From managing robots to optimising an intelligent fleet.
A manufacturer running AMRs and AGVs had limited visibility into route congestion, charging downtime, idle fleet utilisation and cross-zone movement inefficiencies.
The programme and figures on this page come from an earlier work by our advisers. The client name is withheld under NDA. Figures are as reported at the time. Named references are available under NDA when you engage.
What they were up against.
Route congestion
Robotic traffic created congestion hotspots across zones.
Idle & charging downtime
Fleet utilisation and charging were unmanaged.
No fleet-level view
Robots were managed as individual machines, not a fleet.
Earlier work by our advisers. Client name withheld under NDA. Client name withheld under NDA; figures as reported at the time. Named references available under NDA. We model your own numbers when you engage.
Independent, tied to a named KPI.
UWB for real-time robotic positioning, BLE for asset-interaction tracking, AGV/AMR telemetry for route analytics, and AI orchestration for congestion, idle-time and predictive movement optimisation.
What the engagement delivered.
- UWB RTLS real-time robotic positioning
- BLE asset-interaction tracking
- AGV/AMR telemetry route analytics
- AI fleet orchestration: congestion, idle, predictive optimisation
Results in this engagement
“We stopped managing robots as individual machines and started optimising them as an intelligent fleet.”
Director of operations, industrial manufacturing (client anonymised)
The stack behind it
More case studies like this.
Baseline, method, period and what we excluded.
This page describes an engagement our advisers led in earlier roles. The client name is withheld under NDA. Figures are as reported at the time, with published industry sources where noted.
- Baseline: Fleet utilisation from OEM dashboards; deadlocks handled ad hoc by floor staff.
- Method: Fleet orchestration / VDA 5050-style master view; traffic rules; KPI on jobs/hour and deadlock rate under peak density.
- Period: Live aisle simulation then controlled production window.
- Excluded: Single-supplier happy-path demos without mixed traffic.
Programme context
Fleet intelligence overlays location, mission state and exception codes so operations can see robots from several suppliers in one view.
TRACIO stays advisory/integration-led. We don't sell hardware, so our advice stays independent.
Outcomes in this engagement and how they were earned
Outcomes: higher mission success, faster recovery from blocked paths, better CapEx decisions from utilisation truth.
Lessons from this engagement
Lessons: standardise event schemas; avoid supplier lock-in on the control plane; measure blocked-time not just robot count.
Deeper problem framing
Fleet growth outpaces integration. Each OEM brings a cloud. Operations loses a single pane for exceptions.
Approach
Normalise events; location join; exception workflows; avoid control-plane lock-in; TRACIO does not sell robots.
Outcomes and measurement
Time-to-recover blocked missions, utilisation truth for CapEx.
Governance, risk and what we refuse to claim
Figures from this engagement only. Your payback depends on adoption, integration quality and exception labour, not tag unit cost. Named references under NDA on engagement.
We challenge any supplier who asks you to accept demo-day averages as production truth.
Implementation sequence and change management
Inventory OEM consoles and event payloads. Normalise to an operations schema. Join location. Build exception runbooks. Avoid renewing lock-in via 'free' cloud that blocks export.
KPIs: mean time to recover blocked missions, utilisation by mission type for CapEx decisions.
Buyer checklist, supplier challenges and engagement shape
For fleet intelligence, require event export and schema control so OEM clouds cannot trap your operations data.
Buyer checklist before you sign: (1) written system of record for events; (2) acceptance tests with 95th-percentile performance under real interference; (3) integration owner named in IT/OT; (4) privacy or labour consultation path if people are tagged; (5) cybersecurity zoning sketch; (6) five-year TCO including batteries, spares, recalibration and SLA escalations; (7) exit/export terms so you are not hostage to a cloud tenant; (8) a pilot that can fail without political punishment.
Supplier claims to challenge in this pattern: brochure accuracy without production load; 'compliance included' without artefacts; ROI that assumes perfect adoption in 30 days; references that cannot be called under NDA; install partners who have never worked your vertical's overlays; shared support accounts; and any design that dumps locating onto a flat plant or clinical VLAN.
How TRACIO typically engages: stage-1 architecture and measurement design; independent shortlist and RFP language; pilot acceptance criteria; then optional implementation oversight or programme rescue if a prior pilot stalled. Pages like this one exist so you can prepare the workshop. Your numbers replace every planning band when we model payback.
Risk register themes that recur: mute fatigue on alerts; shadow spreadsheets reappearing beside the platform; battery logistics understaffed; master data too weak to support identity; works-council or IG review starting too late; and success declared on demo day before night-shift reality.
Document baseline windows explicitly: what you measured, for how long, which shifts, and what you excluded. Investment committees and auditors both punish fuzzy before/after stories. If your baseline is weak, spend two to four weeks fixing measurement before ordering anchors or portals. That discipline is cheaper than a stranded deployment and is the difference between a locating programme and a technology souvenir.
Last updated:
Line-side truth
Independent advisory means the radio is chosen last. After the job, the constraints and the system of record are clear.
JIS / WIP that procurement can audit
Start from the sequence problem and the stations that break first.
Score radios against the plant
Metal, conveyors and existing MES integrations beat brochure accuracy claims.
Feed the systems the plant already runs
Position and identity events into MES/WMS, not a second glass the line ignores.
Measurement
Baseline: mission success, blocked-path events, and lineside starve/block minutes. Pilot: one loop or cell with fleet telemetry reconciled to MES/WMS moves. Steady-state: mission success rate, mean recovery time, and starvation minutes on the same ops dashboard. Figures are as reported in this engagement.