The Discover Sprint: turning “unknown unknowns” into a data-driven roadmap

l 0–30 days, low disruption, huge clarity

1 . Benchmark true productivity with OEE

We pull shift logs and sensor traces into an automated dashboard to compute Overall Equipment Effectiveness—Availability × Performance × Quality. Anything under ~85 % flags hidden losses, and the factor drill-down shows whether they stem from stops, speed loss, or scrap.

Dashboard overview showcasing OEE Software analytics for production efficiency. The display includes metrics such as OEE, Availability, Performance, and Quality, alongside visual data on production orders, machine activity, and employee status, tailored for evaluating compatibility with regional security policies. Image is an example from https://www.smartfactorymom.com/functionality/oee-software/

2 . Map the production flowsheet

Next we sketch a flowsheet of material, energy, and data paths: which machines feed which, where buffers exist, and how controls hand over. This living P&ID becomes the backbone for both the digital twin and future line-balancing work.

Heat Maps are powerful tools that provide visual insights into data by highlighting areas of significance. Image is an example from https://insightsoftware.com/blog/when-and-why-to-use-heat-maps/

3 . Capture the plant in 3-D (Matterport → Twin)

A two-hour Matterport walk-through creates a point-cloud/mesh that our pipeline auto-ingests, spawning a navigable, photoreal twin in < 30 days. Operators can tag assets, overlay live KPIs, and rehearse change-overs in VR before touching the line.

Example of the immersive digital twins. https://metrology.news/matterport

5 . Identify Asset-Administration-Shells (AAS)

For each critical asset we record its AAS metadata—nameplate, OPC-UA endpoints, maintenance docs—so every motor, sensor, and robot becomes a self-describing node in the Industry 4.0 stack. Think of it as a digital passport that lets the twin “talk” to the shop floor.

6 . Inventory edge devices & comms

Example of Edge device: Jetson Orin Nano

Finally we audit compute at the edge: Jetson Orin Nano, Dragonwing RB3, Advantech HMIs, plus network capacity. The goal is to know where AI models can run locally for sub-100 ms latency and what needs a cloud assist.

Output package

• Baseline deck – OEE scorecard, flowsheet, heat map.

• Interactive twin – hosted preview link for your team.

• Edge readiness report – AAS checklist & hardware gaps.

In a single sprint you move from “we suspect waste” to a defensible list of hotspots and a visual twin that everyone—from ops to finance—can explore. Next stop: the Design phase, where we script the AI playbook to attack those constraints.