How Southern Virginia Organizations Are Using AI to Solve Real Business Challenges

The project presented by Huber Engineered Woods focused on a challenge familiar to many industrial operations: identifying quality issues before they lead to defects, waste, or production disruptions.

Using historical production data, the team developed a machine learning model designed to recognize patterns associated with successful and unsuccessful production outcomes. The project combines historical production data, supervised learning, and AI-assisted recommendations to help identify root causes of quality issues and flag when processes may be trending toward undesirable outcomes.

The team emphasized that data quality is just as important as model quality.

“Just because it’s data doesn’t mean it’s good data,” one presenter explained, noting the importance of filtering out downtime events, outages, and unusual operating conditions that could distort results.

The project also reinforced an important principle shared throughout the cohort. AI should support human expertise, not replace it. While the model provides recommendations and early warnings, operators remain responsible for evaluating the information and making final decisions.

For the organization, the biggest immediate benefit is speed. Analyses that once required days of manual effort can now be completed in near real time, allowing teams to respond faster and make more informed decisions.

The project has progressed beyond experimentation and is preparing for controlled production testing to evaluate prediction accuracy, operator adoption, and overall performance.

The team from Reynolds Consumer Products’ South Boston facility explored how AI could help address one of the most time-consuming challenges in manufacturing: creating and maintaining training documentation.

Developing a single standard work document often requires dozens of hours spread over several weeks as trainers, supervisors, process specialists, and operators collaborate to document procedures. By the time documentation is complete, equipment changes or process improvements may already require revisions.

The result is often a gap between documented procedures and actual operations.

The team’s AI use case focused on accelerating the creation and maintenance of standard work documents while capturing valuable institutional knowledge held by experienced employees.

Their goal is to reduce documentation bottlenecks, improve onboarding, strengthen training consistency, and preserve expertise that might otherwise be lost through workforce transitions.

For many manufacturers facing retirements and workforce challenges, ensuring that critical knowledge remains accessible may prove just as valuable as improving operational efficiency.

Representatives from Dedon and Gloster, two outdoor furniture brands operating under one umbrella, tackled a challenge that affected both employees and customers: freight calculations.

As the organizations became more integrated, freight quotes continued to be generated through multiple tools, by numerous employees, across two brands. The result was inconsistent information, internal inefficiencies, and delays in responding to customer requests.

One example highlighted the challenge. A sales representative on the West Coast might be working to close a significant order after the East Coast freight team had already left for the day. Without access to reliable freight calculations, the sale would have to wait.

To address the issue, the team used ChatGPT as a collaborative design partner, consolidating freight maps, spreadsheets, and standard operating procedures into a single freight calculator.

“I used ChatGPT like a business partner,” one team member shared. “We went back and forth 39 times before we had a version that actually worked.”

The resulting tool standardizes freight logic across brands, simplifies the user experience, and provides faster, more consistent quotes regardless of time zone.

The project also underscored an important lesson about AI implementation.

“Garbage in, garbage out. You have to test it, validate it, and make sure what you’re getting out is accurate.”

The team plans to seek executive approval and begin implementation across customer service, sales, and warehouse operations.

For Morgan Lumber Group, a deck board manufacturer in Charlotte County, the challenge was knowledge management.

Critical operational knowledge existed in multiple places: vendor emails, maintenance records, manuals, blueprints, and the memories of employees with decades of experience.

“Every time we lose an employee, we lose a lot,” one presenter explained. “Once they retire, everything they know is gone.”

The company is developing a secure, internally hosted knowledge system that will allow employees to quickly search vendor communications, manuals, operating procedures, maintenance history, and troubleshooting records.

Instead of spending valuable time searching through binders, inboxes, and disconnected files, technicians would be able to ask simple questions such as, “Have we seen this fault before?” or “What fixed this issue last year?” and quickly locate relevant information.

One example demonstrated the potential impact. A recurring equipment fault that previously required 45 minutes of investigation could eventually surface related emails, manual pages, maintenance photos, and prior solutions in seconds.

The project will be implemented in phases, beginning with vendor communications and service records before expanding to manuals, blueprints, standard operating procedures, and institutional knowledge capture.

The team stressed that ease of use will determine success.

“Whatever we do, it has to be simple and easy to use. If they have to click through a bunch of screens and still can’t find what they need, they’ll stop using it.”