MAIA · Crew & Line-Balancing Engine

Shift Recommendations

Crew rotations, overtime, predictive call-ins and cross-train deployments — proposed against the live demand curve so OEE clears target without manual scramble. Approve a recommendation and the schedule, market and PM systems update in one motion.

Initializing 3 models active 4/6 lines

Pending review

3 high-severity

Auto-approved · today

saved 18 ops-min/each

+2.4 pp

Avg uplift / accepted

OEE delta on accepted

78%

Acceptance · 7d

rolling — 142 / 182

62 hr

Coverage gap closed

this week vs base plan

184 hr

Forecast OT · this wk

$8.4k incl. premium

$48k

$ saved — accepted

30d realized vs projection

Live queue · MAIA proposals

Recommendations awaiting decision

Each card shows the trigger signal, projected impact, model confidence, and the operators affected. Auto-approved cards (≤3 ops, ≤2h, ≥0.95 confidence) execute without sign-off.

5 awaiting 1 auto 1 approved

Crew rotation Awaiting approvalNext 4h · A · DayDemand-Match v3

Move 2 operators from L2-FA to L6-CNC for next 4h

TriggerL6-CNC short 8 on shift; cycle-time +1% vs target

ImpactRestore L6-CNC run-rate to plan, avoid 144 units lost over the window

ModelDemand-match model · seasonality + fatigue weighted

Affected

JG

RL

L2-FAL6-CNC

2 operators

OEE

+3.6 pp

Throughput

+84u

Cost

$1.2k

spend

Confidence91%

Overtime Awaiting approvalSat 06:00 → 22:00Backlog-Burn v2

Approve weekend OT — Body Weld backlog 14h

TriggerL3-WLD backlog accumulated 14h vs schedule; downstream paint queue dry by Tue 06:00

ImpactBurn 92% of backlog in two 8h Sat shifts; protect 2 customer commitments

ModelBacklog-burn model · sequenced WO simulation

Affected

CM

TW

BT

JM

AP

L3-WLDL3-WLD

6 operators

OEE

+1.2 pp

Throughput

+210u

Cost

-$3.9k

save

Confidence86%

Cross-train deploy Awaiting approvalNext swing shiftCross-Train Optimizer

Cross-train pull: Walter Collins certified on CNC, deploy to L6 to relieve M. Garcia

TriggerL6-CNC operator fatigue index 78/100; Walter Collins certified 1.2yr · only 2% utilization on cell

ImpactMaintain L6 run-rate, reduce minor-stop rate during fatigue window

ModelCross-train optimizer · cert-currency + fatigue blend

Affected

WC

L3-WLDL6-CNC

1 operator

OEE

+1.8 pp

Throughput

+44u

Cost

spend

Confidence79%

Predictive call-in In progressTonight · C · NightPdM-Aware v1

Pre-stage 1 maintainer on L1-FA — vibration trending up

TriggerPredictive model flags failure window in 36–52h on ABB Robot 1; covers PM 2026-05-04

ImpactCuts MTTR by ~38min if it triggers; avoid line-stop on next shift

ModelPdM-aware model · vibration + temp residuals

Affected

SM

Maintenance poolL1-FA

1 operator

OEE

+0.9 pp

Throughput

+52u

Cost

-$640

save

Confidence74%

Crew rotation Auto-approvedCurrent hourDemand-Match v3

Slide 1 operator L5-PAINT → L3-WLD for current hour

TriggerMicro-imbalance: L5-PAINT +1 vs plan, L3-WLD −1

ImpactMaintains both lines at plan headcount

ModelDemand-match · within auto-approve band (≤3 ops, ≤2h)

Affected

JW

L5-PAINTL3-WLD

1 operator

OEE

+0.4 pp

Throughput

+14u

Cost

spend

Confidence97%

Overtime ApprovedFri · 14:00–18:00Backlog-Burn v2

Extend Day shift +2h on L1-FA — Friday WO push

TriggerWO-22418 due Fri 18:00; current pace yields 91% completion

ImpactCloses the 9% gap, keeps Toyota TS commitment

ModelBacklog-burn · WO sequencing

Affected

MJ

AD

CR

AH

L1-FAL1-FA

4 operators

OEE

+0.6 pp

Throughput

+96u

Cost

-$1.8k

save

Confidence88%

Cross-train deploy Awaiting approvalTomorrow swingCross-Train Optimizer

Pair mentor: certified L2-FA expert shadows L5-PAINT booth changeover

Trigger2 trainees on L5 changeover; mentor improves SMED time 22% historically

ImpactCuts changeover ~7min, lowers internal-stop split

ModelCross-train optimizer · mentor history

Affected

AH

L2-FAL5-PAINT

1 operator

OEE

+0.5 pp

Throughput

+12u

Cost

spend

Confidence71%

Predictive call-in Awaiting approvalThu · B · SwingPdM-Aware v1

Call-in Rafael Lopez for L2-FA — Garcia PTO + scheduled PM converge

TriggerPM window collides with scheduled PTO on B shift

ImpactPrevents L2-FA from dropping below plan headcount

ModelPdM-aware · roster intersection

Affected

RL

Roster overflowL2-FA

1 operator

OEE

+1.4 pp

Throughput

+38u

Cost

-$480

save

Confidence83%

Crew rotation DeclinedEarlier todayDemand-Match v3

Move Andre Harris from L1-FA to L4-STMP for hour

TriggerMarginal balance gain; supervisor flagged operator coaching session in progress

ImpactWithdrawn — coaching session preserved

ModelDemand-match · superseded by supervisor override

Affected

AH

L1-FAL4-STMP

1 operator

OEE

+0.2 pp

Throughput

+6u

Cost

spend

Confidence62%

Crew utilization · now

Lines × Shifts heatmap

Planned headcount vs assigned. Click a cell to see the operator list.

UnderBalancedOver

DAY

SWG

NGT

L1-FAFinal Assembly

L2-FAFinal Assembly

L3-WLDBody Weld

L4-STMPStamping

L5-PAINTPaint Booth

L6-CNCCNC Machining

Cross-train deployment

Underused 2nd-line certifications

Operators certified on a secondary line but spending <20% of hours there.

Operator	Cert · yrs	Primary	Avail. shift	→ Deploy	Uplift
IJ Ivan JamesUtil 11% on 2nd	4.5y	L1-FA	DAY	L4-STMP	+3.7 pp
GP George PhillipsUtil 18% on 2nd	4.3y	L2-FA	SWG	L3-WLD	+3.6 pp
AR Adam ReedUtil 19% on 2nd	4.1y	L6-CNC	NGT	L1-FA	+3.5 pp
MJ Marvin JamesUtil 5% on 2nd	3.6y	L6-CNC	NGT	L5-PAINT	+2.7 pp
JK Jaspreet KellyUtil 17% on 2nd	3.6y	L5-PAINT	SWG	L6-CNC	+3.6 pp
AM Andre MartinUtil 6% on 2nd	3.5y	L6-CNC	NGT	L3-WLD	+3.9 pp
JJ Jonathan JamesUtil 8% on 2nd	3.5y	L1-FA	DAY	L6-CNC	+3.8 pp
PG Paul GriffinUtil 14% on 2nd	3.5y	L4-STMP	DAY	L6-CNC	+3.7 pp

7-day backlog burn-down

Scheduled WO vs capacity · with recommended OT

If all OT recommendations are accepted, capacity meets schedule by Saturday.

14-day shift gap

Forecasted operator shortfall by line

Driven by PTO, training windows and scheduled PMs. Negative = over-staffed.

W

T

F

S

S

M

T

W

T

F

S

S

M

T

Σ

L1-FA

·

·

−1

−1

−1

−1

−1

·

+1

+1

+1

+2

+1

+1

+2L2-FA

+1

·

−1

−2

−2

−1

−1

−1

·

+1

+1

+2

+2

+2

+1L3-WLD

−1

−1

−2

−2

−3

−2

−2

−1

−1

·

·

+1

+1

+1

−12L4-STMP

·

−1

−1

−1

−1

−1

−1

·

+1

·

+1

+2

+2

·

·L5-PAINT

+1

·

−1

−1

−2

−1

−1

·

·

+1

+2

+2

+1

+2

+3L6-CNC

·

−1

−2

−2

−3

−3

−1

−1

−1

·

·

·

·

·

−14

ShortfallSurplus

Drivers: PTO · training · scheduled PMs

Feedback loop · last 7 days

Acceptance, rejection, and realized vs projected impact

MAIA grades itself: realized OEE uplift on accepted recommendations vs the model's projection. Drift > 0.5pp triggers re-training.

Crew rotation

89%64/72 accepted

Realized

+2.1 pp

Projected

+2.4 pp

Drift -0.3pp · model in tolerance

Overtime

81%22/27 accepted

Realized

+1.6 pp

Projected

+1.8 pp

Drift -0.2pp · model in tolerance

Cross-train deploy

82%14/17 accepted

Realized

+1.9 pp

Projected

+2.2 pp

Drift -0.3pp · model in tolerance

Predictive call-in

82%9/11 accepted

Realized

+1.1 pp

Projected

+1.3 pp

Drift -0.2pp · model in tolerance

Why these recommendations?

The three models behind the queue

Every card on this page is the output of a model that consumes upstream signals. Click through to inspect the source intelligence.

Demand-Match v3

Reads the live demand curve from /forecasting and the per-line OEE trace from /performance-intelligence. Proposes rotations when the marginal OEE gain on the receiving line exceeds the loss on the source line by ≥1pp.

Backlog-Burn v2

Sequences open work-orders against current capacity and scheduled changeovers. Proposes OT when the cheapest premium-hour beats the cost of a missed customer commitment.

PdM-Aware v1 + Cross-Train Optimizer

Predictive maintenance windows from CMMS combine with the operator certification matrix to pre-stage maintainers and pull cross-trained operators where bench depth exists.

MES · CMMS · L&D · Roster · Demand-Forecast · Last-30d outcomes

See demand forecast See OEE intelligence