Instro AI trials cut engineering response times by 67%

Instro AI Solutions has published results from manufacturing trials conducted with AMRC Cymru and involving several UK engineering and manufacturing businesses.

The strongest gains were reported at Colchester Machine Tool Solutions, Poeton Industries and Star Micronics. In these trials, the software helped staff find technical information, handle customer enquiries and support engineering decisions. The work was carried out with AMRC Cymru, part of the University of Sheffield Advanced Manufacturing Research Centre.

At Colchester Machine Tool Solutions, engineers used the system during service and maintenance work on CNC and manual combination lathes. Average time to find and respond to technical information fell from 5.5 minutes to 1.8 minutes, a 67.3% reduction.

Poeton Industries used the software to manage technical and commercial enquiries. It recorded a 40% to 65% reduction in first-response times, while manual effort for triaging and drafting replies fell by 20% to 35%.

At Star Micronics, engineers used the system 1,222 times during the trial to diagnose alarm codes and search manuals and service records. Decision-making during technical support tasks was 44.6% faster, which Instro rounded to 45%.

Trial Focus

The projects were structured as proof-of-value trials built around specific business outcomes. The results were reviewed with AMRC Cymru to validate the findings.

Each use case addressed a different operational pressure point. At Colchester, the focus was engineering knowledge retrieval. At Poeton, it was the speed and consistency of responses to incoming customer requests. At Star Micronics, the aim was to help international engineering teams resolve technical support issues more quickly.

Poeton handles up to 4,000 customer emails a month and receives around 1,400 requests for quotations each year asking how surface processes should be carried out. For the trial, the system was configured to analyse incoming enquiries, identify relevant knowledge and produce draft responses for staff review.

Lee Mason, Group Digital Transformation Manager at Poeton, said: “Phase 1 showed strong early value, especially in faster, more consistent technical responses, and the tool was well received by our teams. Phase 2 will scale that progress, deepen the use cases, and test how it embeds into daily operations. We’re pleased to continue the partnership.”

Data Problem

A central finding across the trials was that older, fragmented data remains a major obstacle to wider use of generative AI in manufacturing. According to Instro and AMRC Cymru, the issue is less about the models themselves and more about how information is stored across documents, systems and service records built up over many years.

Instro said its software draws information from multiple sources and file formats, including PDFs, then standardises terminology and identifies authoritative material. The aim is to create a usable knowledge layer without large data-preparation projects.

Pritesh Patel, Industrial Digitalisation Technical Lead at AMRC Cymru, said: “These proof-of-value trials acted as the ground truth in witnessing the impact of adopting generative AI technologies, filtering out the noise and hype we experience in society today. It provided manufacturers with an understanding of how AI works under the hood. The trials showed that while the impact of generative AI is massive, the real challenge lies in the ‘reality of data.’ The biggest hurdle that manufacturers face is not utilising AI, but the fragmented legacy data that they have carried for decades. By properly organising this knowledge, generative AI systems such as Instro AI provides engineers with more time to focus on value-added tasks whilst ensuring that they remain as the final decision-makers in an AI-assisted workflow.”

Wider Context

The results come as manufacturers test where generative AI can be used in day-to-day operations rather than in isolated pilots. In these trials, the main uses were internal knowledge access and assisted drafting, areas where companies often hold large volumes of technical material but struggle to retrieve it quickly.

Instro describes itself as a UK software company focused on tailored generative AI systems for organisations with complex operational data. According to background information it provided, the AMRC-linked work covered manufacturers in sectors including automotive, aerospace, medical and construction.

Phil Sanders, Commercial Director at Instro AI Solutions, said: “These outcomes demonstrate how generative AI is moving beyond experimentation and delivering measurable operational improvements across engineering support, enquiry handling and technical decision making for organisations of all sizes. Over the years, even small companies can generate huge volumes of data and expertise that becomes locked away. We help them put that knowledge to work quickly.”