Case Study: AI Boosts Takt Times

San Franciso-based Invisible AI illustrates what AI can do for manufacturers in a case study of an OEM supplier to the auto industry that was having trouble keeping up with post-pandemic demand. The vendor’s combination of digital twinning and changes on the shopfloor helped the customer increase manufacturing throughput by more than 60 percent.

According to the case study, reported by the Association for Advancing Automation1, Invisible AI deployed Edge AI devices at points in the supplier’s production line to find the reasons why production volumes lagged their takt targets, i.e., the duration of product assembly needed to meet demand. In the case of the automotive and aerospace specialist’s client, per-shift production had to rise from 320 units to 520 units.

Edge AI captured production metrics in real time and Invisible AI then used anomalous cycle detection to analyzes whether a human or machine is performing within an expected range during a process cycle. Using a digital twin, thevendor pinpointed process and design bottlenecks, allowing the manufacturing to implement solutions.

These included introduction of autocycling at overloaded stations to accommodate for faults by less-skilled operators. Personnel were retrained using video to improve station cycle times and moving materials closer to stations saved steps (and time).

The improved better visibility identified spikes in cycle times at some stations that AI tools registered in dashboard monitors. The changes allow shift managers to spot issues in real time and correct them without interuptiing workflow.

According to Invisible AI’s case study, the modified workflow produced immediate benefits, including the rise of throughput to per-shift production targets in just four hours2.