Administration Challenges, Part 3: Technology, Standards, and Shared Incentives

Part 1 of this series described the surge in fraudulent claim filings in class action settlements and the administrator responses it produced: an increase in the number of documentation audits, greater claim scrutiny, compressed time for claimants to respond, and vague rejection codes that can make it difficult for legitimate claimants to cure deficiencies. Part 2 examined the legal architecture governing those responses — Rule 23(e) and the settlement agreement itself — and found that class member protections are real but largely untested and underdeveloped.

The legal avenues explored in Part 2 provide a basis for legitimate claimants whose claims are unfairly rejected to seek redress.  Litigation, however, is slow, and outcome can be uncertain. Such cases require a claimant with the appetite to litigate. Even if successful, a favorable ruling can take years, with the decision arriving after the claims it concerned have already been paid or forfeited.

The better solution for preventing fraud, while allowing for legitimate claims to recover, requires changes to the administration process: better fraud screening, clearer contractual terms, and coordination among parties who all lose when fraudulent claims are paid or legitimate claims are rejected. This final part proposes such a framework.

Increase the use of available fraud-screening techniques

Fraudulent bot-driven filings leave signatures that are distinct from the trade data itself, including volume clustered in a single IP address or adjacent ranges, submissions arriving at machine speed (often overnight), repeated device fingerprints and browser profiles, disposable email domains, recycled payment instructions across nominally unrelated claimants, and form responses that vary just enough to defeat exact-duplicate matching.

Banks, payment processors, e-commerce platforms, and ad networks have been mitigating exactly these patterns for two decades using rate limiting, device fingerprinting, velocity checks, anomaly scoring, and reputation data. It is not that such tools are unavailable for securities actions; it is that securities claims administrators may have less experience with them, combined with less time and capacity to develop and employ such systems when constrained by fixed fees and compressed timelines.

Nevertheless, with greater adoption and use of these tools and a risk-based posture, administrators can save themselves the time and effort they would have spent on additional audits by more easily weeding out fraudulent claims based on these signatures. Administrators should score submissions, triage claims, and reserve additional documentation demands for the small share of claims that fail these metrics or carry a bot signature. That is a better use of a limited review budget than auditing more claims and applying blanket scrutiny and focuses efforts on the truly fraudulent claims.

Dealing with AI-enabled fraud

Part 1 described how generative AI has made it easier and cheaper for bad actors to forge trade confirmations and brokerage statements. The same technology can also be used against them: AI tools that use document metadata and provenance analysis, detection of template and typography inconsistencies, and automated cross-checking of statement figures against historical market data can flag fabrications that would pass a manual review by an overextended analyst.

The more durable fix is to get out of the document business altogether. PDFs are easy to forge. Structured data delivered through verified channels — direct custodian and broker feeds, submissions from filers under contract, machine-readable formats, and consistent entity identification through Legal Entity Identifiers (“LEIs”) — is considerably harder to fake and cheaper to process. A claim with data arriving from a verified filer of record does not need letterhead to be credible.

Contractual protection

Part 2 showed that the standard deficiency clauses in securities settlement agreements are one-sided and, in key respects, undefined.  This makes it difficult for class members to cure deficiencies when their claims are audited. A few obvious improvements to such agreements are available that can make it easier for legitimate claims to survive:

• Define “curable.” List the categories of defect that entitle a claimant to notice and an opportunity to cure. As drafted today, claimants cannot tell whether a given defect entitles them to notice.
• Allow claimants to toll deadlines. Instead of giving a hard deadline from the mailing of a notice, allow claimants more time by the ability to acknowledge receipt. This has the advantage of also giving the administrators a heads-up of how many claims they will be receiving before they flood in.
• Bind the administrator to a timeline as well. A deadline to acknowledge a cure submission and a date by which the administrator must issue its decision will keep claimants from waiting beyond distribution to learn if their claim survives.
• Publish a documentation safe harbor. Specify more clearly what will suffice — for example, a custodian or broker statement on institutional letterhead whose holdings and trades reconcile — so claimants know the standard before they file rather than after they are rejected.
• Define the rejection codes. A code paired with a plain-English definition and a stated cure path would eliminate many disputes.

Administrators and law firms would also benefit from many of these practical changes. Precise terms produce fewer disputed rejections, fewer requests for court review, and less work late in the administration when reopening or revisiting claims is most expensive.

The case for a coalition

The obstacle to these solutions is not technical difficulty but divided incentives. Claimants and third-party filers bear the cost of substantiating and curing claims and face a reduced recovery for every fraudulent claim paid from a finite pool. Administrators work under fixed fees and shortened timelines, and face reputational exposure if fraudsters are paid, which makes a strict documentation posture the easiest defensible course. Class counsel wants efficient distribution and finality. Left alone, these pressures produce exactly the outcome this three-part series has described.

One interest, however, is shared: no participant wants fraudulent claims paid out of the settlement fund. That overlap is enough to support collaboration.

Solutions can be achieved.  Claims administration challenges are similar to those that securities firms have solved in the past including standardized messaging formats and corporate-actions data conventions.  As here, these involved problems that no single industry participant could solve alone and from which they all benefited through coordination.

Third-party filers are well positioned to assist the conversations and solutions. Firms like FRT file claims in volume, see deficiency patterns across different administrators and cases, and must deal directly with anti-fraud measures and arguably over-broad audits. FRT has advocated for systemic change on behalf of investors abroad, submitting affidavits and publishing commentary in support of robust recovery processes in Australia, the United Kingdom, Japan, the Netherlands, and Germany. Domestic incentives are potentially less aligned, but the overlap is hopefully sufficient to begin a discussion among claimants, filers, class counsel, and administrators — and perhaps defendants, who fund these settlements, have no interest in compensating criminals, and are parties to the settlement agreements.

What claimants can do now

While we hope positive changes in the industry will reduce the work faced by claimants, there are several concrete ways to protect claims in the meantime.

• Preserve trading data across time and custodians. Retain transaction and holdings records through custodian changes, restructurings, and fund closures, and document account lineage. Claims most often fail because records are unavailable, not because they never existed.
• Keep entity identification consistent. Naming discrepancies across accounts and legal entities are a recurring source of rejections that consistent use of legal names and LEIs largely prevents.
• Centralize notice intake. A notice with a short response time is easily lost in an individual’s email inbox. A monitored, shared email address and a tracked queue keep cure deadlines from expiring unnoticed.
• Select filing providers on their skills in handling deficiencies, not just those charging the lowest price. As Part 1 noted, the cheapest provider may not get the best outcomes. What matters is how a provider verifies data before filing and how it responds when an administrator pushes back. Claims don’t get paid unless they survive the deficiency process.

Conclusion

Fraud is not going away, and administrators are right to take it seriously. The question this series has posed is whether the response will be coordinated and fair rather than reactive and indiscriminate, and whether it can be implemented without bona fide claimants paying the cost.

The technology tools needed to separate fraudulent submissions from legitimate claims largely already exist, and most of them impose nothing on the claimant. What is missing is coordination, and a willingness to treat the deficiency process as infrastructure worth developing further rather than a cost to be minimized. Absent that, the legal routes described in Part 2 remain the backstop — and legitimate claimants may suffer adverse consequences in the meantime.