At the latest Inspection Expo & Conference (IEC) in Austin, TX (February 3–4, 2026), one message came through clearly across multiple sessions: artificial intelligence, like in everything else, is being more integrated in welding automation and inspection workflows. Its value, however, depends entirely on how well it supports qualified professionals and the disciplined procedures and code compliance requirements we’re all tied to.
The conference did not present AI as a replacement for inspectors or welding engineers. Instead, the strongest speakers emphasized something more practical: AI can make good work sharper, faster, and more consistent, but only when it is built on sound data, qualified judgment, and structured documentation.
From a welding inspection perspective, CEI’s CWI attended several of the leading AI-focused sessions and came away with a clear takeaway: the future of AI in welding will not be determined by how advanced the tools become, but by how well they integrate with procedure control and inspection accountability.
Jeff Noruk, President of Servo Robot Corp., opened with a compelling blend of history and realism in his session, “AI-Integrated Welding Automation and Inspection: History Doesn’t Repeat Itself, but It Often Rhymes.” With experience spanning Fanuc Robotics, Newport News Shipbuilding, and automotive R&D, Noruk brought a practitioner’s perspective to what AI is actually doing on the shop floor today.
A major focus of his talk was the use of AI with 3D laser camera systems to scan weld joints, identify seam location, and adjust weld paths in real time. This type of arc seam tracking allows automated systems to stay on course even when parts vary slightly — a meaningful step forward for consistency in mechanized and robotic welding environments.
Noruk also discussed how laser scanners can be used after welding to perform surface-level inspection. By scanning finished weld geometry, inspectors can measure weld size, shape, and location more precisely than traditional hand tools or gauges.
At the same time, he was careful not to oversell the technology. These systems are highly effective at evaluating geometry, but they are not a complete replacement for qualified inspection. Conditions such as porosity, lack of penetration, cold lap, or subsurface discontinuities still require additional validation and experienced interpretation.
In many ways, Noruk’s message set the tone for the entire conference: AI can improve repeatability and visibility, but compliance still depends on qualified oversight and traceable procedure control.
From there, the conversation broadened beyond inspection hardware into workforce realities.
In “Giving Birth to ‘AWStin:’ AWS’s Journey to Implementing Artificial Intelligence,” Carey Chen, CEO of the American Welding Society, addressed what many in the industry are calling the “Silver Tsunami” — the retirement of highly experienced, skilled labor and the growing urgency to train the next generation.
Chen discussed the rise of digital and hybrid learning models, including VR/AR welding simulators and micro-certifications designed to train workers for specific task competencies. Automation and robotics, he noted, will increasingly handle repetitive or hazardous work, but skilled people remain essential to guide, validate, and oversee these systems.
To ensure the audience understood what “AI” actually encompasses, Chen broke down the primary approaches driving its maturity in industry. Rather than treating AI as a single monolithic tool, he explained it as a spectrum of capabilities:
Machine Learning (ML): Algorithms that learn from data and make predictions to automate tasks or identify trends.
Natural Language Processing (NLP): Systems that understand and generate human language, enhancing communication, documentation, and information retrieval.
Neural Networks: Models that mimic the brain’s structure to learn patterns and solve complex problems.
Computer Vision: AI that analyzes images and visual inputs — particularly relevant to welding inspection, where geometry recognition and discontinuity detection are critical.
Chen emphasized that meaningful AI adoption is not about simply adopting the latest technology label, but about matching the right approach to the right challenge--and investing in the disciplined data practices, including proper structuring and labeling, that make reliable outcomes possible.
The takeaway was consistent: AI can accelerate processes, but qualified oversight and strong data foundations remain non-negotiable.
For our shop owners and inspection professionals, this matters directly. As training pathways evolve and qualification complexity increases, maintaining accurate WPQs, continuity records, and expiration tracking becomes even more critical. Tools may help us move faster, but compliance still rests on disciplined documentation and accountable professionals.
Lee Pielaet, President of Pioneer Steel Services, Inc., continued the inspection-focused discussion in his session, “Artificial Intelligence in Welding Inspection: Promises, Pitfalls, and Practical Pathways.”
He outlined how inspection methods are progressing beyond traditional surface evaluation into more advanced techniques, including ultrasonic testing (UT) and AI-assisted interpretation of UT data.
He also highlighted how AI-enabled tools can improve safety and efficiency, such as drone-based inspection in hazardous or hard-to-reach environments.
Importantly, he emphasized that even when AI is used to draft baseline documents or review audit notes, qualified individuals must still validate results. AI can assist, but it cannot own compliance.
As inspection tools become more data-driven — whether through UT interpretation or drone-enabled scanning — the volume of information will only increase. That makes documentation clarity even more important. Inspectors will need to validate not only what the system detected, but how that detection aligns with acceptance criteria, procedure variables, and code requirements.
AI may help interpret signals. It does not carry the responsibility of signing off on compliance.
Calvin E. Pepper of AWS brought a human-centered lens in “AI and the Inspector’s Eye.” He addressed the resistance some inspectors feel, not because they reject technology, but because they understand what is at stake.
Inspection is not simply pattern recognition. Inspectors become experts through education and, most importantly, experience. AI can analyze data, but it does not carry sensory intuition or accountability.
Pepper stressed that the industry will continue to rely on inspectors not only to perform work, but to validate AI outputs, justify conclusions, and maintain trust in the inspection process.
This perspective is critical for long-term adoption. Trust in AI will not come from capability alone; it will come from transparency and accountability. Inspectors must be able to explain and justify conclusions, whether those conclusions originate from personal observation or AI-assisted analysis.
The human role is not diminishing. It is evolving toward validation, interpretation, and control.
Luke Faulkner, Director of Technology at AISC, shared how the organization approached AI implementation through a structured, reference-backed chatbot called “Clark.”
Rather than deploying AI broadly, AISC began by defining the problem, structuring the data, and ensuring that answers were supported by logic and citations.
That model is instructive for welding compliance: AI becomes most valuable when it operates inside a framework of traceability, standards, and defensible documentation.
The lesson from AISC’s structured approach is instructive: define the problem, structure the data, and require references. Welding inspection environments demand the same rigor. AI becomes most powerful when it operates inside a clearly defined framework of standards and traceable documentation.
Technology should strengthen compliance infrastructure, not bypass it.
The Inspection Expo & Conference made one thing clear: AI in welding automation and inspection is not about replacing qualified professionals. It is about equipping them.
When applied correctly, AI tools can help inspectors and engineers work with greater precision, identify issues faster, and document results more efficiently. But those gains only hold when they are anchored in procedure control, code compliance, and structured records.
That is why solutions like ProWrite matter. ProWrite helps teams create, organize, and electronically control welding procedures and qualifications with built-in code assistance and standardized reporting.
AI may sharpen the tools of inspection. But disciplined documentation sharpens compliance.
For more information about ProWrite or the latest in welding code compliance, click the image below.