Asset managers in the alternatives space are increasingly overwhelmed by fragmented, unstructured data flowing in from multiple accounting systems, according to Laurent Louvrier, VP of Product, Artificial Intelligence at Confluence Technologies.
As the industry turns to AI-driven solutions to modernize legacy infrastructure and meet growing reporting demands, the focus, according to Louvrier, is not on full automation but on using GenAI to augment human review and build faster, more reliable data pipelines with “precision over blind automation”.
According to Louvrier, at the heart of the AI transformation are three major trends that are reshaping how financial institutions approach data and operations.
“First, the breakthrough in unstructured data processing: GenAI now handles PDFs, documents, emails, research reports, and other content as seamlessly as traditional structured datasets,” he told Traders Magazine.
This alone addresses one of the most pressing challenges in financial services—making sense of high volumes of text-heavy content, he said.
“Second, AI systems now create adaptive ETL processes that are more flexible when encountering schema changes, automatically detect data anomalies, and generate new integration workflows on demand. This represents a fundamental shift from rigid, rule-based automation to flexible, context-aware processing.”
The third, and arguably most disruptive trend, is the rise of agentic AI, Louvrier said: “Agentic AI is emerging as the most significant paradigm shift, enabling autonomous execution of complex, multi-step workflows.”
The result is a new model for how financial firms manage data—moving away from brittle, manual architectures to systems that can dynamically adapt.
“GenAI is fundamentally transforming data pipeline architecture from rigid, manually coded processes to adaptive, intelligent systems that mitigate traditional ETL limitations,” said Louvrier.
These limitations—such as schema dependency, manual intervention requirements, and batch processing constraints—are now being addressed with GenAI’s ability to “enable intelligent data extraction from previously challenging sources, using natural language processing to understand context within unstructured documents and emails, while adapting to schema modifications without manual intervention.”
Louvrier added that GenAI not only automates extraction but also transformation, as it “identifies and corrects inconsistencies, fills missing values, and enriches datasets by inferring additional information from context.”
Looking forward, Louvrier sees agentic AI as the biggest technological force likely to impact finance: “Agentic AI represents the most transformative force in finance automation, enabling systems that don’t merely assist but actively orchestrate complex workflows.”
These intelligent agents collaborate across processes, optimize collectively, and, crucially, learn from feedback, he said. “The key benefit here is that agentic AI systems learn and adapt from environmental feedback, continuously refining their decision-making processes while ensuring compliance with regulatory requirements.” In his view, this is more than task automation—it’s the “autonomous orchestration of entire business processes.”
But with innovation comes regulation, and Louvrier emphasized that regulatory expectations are rapidly evolving: “Regulatory scrutiny operates across two critical dimensions, each requiring distinct compliance approaches.”
At the data governance level, regulators are honing in on how training data is sourced—particularly for proprietary or fine-tuned models, he said: “The ‘fair use’ debate in jurisdictions like the United States exemplifies the complexity of ensuring appropriate data rights for AI model development.”
He added that at the operational level, deployment and oversight are under the microscope. “The EU AI Act classifies many financial AI applications as ‘high-risk,’ imposing strict requirements for transparency, accountability, and continuous monitoring.”
Financial firms must manage model governance comprehensively—from development through deployment to post-launch performance. According to Louvrier, cross-border firms also face growing regulatory complexity: “The UK’s principles-based approach emphasizes safety, fairness, and accountability, while the EU’s prescriptive framework mandates specific technical documentation and testing protocols. US regulators are developing sector-specific guidance, creating a patchwork of compliance requirements that global financial institutions must navigate simultaneously.”
In parallel, asset managers face mounting pressure to increase transparency with regulators and limited partners. “The transparency challenge goes beyond AI technology itself, focusing on data accessibility and openness,” said Louvrier.
“Limited partners and investors increasingly demand customizable access to underlying investment data, real-time reporting capabilities, and transparency into both investment processes and reporting,” he noted.
He said that AI plays an enabling role by providing the technical infrastructure that supports open access: “AI-powered platforms can automatically generate personalized investor portals, create customizable reporting dashboards, and provide access to performance metrics and risk analyses.”
For Louvrier, the most transformative opportunity lies in “democratizing data access, where AI-powered tools enable seamless data exchange between asset managers, regulators, and limited partners.”
