A joint defense agreement lets separately represented parties who share a common legal interest exchange privileged information and strategy without waiving attorney-client privilege or work product protection as to outside parties. It is built on the common interest doctrine, which extends privilege protection to communications shared among parties and their counsel who share a common legal interest in litigation or anticipated litigation — but that protection is not automatic, and a poorly documented or poorly scoped arrangement is one of the more common ways privilege gets waived by accident.

Because the whole point of the agreement is to allow sharing that would otherwise waive privilege, the drafting has to be precise about what is shared, who it can be shared with, what happens if a member's interests diverge from the group's, and what happens to previously shared material after a member leaves, settles, or is dismissed. A vague or informal understanding — 'we're all on the same side, so we'll just talk freely' — is exactly the pattern that later gets challenged successfully by an opposing party seeking to compel disclosure.

The common interest doctrine is not an independent privilege — it is an exception to the general rule that disclosing privileged material to a third party waives the privilege. It permits parties who share a common legal interest to share otherwise-privileged communications with each other's counsel without waiving privilege as against parties outside the group, provided the communication is made in furtherance of that shared legal interest and kept confidential from outsiders.

The interest shared has to be legal, not merely commercial or a shared general desire for a favorable outcome. Courts scrutinize common interest claims closely, and a poorly defined or overly broad claimed common interest is a frequent basis for a court finding privilege was waived when the arrangement is later challenged in discovery.

State the specific common legal interest the agreement is built around — defending against a specific claim, responding to a specific investigation, opposing a specific motion — rather than a vague reference to 'this litigation' generally. A narrowly and accurately defined common interest is far more defensible than a broad one asserted after the fact.

Specify what categories of information may be shared (legal analysis, strategy, draft filings, factual investigation results) and explicitly exclude anything a member does not want shared, such as information relevant to potential cross-claims between members that may exist alongside the common defense.

A joint defense group's interests can diverge over time — one member may want to settle while others want to continue litigating, or evidence may emerge implicating one member more than others. Build in a mechanism for a member to withdraw from the agreement, and specify in advance what happens to information already shared: it typically remains subject to the confidentiality and privilege obligations of the agreement even after withdrawal, since the point of the agreement was never contingent on continued membership.

Address directly what each member's counsel can and cannot do with jointly developed strategy if the group's interests later diverge — for example, whether counsel who received joint defense material can later use it adverse to a former member's interests, which most agreements expressly prohibit.

A member who settles or is dismissed from the case does not automatically lose their obligation to protect previously shared privileged material — nor does that member automatically retain a right to use it for their own purposes going forward. The agreement should state explicitly that confidentiality obligations survive settlement, dismissal, or termination of the agreement itself, and specify whether the departing member's counsel must return or destroy shared materials.

This provision matters more than it might seem: a departed member with continuing exposure (for example, in a related but separate proceeding) has a real incentive to want to use previously shared strategy, and a well-drafted survival clause is what prevents that.

The most frequent failure is treating the agreement as a formality rather than an operative document — signing a boilerplate agreement and then sharing far more broadly, or with people not covered by it, than the agreement actually authorizes. Privilege protects what the agreement actually covers and how the parties actually behaved, not what a signature page says in the abstract.