Data quality audit and profiling of existing systems
Key Takeaway
Profile the existing data before you design, migrate, or clean it. Measured fill rates and recent use show the real condition of the data, and interviews and cleanup don't replace that measurement.
A data quality audit checks how reliable the data in an existing system really is before you design a solution on top of it. Data profiling is the core technique. Profiling analyzes the structure, content, and quality of data to uncover patterns, anomalies, and metrics. The result is measured evidence that replaces what stakeholders believe about their data.
Why interviews aren't enough
Discovery interviews tell you which objects support a process, which fields matter, and how records relate. Interviews define scope and intent. However, they don't always reflect the actual condition of the data.
Systems change over time. Fields become unused, integrations change, and records get fragmented across systems. Data profiling adds measurable evidence and concrete examples of these data quality risks. Combining interviews with profiling leads to better design decisions.
The amount of data to assess grows fast. Even a narrowly scoped agentic AI project that covers only a few objects can span hundreds or thousands of fields and millions of records. In CRM orgs older than 5 years, a significant percentage of custom fields can be unused or abandoned.
During discovery at an industrial pump manufacturer, the service manager says technicians record a "Failure Code" on every case. A profile of the Case object tells a different story. The field is filled in on 35% of records, and most of those values were entered more than a year ago. The interview described the process as it was designed. The profile shows how the field is really used today.