Investment Operations & Securities Processing · Financial Services & Insurance
Should you build or buy Securities Reconciliation (Intersystem / Position-Cash)?
Securities reconciliation software matches positions, cash balances, and transaction records across internal systems, custodians, and prime brokers, identifies discrepancies (breaks), and manages the workflow to investigate and resolve them. It is the operational control layer that ensures a firm's books agree with its counterparties before settlement deadlines.
The build-vs-buy decision for securities reconciliation turns on how rapidly AI and open custodian APIs are lowering the cost of building capable matching engines, weighed against how much of the vendor's feature surface a given firm actually uses; the calculus is shifting and break volume, feed complexity, and vendor pricing all factor in.
Build it, buy it, or bridge?
When building makes sense
Building a securities reconciliation engine has become genuinely tractable for scaled asset managers and fintechs with data engineering teams, in a way that it was not a few years ago. ML-based matching and anomaly detection for financial data is well-documented territory. Duco's platform is AI-native, but fintech firms have also shipped in-house reconciliation engines covering the primary position-cash matching use case, which is evidence that the core function is buildable by a competent data team. The build case gets serious when custodian feed mappings and exception tolerance rules are complex enough to justify owning the pipeline, vendor pricing reflects an enterprise contract rather than actual utilization, and the firm has the capacity to maintain the matching logic over time. Open custodian APIs have reduced the integration cost that previously made self-build impractical. The function is still operationally inert as a differentiator, but cost and control can both favor building at sufficient scale.
When buying makes sense
Buying securities reconciliation makes sense when the firm lacks a data engineering team to build and maintain matching logic, when break volume is modest enough that vendor pricing is proportionate, or when the speed of deploying a platform with custodian feed connections already built outweighs the ongoing licensing cost. Platforms like SmartStream TLM, Gresham Clareti, and AutoRek have built large businesses on the operational necessity argument, and for firms without the scale or team to justify an in-house build, that argument holds. The risk of buying is over-committing to enterprise pricing for features that never get used. Reconciliation is pure operational hygiene; the goal is accurate books before settlement, not a competitive advantage. If a vendor's core matching and exception-queuing handles the primary workflows, buying the base product and accepting its limitations is often more cost-effective than building. Evaluating actual feature utilization before signing an enterprise contract reduces that risk.
The desk read
Securities reconciliation, matching positions and cash across custodians, prime brokers, and internal systems, is operationally necessary and strategically inert. Nobody wins market share from faster break resolution. That combination historically made it a default-buy, and platforms like SmartStream TLM and Gresham Clareti have built large businesses on the operational necessity argument.
The AI shift is meaningful here. ML-based matching and anomaly detection for financial data is well-documented, and Duco's own platform is built on AI-native matching logic. Fintech firms have shipped in-house recon engines covering the primary use case. The build case gets serious for scaled asset managers with data engineering teams, where the exception tolerance rules and custodian feed mappings are complex enough to justify owning the pipeline but the vendor pricing reflects an enterprise contract rather than actual utilization. Open custodian APIs have dropped the integration cost that previously made self-build impractical. Whether buying or extending depends heavily on break volume, feed complexity, and how much of the vendor's feature surface is actually in use.
Frequently asked
What is securities reconciliation (intersystem / position-cash)?
Securities reconciliation software matches positions, cash balances, and transaction records across internal systems, custodians, and prime brokers, identifies discrepancies (breaks), and manages the workflow to investigate and resolve them. It is the operational control layer that ensures a firm's books agree with its counterparties before settlement deadlines.
When does building securities reconciliation make sense?
Building is tractable for scaled asset managers and fintechs with data engineering teams, since ML-based matching is well-documented and production self-builds exist. It gets serious when custodian feed complexity is high, vendor pricing exceeds utilization, and the team can maintain the matching logic over time.
When does buying securities reconciliation make sense?
Buying is right when the firm lacks a data engineering team, break volume is manageable, or speed to deployment matters more than long-term cost optimization. The main risk is over-committing to enterprise pricing for features that go unused, so evaluating actual utilization before signing matters.
What are the main securities reconciliation vendors?
Representative vendors include SmartStream (TLM Reconciliations), AutoRek, Gresham (Clareti / CTC), Duco. B4 Pro scores the full set.
How is AI changing the reconciliation market?
ML-based matching and anomaly detection have made self-build meaningfully more feasible over the past few years, and Duco has built its platform as AI-native from the ground up. This is shifting the build-or-buy calculus for data-capable firms, particularly where vendor pricing does not reflect actual usage.