Unitary: How AI-Powered Virtual Agents Are Transforming Delegated Authority Operations

Bordereaux management is one of the most operationally intensive processes in delegated authority. Data arrives from coverholders in varying formats, which then must be extracted, reformatted to meet carrier and reinsurer specifications, validated against policy records, and submitted on time, with exceptions investigated and resolved along the way. Multiply that across every active binder and the result is a team spending most of their time on work that is critical but repetitive. For firms growing YOY, costs scale linearly with no efficiency gains.

At Unitary, we’ve spent years building Virtual Agents that can take over this work entirely – not by replacing your systems or requiring months of integration, but by working inside the tools you already have, the same way a bordereaux analyst would. We have deployed this in the Lloyd’s market with significant results: 30% cost reduction, 99% of submissions handled without human intervention, and deployment timelines measured in weeks rather than months.

Delegated authority now accounts for 40% of Lloyd’s volumes and over £5 billion in UK premiums. While underwriting models and distribution have evolved rapidly, the operational machinery supporting delegated authority has lagged behind.

Many MGAs and managing agents still depend on the same mix of spreadsheets, inboxes, manual checks, and rekeying that they relied on a decade ago. Attempts at change have often taken two forms: robotic process automation (RPA) or large-scale platform replacement.

The limitations of RPA are well known. A bot is created to extract data from a spreadsheet or upload a file to a portal. It functions until a coverholder alters their layout or a submission screen changes, at which point the automation fails. RPA depends on interfaces remaining exactly the same and cannot perform the reasoning required to manage bordereaux end to end.

Platform replacements promise more, but come at a high cost. Data migrations, system integrations, retraining, and months of operational disruption are common. These initiatives take months and cost six figures before delivering tangible value. Even when they succeed, they frequently automate only part of the workflow, leaving teams to manually bridge the gaps.

As a result, firms are left with two options: tie up experienced staff with low-leverage work, or hire an ever growing BPO team that requires constant oversight from internal staff who never have the opportunity to focus on the business itself, and introduces additional operational risk. Both options are expensive and do not scale sustainably, with the problem compounding as books grow year on year.

Unitary’s Virtual Agents are robust AI workers designed to operate tools and interfaces the same way humans do. Rather than requiring an API connection to a policy administration system, a Virtual Agent logs in through existing portals, navigates to the appropriate screens, and extracts or enters data exactly as a team member would. This means no integration debt, no dependency on IT resources, and no risk to systems that are already functioning.

For bordereaux operations, Virtual Agents handle the full process end to end. They receive submissions in any format, extract and interpret the data, configure and run ETL pipelines, validate against policy administration systems, troubleshoot data errors, and produce outputs ready for submission. Where confidence is high, they proceed autonomously. In the small minority of cases where ambiguity remains, they escalate to a human with full context.

One Lloyd’s market firm we work with processes several hundred bordereaux each month across multiple lines of business. Before automation, a dedicated team of twelve handled file intake via inbox, ticket creation, configuring and running ETL and reconciliation. As volumes increased, scaling this model was becoming increasingly difficult and costly.

Within five weeks of deploying Virtual Agents, 99% of bordereaux were being processed without human involvement. Turnaround times dropped from days to minutes, and operational costs were reduced by 30%. Importantly, this was achieved without replacing any systems or changing established workflows. The only requirement was user access to existing tools.

The agents simply joined the team and began doing the work—faster and with higher accuracy.

The regulatory environment for delegated authority is tightening. Lloyd’s has made claims a hurdle principle, and capacity providers across the market are demanding higher standards of data quality and reporting. At the same time, MGAs and managing agents are under pressure to grow efficiently, without expanding headcount at the same rate as premium volumes.

Virtual Agents offer a practical path forward. They deploy quickly, operate within existing infrastructure, and deliver accuracy levels that can be contractually guaranteed. Most importantly, they free experienced teams to focus on underwriting support, oversight, and growth rather than manual processing.

If bordereaux is consuming too much of your team’s capacity, Unitary’s Virtual Agents are already proving there is a better way.

Book a demo to see how end-to-end automation works in practice.