Page 58: of Maritime Reporter Magazine (August 2026)

SHIPYARD TOOLS

PITFALL 14: Fearing AI move parts, consumables and tools are common

Some in the industry still fear automa- culprits of slow cycle time).

tion will take their jobs. The reality is that • Weld sequence software can recognize when an automation changes the nature of work and operator did not weld the correct seam; if a 10-ft. seam opens new opportunities, and industry should treat AI the took 10 seconds to weld, something is wrong. same. The shipyards that win more contracts and operate more pro? tably will be the ones that make best use of AI PITFALL 17: Relying on Clipboards and automation. AI agents (or “agentic AI” solutions) act au- Unless you have stock in an orthopedic company, data tonomously on behalf of a human user and can interact with management by walking around a shipyard makes no sense other software and AI solution. when digital solutions exist. Manually checking parameters

There are already “AI agents” that layer on top existing on hundreds (more probably of thousands) of welding systems

CAD/CAM, ERP, PLM and QC systems for manufacturing, consumes thousands of hours and opens the door to data entry such as for automating process planning, estimating, inven- errors. There is no real way to catch out-of-spec situations in tory management and customer service (the dreaded chatbot). real-time, and any analytics that occur come long after the

Expect to see AI-driven solutions that further enhance vision weld metal gets deposited.

systems, enable autonomous decision-making to drive robot As an example, consider hardware/software solutions that adaptability and ? exibility, plans weld sequences to manage monitor shielding gas ? ow. Everyone knows that even though distortion on complex weldments, and monitor machine data a FCAW WPS may call for 35 CFH, operators tend to set gas and QC analytics to optimize performance. As with any other ? ow on the high side. An extra 5 CFH on hundreds of weld- technology, the best sources of advice will come from people ing stations adds up to hundreds of thousands of dollars a year who know the industry and their customer’s operation. (and possibly millions if it causes weld defects associated with excess ? ow). The ability of digital systems to identify when

PITFALL 16: Drowning in Data gas ? ow goes outside of the WPS delivers immediate bene? ts.

Don’t confuse data collection with actionable intelligence. Detecting leaks (gas ? owing when or where it shouldn’t) has

Avoid systems that merely ? ll spreadsheets; rather, focus on signi? cant safety implications, plus it could prevent unplanned systems that interpret what the data means, such as pattern work stoppages if a cylinder or bulk tank leaked and ran empty.

recognition, operational guidance and alerts when a process goes/trends out of speci? cation. Examples include: PITFALL 18: Not Automating for Safety • QC software that sends an alert about undersized Hazards in shipyards include injuries to hands (crushes, welds. The solution may to be work with the operator cuts), skin (burns), eyes (arc ? ash, spatter, grinding sparks), to improve gun mechanics and travel speed. strain/repetitive stress, fumes, combustible gases, and trips/ • Identify top-performing areas to replicate success elsewhere. falls. Automating any process that reduces these injuries im- • Uncover hidden bottlenecks (waiting for cranes to proves workplace wellness. For example, automated data col-

PITFALL 17: Manually collecting data, such as gas ? ow rates or welding pa- rameters, wastes valuable time. Connected hardware and software systems free valuable time, eliminate manual data entry errors, can identify trends quick- ly and prevent wasting shielding gas.

58 Maritime Reporter & Engineering News • August 2026

MR #8 (50-65).indd 58 MR #8 (50-65).indd 58 8/6/2026 6:42:37 PM8/6/2026 6:42:37 PM