan v. V2X. The attorney pulled the vendor's data processing addendum and discovered a sub-processor list that included a generative AI partner never identified at engagement. The platform's standard terms also permitted aggregated analytics on processed data — a clause the attorney had not read closely at signup.

How the hypothetical response was framed

1. The attorney engaged outside counsel to evaluate the privilege exposure and the protective-order posture before responding to opposing counsel.
2. A 5-question vendor-vetting framework was applied to the existing platform: physical location, AI processing, contract language, affidavit capability, and migration mechanics.
3. The vendor declined to sign an affidavit confirming compliance with the operative protective order. That single answer drove the migration decision.
4. The matter was migrated to on-premises processing on firm-owned hardware. The export preserved coding work and metadata; the prior vendor certified deletion across backup systems within thirty days.
5. The migration, the sub-processor history, and the deletion certifications were documented in a sworn affidavit produced in response to opposing counsel's discovery request.

Why that mattered

In the hypothetical, the attorney preserved the privilege claim, avoided a sanctions motion, and produced a clean chain-of-custody record. The judge accepted the migration narrative because the affidavits were detailed, the deletion certifications were sworn, and the on-premises architecture removed any future sub-processor exposure.

The structural takeaway is that the eDiscovery vendor is part of the privilege strategy. The 5-question framework is not a substitute for the protective-order analysis; it is the architecture check that decides whether the vendor can even satisfy the order. When the answers are wrong, the right time to migrate is before the discovery request lands — not after.