A small personal-injury firm uses an AI voice intake system. The agent captures four after-hours calls one Friday night. One of them is a prospective client whose adverse party is a long-standing client of the firm in unrelated transactional matters. The agent's conflict check is set to "next-day human review." Monday morning, the intake coordinator catches the conflict. The firm calls back to decline.
The prospective client has already retained another firm. The other firm files a notice of representation. The firm of the long-standing transactional client (the would-be-conflicted firm) ends up in the matter as defendant's counsel, opposite the now-represented former-prospective client. Three months in, the prospective client tells their new counsel that they had disclosed factual information to the AI agent during the after-hours call. The new counsel files a motion to disqualify.
The firm settles the disqualification motion. The malpractice carrier opens a file. The firm's renewal premium goes up.
This pattern is one of the cleanest examples of AI-intake malpractice exposure in 2026. It is not catastrophic. It is also not rare. This post is the practical breakdown of how the exposure works, what carriers care about, what they pay, and what to put in the file before you ship.
This is not legal advice. It is a working overview written for managing partners and intake managers running AI voice intake.
How conflict-driven malpractice claims work
Conflict-related malpractice claims account for somewhere between 12% and 18% of legal-malpractice claims by frequency, depending on firm-size mix and practice mix. They are rarely the highest-severity claim category by total payout, but they are reliably in the top five frequency categories tracked by the major carriers (CNA, Hanover, Axis, ALPS, Markel).
The structural shape of these claims:
- Intake occurs (with or without AI).
- Engagement is opened.
- Conflict is discovered (sometimes by the firm, sometimes by opposing counsel, sometimes by the prospective client themselves).
- The firm withdraws or is disqualified.
- The original client suffers downstream harm: delayed matter, increased cost of replacing counsel, lost evidentiary advantage, settlement delay.
- The original client files a malpractice claim.
The claim alleges that the firm failed to identify a conflict that proper diligence would have caught. The firm's defense rests on showing reasonable diligence. The carrier's exposure is the cost of defending the claim plus any settlement.
AI intake does not shift the duty
The most common firm misconception about AI voice intake and malpractice exposure: "If the AI missed it, the vendor is responsible."
The duty under ABA Model Rules 1.7, 1.9, and 1.18 is the lawyer's duty. The vendor is not a party to the engagement. The lawyer cannot delegate the conflict-check duty to the vendor in a way that absolves the lawyer.
Three places this principle shows up in practice:
- Carrier renewal questionnaires. Carriers ask whether the firm uses third-party intake or conflict-checking systems and how the firm verifies them. The firm's answer is part of the underwriting analysis. "We use AI intake and trust the vendor" is a worse answer than "we use AI intake and verify conflict checks against our internal system on every matter."
- Disqualification motions. Courts evaluating disqualification motions evaluate the firm's process, not the vendor's. The vendor's failure to detect a conflict is the firm's failure to verify.
- Malpractice defense. "The vendor missed it" is not a defense to malpractice. It is a fact pattern within the larger question of whether the firm exercised reasonable care.
The functional rule for AI intake and malpractice: the AI is a tool the firm uses to satisfy its duty. The duty itself does not move.
What carriers actually pay, by claim shape
Conflict-driven malpractice claims have a recognizable severity distribution. Drawing from carrier-published claim data and from claim-counsel reports:
Low-severity (modal): $5,000 to $50,000 in defense and settlement. The conflict is identified mid-engagement. The firm withdraws. The original client retains replacement counsel. The disruption cost is moderate. The malpractice claim settles for the cost of disruption plus modest damages. Most carriers can resolve these inside the deductible or with a single-digit-thousands payment.
Medium-severity: $50,000 to $250,000 in defense and settlement. The conflict is identified late, after substantive work product has been produced. The withdrawing firm cannot effectively transition the matter. Replacement counsel has to redo significant work. The settlement covers the cost of duplicated work plus a portion of fee disputes. Defense costs run six figures.
High-severity: $250,000 to $1M+ in defense and settlement. The conflict creates a material disadvantage in the matter (e.g., privileged information disclosed during the AI intake is later argued to have shaped opposing counsel's strategy). Disqualification cascades through related matters. The plaintiff alleges substantive harm, not just disruption. These are rare but not impossible.
The typical AI-intake-related conflict miss lands in the low-severity bucket because the conflict is usually identified within the first month. The exposure is real but bounded. The reputational cost (and the renewal-premium impact) is often higher than the direct settlement cost.
What carriers ask in the renewal questionnaire
Most legal-malpractice carriers ask about intake processes in their renewal underwriting. The questions evolve, but a 2026 renewal questionnaire typically includes some version of:
- Do you use third-party intake or screening services? If yes, name the provider and describe the service.
- How do you verify conflicts before engagement? Describe the workflow.
- Do you record intake calls? If yes, what is the retention policy?
- Have you had any conflict-related claims, near-misses, or disqualification motions in the last 36 months?
- Do you have a written intake protocol that includes conflict-check timing?
The answers shape the renewal premium. A firm that uses AI voice intake without a documented protocol and without verification against the firm's internal conflict system is materially riskier than a firm with the same tools and a clear process. Carriers reward documented process. They penalize ambiguity.
The renewal-questionnaire honest answer for an AI-intake firm should be:
- "We use [vendor name] for AI-handled intake on inbound calls."
- "Our workflow runs a real-time conflict check during the intake call against our internal conflict-check system. If the check returns ambiguous, the call is flagged for partner review before any engagement decision."
- "All intake calls are recorded with timestamped audit trails for compliance."
- "We have not had any conflict-related claims related to AI intake. [Or: we had one near-miss on [date], described below, with corrective action taken.]"
- "Our written intake protocol requires conflict check at first opposing-party disclosure. Vendor provides the technical capability."
A firm that can give this answer is materially less likely to be paying a higher premium than a firm that cannot.
Documentation that survives the claim
If a conflict-related claim is filed against the firm, the documents that determine outcomes are:
- The intake protocol. Written, dated, signed by the responsible attorney. Specifies when conflict checks run and what triggers an escalation.
- The vendor agreement. Includes data-use commitments, architecture commitments, and (ideally) language treating the vendor as an authorized agent under the privilege analysis.
- The per-call audit trail. For the call in question: evidence of disclosure timing, recording timing, conflict check performed, conflict check result.
- The training record. Documentation that intake staff (paralegals, intake coordinators) are trained on the AI's behavior and the escalation protocol.
- The post-call review. Documentation that the firm reviewed the AI's behavior periodically (typically monthly) and verified that the conflict-check workflow ran as designed.
A firm with these five documents in place can defend a malpractice claim with a strong reasonable-diligence argument. A firm without them cannot.
Five things to put in the file before you ship
The 30-minute pre-launch checklist for any firm deploying AI voice intake:
- Document the intake protocol in writing. Include conflict-check timing, escalation criteria, and the AI's role.
- Sign a vendor agreement that addresses authorized agency, data use, and audit trail. The vendor's standard MSA usually does not. Negotiate.
- Verify the conflict-check integration runs in real time, not next-morning. Conflict checks at the intake stage covers the workflow patterns.
- Capture a sample audit trail and review it. If the audit trail does not show disclosure-timestamp, conflict-check-timestamp, and conflict-check-result fields, the carrier review will surface the gap.
- Update the carrier on the new intake process. Most malpractice carriers want to be told about material changes to intake or conflict-check workflow. A 30-day-advance notice is reasonable.
If you ship AI voice intake without these five steps, the malpractice exposure shifts from "minor" to "the carrier will ask why we did not know."
For the workflow detail on conflict checks, see conflict checks at the intake stage. For the privilege architecture that interacts with vendor selection, see attorney-client privilege when AI takes the first call. For the Rule 1.18 framework, see ABA Model Rule 1.18 and AI voice intake. To see what a malpractice-defensible audit trail looks like in a working demo, book a call.