Ultrasonic inspections are undertaken on a routine basis during outages of CANDU reactors. Expert interpretation of the data generated during these inspections is used to confirm the integrity of zirconium alloy pressure tubes which house fuel bundles. The inspections form an essential component of the outage programme for CANDU reactors and as such are part of the critical outage path. Analysis of the large volume of data generated is undertaken by a small team of engineers who apply expert knowledge and interpretation through a robust review process that involves multiple human experts reviewing datasets before a consolidated view is determined. Recent work on automated defect detection and analysis on historical ultrasonic data has demonstrated the viability of a decision support tool which can aid the existing analysis process. Testing of the prototype system with engineers and domain experts has enabled improvement in accuracy and capabilities while highlighting tasks within existing procedures where explicit descriptions of all scenarios cannot feasibly be specified and where therefore human judgement plays a more important role. Development of the software, as a support tool to enable more rapid analysis of typical defects, is intended to enable increases in the time available for experts to consider more complex defects. This paper discusses the development of the software prototype, potential deployment routes for the prototype system, considering the interactions with the existing process, benefits which may be realised and limitations that may be addressed through the use of data-driven analysis techniques.