Speedy data reconciliation across systems

Smart data reconciler

Reconciliation of systems is hindered by contradicting data to obtain one version of the truth and ensure compliance, reconciliation of multiple systems is a common approach. As records across systems will contradict each other, reconciliation involves hunting across systems for those records that match the truth best: Golden Records. Finding Golden Records manually is costly, time-consuming and generally not feasible.

The challenge

How PA helped

We tailor the algorithm in 6 to 8 weeks. We work in an open source Hadoop sandbox environment with multiple data sources loaded onto the platform, allowing for rapid prototyping.
We ensure progress by validating the results along the way, in biweekly sessions with client experts. This is how we know where to focus our efforts to improve the outcome the most.
We ensure relevance: results that can be used. By working closely with clients throughout the process and understanding their needs.

Benefit for the clients

The client, a Dutch bank, needed to reconcile 100 CRM systems for legal purposes. Manual identification of the Golden Records in conflicting customer data would have required 24,000 man hours. We delivered our solution in 6 weeks, tailor-made to identify customer Golden Records across our client's systems, reducing the resources required by 20,000 man hours.