The Additional Insured Labyrinth: Pleadings and the Duty to Defend

Pleadings

The first decision for anyone entering a maze is which way to turn at the initial crossroad. Similarly, with AI coverage, an early decision point generally occurs when an insurance adjuster receives a tender from an upstream party requesting AI coverage. To respond to the tender, the adjuster will need to review and compare the allegations in the pleadings—specifically the claimant's complaint—with the AI endorsement wording.

Often claimants' attorneys intentionally draft complaints to be vague with the allegations broadly worded to cast a wide net of coverage, and the parties can be united in interest in seeking to maximize insurance. In addition, while pleadings will typically contain allegations against the putative AI as a defendant, the complaint may or may not include the adjuster's named insured as a defendant. And sometimes, the pleadings will refer to the named insured as a nonparty but not make any direct allegations against that entity.

Accordingly, the pleadings, allegations, and procedural posture can have an early and important role in AI coverage analysis. After reviewing the allegations, the adjuster must then consider whether the pleadings trigger AI coverage.

Since 2004 and beginning with the Additional Insured—Owners, Lessees or Contractors—Completed Operations (CG 20 37 07 04) form, the Insurance Services Office, Inc., AI endorsements provide coverage to putative AIs "only with respect to bodily injury or property damage caused in whole or in part by the acts or omissions" of the downstream named insured. Many customized or "manuscript" AI endorsements use similar trigger wording, and as a result, this language is now largely industry-standard.

Burrowing Deeper

When an owner or GC impleads subcontractors into a lawsuit, can their third-party or cross-claim pleadings trigger a duty to defend under standard AI wording? Would a reference in the complaint to a nonparty subcontractor be sufficient to trigger the subcontractor's AI coverage? Another consideration is whether the downstream insurance adjuster should consider facts it knows or should know that implicate the named insured's responsibility (i.e., extrinsic evidence beyond the allegations in the complaint).

A recent reported federal court decision answered these three questions in the affirmative to find AI coverage. In United States Specialty Ins. Co. v. American Empire Surplus Lines Ins. Co., No. 25-cv-7415 (JSR), 2025 U.S. Dist. LEXIS 223363 (S.D.N.Y. Nov. 12, 2025), an injured construction worker sued the owner at a site where she tripped and fell. The owner sued the construction manager, MDG, which, in turn, sued its subcontractor AMB. Thus, while the claimant did not sue AMB, the owner and construction manager's third-party complaints alleged AMB caused the loss because it was responsible for the debris on which the worker tripped. The owner and construction manager then tendered to AMB's Empire policy, whose AI endorsement contained standard wording.

The court held that allegations in the third-party complaints were sufficient to trigger AI coverage. But the court also found that the third-party complaints "read in the context of the Underlying Action" raised a reasonable possibility that AMB caused the loss, thus triggering AI coverage for the owner and construction manager.

Next, the court found that the underlying complaint's reference to the negligence of the owners' "agents, servants and/or employees" also, "though somewhat less directly" than the third-party complaints against AMB, raised a reasonable possibility that AMB caused the loss, thus triggering AI coverage. See also, Chemours Co. FC, LLC v. National Vacuum Env't Servs. Corp., 673 F. Supp. 3d 490 (D. Del. 2023) (finding AI coverage from a subcontractor's policy even though the subcontractor had not been sued).

Finally, the court considered deposition testimony (i.e., extrinsic evidence beyond the pleadings) to find a reasonable possibility that AMB caused the loss.

Complaint Dismissed

Moreover, there is a growing line of authority holding that AI coverage may be proper even where the named insured was sued but later was dismissed entirely from the underlying claim by motion. See also, Federal Ins. Co. v. Mt. Hawley Ins. Co., 2025 U.S. Dist. LEXIS 56600 (S.D.N.Y. Mar. 26, 2025); WDF, Inc. v. Harleysville Ins. Co. of N.Y., 2021 NY Slip Op 02621, 193 A.D.3d 667, 146 N.Y.S.3d 128 (App. Div. 1st Dept.); Old Republic Gen. Ins. Corp. v. Consolidated Edison Co. of N.Y., 2021 NY Slip Op 02466, 193 A.D.3d 595, 146 N.Y.S.3d 620 (App. Div. 1st Dept.).

These cases are in accord with the principle that there is no requirement that the named insured be negligent for AI coverage to apply; the named insured needs only to have proximately caused the loss. See my Expert Commentary, "Additional Insured Obstacles Fall: Coverage Is Not 'Premature,'" IRMI, October 9, 2025.

Moreover, prior notice of a dangerous method of work or instrumentality to the named insured can establish foreseeability and, thus, arguably trigger AI coverage. See Julian D. Ehrlich, "'Caused in Whole or in Part': Foreseeable Additional Insured Issues," New York Law Journal, February 14, 2020.

The subtle distinction between negligence and proximate cause will be further explored in future discussions.

The Duty to Defend

The lessons learned from the United States Specialty case include that owner and GC defendants should start third-party actions against subcontractors if there is a good-faith factual basis to believe that the subcontractor caused the loss and that third-party pleading may be relevant to trigger AI coverage.

Moreover, third-party pleadings should be analyzed in the greater context of the claimants' allegations, particularly where the underlying complaint refers to the defendants' "agents, servants, or employees." This is especially true when the named insured is an injured worker's employer, a position the claimant is barred by statute from suing directly. Finally, all available facts should be considered when forming AI coverage arguments, subject to the local jurisdiction rules on how extrinsic evidence affects coverage.

However, parties should be mindful that coverage analysis rules of local jurisdictions vary, and courts, while trending toward finding AI coverage, are not yet consistent as to whether (1) upstream parties can rely solely on their own pleadings or whether such reliance is too self-serving, (2) the extent to which facts extrinsic to the pleadings can be considered trigger coverage (i.e., four-corners or eight-corners rules), and (3) indirect references in allegations against nonparties like "agents, servants, and/or employees" are sufficient to trigger the named insured's AI coverage.

Conclusion

The World-Wide Labyrinth Locator has tracked over 6,600 mazes in 90 countries across cultures, history, religions, artwork, gardens, and other forms—but not counting AI coverage disputes. Labyrinths might be good fun at the right time and place, but as this examination and prior discussions have highlighted, for parties seeking certainty in insurance, the search for AI coverage can be a game of whack-a-mole filled with twists and turns. See my Expert Commentary, "Additional Insured Obstacles Fall: Is Coverage Restricted?," IRMI, October 10, 2025.

Because the duty to defend is broader than the duty to indemnify, some insurers are more inclined to offer putative AIs a defense under a reservation of rights rather than a complete tender acceptance of both defense and indemnity. A future discussion will examine how insurers issuing tender acceptances offering the putative AI a defense only under a reservation of rights often miss seeing the big picture of the maze, which can haunt such decisions.