Untangling corporate actions data complexity

Amid fragmented sources, inconsistent formats and growing demand for intraday updates, LSEG is developing artificial intelligence capabilities and leveraging expert validation to bring structure and speed to corporate actions data.

As corporate actions become more frequent and financially significant, firms face mounting pressure to process them with speed and accuracy. Yet the data remains highly fragmented – arriving in unstructured formats, from diverse sources and with little standardisation. This fundamental complexity lies at the heart of the industry’s ongoing challenge.

According to LSEG’s Fausto Marseglia and Paul Barnes, this is not just a technical headache but a strategic risk factor that touches on compliance, operations and front-office performance.

Fragmentation and slow-moving standards

For many capital markets participants – particularly data vendors supplying corporate actions data to clients – this remains a high-risk, resource-intensive domain. From dividend announcements and stock splits to mergers, bankruptcies and rights issues, these events carry material financial, legal and reputational implications. But, despite their significance, the industry continues to grapple with one foundational issue: the fragmented, unstructured nature of source data.

Fausto Marseglia, head of product management for fixed income and derivatives data at LSEG, explains:“HTML and PDF contribute close to 80% of document types used in corporate action announcements.” Issuers, exchanges and regulators each publish announcements in different formats – often inconsistently and even within the same organisation – necessitating widespread manual intervention in the sourcing process. This fragmentation means that firms, particularly data vendors, must devote significant resources to reviewing, interpreting and standardising incoming content before it can be processed.

Even where there are standards in place, such as Swift’s ISO 15022 and its successor ISO 20022, they are not always sufficient. Marseglia points out that ISO 15022 lacks structured fields for common bankruptcy attributes, such as filing type or procedure. “Those arguments are thus relegated to narrative text and/or exposed via proprietary formats,” he says. The absence of structured fields undermines automation and introduces manual interpretation at critical points.

Paul Barnes, director of corporate actions at LSEG, adds that, even within structured delivery formats, “there’s always some overhead”. Corporate actions often involve exceptions, and clients may be forced to interpret narrative data or reconcile inconsistencies. While automation has come a long way, Barnes notes: “We don’t wait until we know everything before we report anything – clients want timely updates, even if they’re partial.”

You can’t wait until everything is known. Speed matters just as much as accuracy. And clients want both

On broader industry reform, Marseglia is pragmatic. He acknowledges that, while ISO 20022 “was established more than 20 years ago, take-up remains modest in the global data vendor community. Standards are slow to evolve and far from adequate.”

While LSEG participates in the Market Data Provider User Group (MDPUG), progress is slow. Barnes shares a similar view: full standardisation remains an admirable goal, but “it won’t happen in my lifetime”, he remarks, highlighting the persistence of legacy systems and varying regional adoption.

Intraday delivery

While standardisation evolves slowly, client expectations are accelerating. Financial institutions are no longer satisfied with end-of-day data updates. Increasingly, they want intraday delivery – with corporate action information distributed multiple times a day and with minimal delay from the moment relevant events occur, ideally as close as possible to real time.

LSEG has responded to this pressure by releasing corporate actions data every 15 minutes, on a 24/7 basis. The shift to high-frequency delivery supports a range of workflows, from portfolio valuation and risk monitoring to order management and trading. According to Marseglia, “most clients are expecting intraday corporate actions”, especially those managing portfolios with multi-market exposure where the timing of corporate events can impact rebalancing strategies.

The demand for speed is particularly acute in the Asia-Pacific region (Apac). “In Apac, we see a request for lowering that time lag to one hour,” he adds. Institutions with asset-servicing responsibilities – such as custodians and fund administrators – are especially sensitive to timing. Short-notice events, particularly those involving elections or mandatory choices, can introduce material operational, financial and reputational risk if not acted on swiftly.

To meet this demand, LSEG has adopted an iterative publishing model. Rather than waiting for all information to be confirmed, the firm pushes partial updates to clients as soon as details become available, then refines the data with subsequent releases. “You can’t wait until everything is known,” says Barnes. “Speed matters just as much as accuracy. And clients want both.”

He adds that LSEG’s service is “timely, but non-streaming” by design – providing frequent updates in batch form while preserving stability and clarity for downstream systems.

Bringing AI into the workflow

To manage both the scale and complexity of corporate actions data – and to improve automation and data quality – LSEG is investing in AI and machine learning. These technologies are being tested in a series of pilot programmes across selected markets and event types.

The objective, Marseglia explains, is to evaluate “to what extent these technologies can enhance sourcing and validation activities … reduce human intervention and, at the same time, increase data accuracy and completeness”. These exploratory projects assess how well machine learning models can extract data from unstructured documents, classify event types and identify discrepancies.

That said, automation is not the endgame. LSEG’s approach is to embed AI into the existing workflow in a way that complements – rather than replaces – human expertise. “The majority of sourcing, extraction and processing of data is performed by hundreds of specialised content analysts,” Marseglia stresses, indicating that these professionals play a critical role in ensuring the accuracy, completeness and timeliness of corporate actions information.

“In a future scenario where we will embed AI and machine learning into our corporate actions workflows, we believe the involvement of specialised content analysts will still be required for oversight and validation purposes,” he adds.

Barnes reinforces that view; AI can play a useful role in triaging inbound data, identifying anomalies and surfacing high-risk events for closer human review, he says. “The technology should help us make smarter decisions about where to focus human attention – that’s where the real value is, not just in removing workload, but in improving judgment.”

Quality and responsiveness

Even as LSEG pursues automation, it continues to invest heavily in validation and client support. The firm sees data quality not just as a technical benchmark but as a service imperative. According to Marseglia, validation processes include “pre- and post-validation checks to detect the errors and fix them before data delivery”, and these checks are built into multiple layers of the workflows.

In a future scenario where we will embed AI and machine learning into our corporate actions workflows … the involvement of specialised content analysts will still be required for oversight and validation purposes.

Barnes describes how operational commitments help sustain client trust – not just during incidents, but in how the service scales and adapts over time: “We log everything and we learn from it.” That feedback loop, combined with structured internal reviews, helps the team improve consistency and reliability at scale. He also notes that “our maintenance interfaces are defensively engineered”, with built-in exception handling and business rules to prevent systemic errors.

The firm’s service model includes a 24/7 global help desk, with initial response times guaranteed within 15 minutes and more than 80% of queries resolved within three hours. These service-level agreements are supported by a network of dedicated data specialists with local market knowledge and language skills who offer operational and technical assistance to clients across regions.

Understanding client workflows

LSEG also sees the role of data vendors shifting. As client workflows grow more complex and integrated, expectations expand beyond simple delivery. Marseglia believes vendors must be “fully knowledgeable about the relevant client workflows … and maintain an ongoing dialogue” to anticipate needs and develop solutions.

This means supporting custom integrations, responding to ad hoc event analysis requests and advising on governance models for corporate actions handling. In practice, that can require walking clients through intricate event types, comparing vendor feeds or adapting output formats to meet specific compliance requirements. Barnes echoes this point, adding that responsive support plays a crucial role in these relationships: “The best client relationships are the ones where there is a two-way conversation – where we understand their workflows and they know how we operate.”

Looking ahead

LSEG’s strategy for corporate actions includes expanding asset class and event-type coverage – including a newly launched service for roughly one million government and corporate bonds – and deepening automation across the sourcing-to-delivery pipeline. The firm continues to enhance tools that streamline workflows, eliminate bottlenecks and support faster data ingestion.

On the question of industry-wide standardisation, LSEG remains engaged but pragmatic. Marseglia sees value in such ongoing initiatives as ISO and MDPUG, but notes that the real friction lies at the source – where issuers and regulators still publish announcements in unstructured, inconsistent formats. In many cases, automation can help navigate this complexity more effectively than waiting for standards to catch up.

It’s a philosophy that reflects the realities of today’s corporate actions environment – fast-moving, highly unstructured and high stakes – where adaptability may matter more than uniformity.