All tickets undergo a quality control process, which involves both systematic and human checks. Multiple fields are reviewed by Qflow to flag potential issues, from when a ticket first arrives in the system, all the way through to full data extraction. For example:
- Timestamp fields are validated to ensure dates are plausible and not erroneously recorded in the future
- Weight fields are assessed for anomalous values, such as figures exceeding 1,000
- The 'sparkle' next to certain data fields in your records indicates the adjustments that Qflow has made on your behalf, when the system has amended original values from tickets to give increased accuracy and consistency
Spot checks are also conducted by our internal QA team. We look for but are not limited to:
- Unusual weight/volume entries for specific material types
- Inconsistent supplier or project data
- Missing or mismatched permit information
- Logical inconsistencies across fields (e.g. zero weight but a high volume logged)
Despite these numerous checks, there is still the possibility that certain data errors or inconsistencies may pass through both our AI-driven and human quality control checks. This is particularly the case when it is not immediately clear whether an inconsistency stems from an omission in the original ticket data (e.g., “18,500” listed without a corresponding unit of measure), or from an issue introduced during the digitization process, i.e. is the value simply anomalous, or erroneous? While our systems and teams strive for the highest level of accuracy, both automated and manual methods are not infallible.
The quality of the original ticket, and image of the ticket, also plays a significant role in the quality of data output. For example, photos with a significant amount of blur can make it more difficult for the system to read and extract.
We strive to extract data from your documentation in a highly accurate manner, and we consistently deliver ~95% accuracy. However, if you spot what you believe to be an error in your data, please let the Customer Support team know so we can investigate and make corrections where necessary to improve your data accuracy moving forwards.
If you’d like further information on the state of data quality across the industry and to find out more about Qflow’s data quality requirements, please visit The State of Data Quality in Construction report available here.