A discovery brief from Enmovil for
PepsiCo × Site Operations

Two WMS.
One truck.

Zero precedent in sixty years.

PepsiCo started co-loading Frito-Lay snacks and PBNA beverages on the same Texas trucks this year — for the first time in sixty years. Two BUs. Two pick systems. One outbound dock. One OTIF scorecard at Walmart. A hypothesis for the site director coordinating the merge.

(Signal: PepsiCo co-loading snacks and beverages in Texas, first time in sixty years of separate BU distribution. San Antonio Site Operations Director, Sep 2024.)

The co-loading execution loop

What we've read.

Two systems. One dock window. One truck. The gap lives between them.

Snacks WMS Pick queue · separate order cycle PBNA Beverages WMS Pick queue · different SKU profile ? Co-load dock who sequences when one BU lags? Single truck snacks + beverages, one departure Texas lanes Retailer OTIF scorecard one score — both BUs Walmart 98/95 · Target OTFR · Kroger one late BU tanks both
Frito-Lay WMS Snacks pick queue. Separate order cycle, separate SKU profile. Has never shared an outbound dock with PBNA.
PBNA WMS Beverages pick queue. Different weight, cube, and temperature profile from snacks. Different planning cadence.
Co-load dock [?] Who decides sequencing when one BU's pick finishes before the other? Neither WMS was built to answer that question.
Retailer OTIF Walmart, Target, Kroger score the combined truck. Frito-Lay on time + PBNA late = one miss on the scorecard.
The coordination gap

The gap.

Not a pick-efficiency problem. A coordination problem. Two systems, one truck, no shared language at the dock.

Hypothesis: the co-loading execution gap is not algorithmic — it's structural. Frito-Lay WMS and PBNA WMS each optimize for their own BU's fill rate and departure window. Neither system sees the other's state. The site director absorbs the mismatch manually, every shift. Enmovil's hypothesis: a cross-BU execution view that closes the dock coordination gap before it hits the retailer scorecard.
EDP (Site Operations)

% of co-loaded trucks departing on-time with both Frito-Lay and PBNA pick complete, measured per shift, per retailer lane, per week.

Numerator: trucks where both BUs at pick completion before departure commit.
Denominator: all co-loaded outbound trucks from the site.

If this number doesn't climb, co-loading is a coordination tax on the site director's day — not an operational improvement.

60 yrs
Frito-Lay and PBNA operated completely separate outbound lanes. This is new for everyone.
2 WMS
Two pick systems now sharing one outbound dock. Neither sees the other's completion signal.
98/95
Walmart OTIF threshold: 98% in-full, 95% on-time. One BU's miss = combined truck's miss.
TX first
Texas trucks are the live pilot. Coordination model built in real time, under retailer scorecard pressure.

This is a hypothesis, not a characterization. We'd love to test it against your site's actual coordination cadence.

Three factors that make this acute

Why now.

Co-loading is live. Texas is the live pilot. The playbook is being written this year.

01

The sixty-year window

First operational integration between Frito-Lay and PBNA distribution in six decades. No inherited playbook.

PepsiCo co-loading initiative, Texas, 2025.

02

Your site, your problem

San Antonio. Site Operations Director. The coordination gap lives in your shift, not in a strategy deck.

San Antonio site, Frito-Lay + PBNA co-load lanes, 2025.

03

Retailer scorecard

Walmart's OTIF window doesn't separate Frito-Lay late from PBNA late. The combined truck either makes it or it doesn't.

Walmart 98% IF / 95% OT · Target OTFR · Kroger ORAD.

If this is half-right, it's worth a working session before Dallas.

Where fit could live

The leverage.

Co-loading created a new execution surface. Neither WMS was designed for it. A cross-BU coordination layer is where we'd test fit.

Outside-in read

Two pick queues. One dock window. One departure. Neither WMS sees the other's state.

What "coordinated" would look like

  • Dock sequence respects both BUs' pick completion — not just whichever finished first.
  • Truck departure commit visible to both planning systems before the outbound window closes.
  • Exception — one BU lagging — routed to the right operator before the shift absorbs the delay.

The new execution surface

Texas co-load lanes → 40+ PFNA sites if it works here.

The coordination model built in San Antonio becomes the template.

Frito-Lay pick → dock ready
Snacks complete signal
PBNA pick → dock ready
Beverages complete signal
Walmart
98 IF / 95 OT · 3% CB
Target
OTFR · 5% COGS penalty
Kroger
ORAD · $500/order
If co-loading coordination is solved in San Antonio → it's the playbook for every co-load lane PepsiCo opens next.
One platform, every decision layer

What Enmovil is.

The intelligence layer that sits above SAP and both WMS systems, not in place of them.

PREDICT

Available for expansion

  • Forecast & Demand Intelligence
  • Inventory Planning & Optimisation
  • Scenario Modelling · S&OP
  • Warehouse & Node Operations

PLAN

Available for expansion

  • Dispatch & Route Optimisation
  • 3D Load Building
  • Network & Capacity Planning
  • Carrier Allocation

EXECUTE

Best-fit wedge for this conversation

  • Cross-BU Dock Coordination · Yard Scheduling
  • Control Tower & Resilience
  • Multimodal Orchestration (OTR · Rail · Ocean · Air)
  • ePOD & Freight Settlement
CADDIE AI: the cross-suite intelligence layer spanning Predict, Plan, and Execute. Sits above SAP S/4HANA, the Frito-Lay WMS, and the PBNA WMS — reads both pick queues, surfaces the coordination gap before it becomes a departure miss. Conversational. Agentic. No rip and replace.
25+ Fortune 100 customers · 150+ team · 10+ years · ISO 27001 · SOC 2 Type II · SAP-ready · AI-native
Proof-by-geometry

How we've helped a peer.

A Fortune 500 SAP-native manufacturer. Plant to DC to retailer at scale. Selected over SAP, Kinaxis, EY, Blue Yonder.

Before

40-minute reports. 16-hour overnight forecast.

  • Reports took 40 minutes per query, one at a time.
  • Forecast jobs ran 16 hours overnight, refreshed once daily.
  • SAP extract → spreadsheet → email → rework, three to four times per decision.

After

90-minute SAP-to-decision loop.

  • Reports down to 5–6 minutes, writeback-ready.
  • Forecast jobs retired. Agentic planning on demand.
  • Single Command Hub. One KPI view across plants.

90 min

SAP extract to decision to writeback.

6× faster

Reports 40 min → 5–6 min.

16 hrs

Overnight forecast job retired.

Selected

Over SAP, Kinaxis, EY, Blue Yonder.

Different product category. The geometry of a cross-system execution gap translates. Percentages to be validated together.
The stack view

CADDIE AI above both WMS, not in place of them.

The cross-BU intelligence layer that sees both pick queues and surfaces the dock coordination decision before the departure window closes.

CADDIE AI · conversational · agentic · cross-BU · cross-suite
Predict
Plan
Execute
SAP S/4HANA · Frito-Lay WMS · PBNA WMS · yard systems · retailer EDI
Co-load dock · pick completion signals · departure commit · OTIF feed, real-time
No rip and replace. Frito-Lay WMS stays. PBNA WMS stays. CADDIE AI queries across both in your language: "show me every co-loaded truck this week where PBNA pick closed more than 15 minutes after Frito-Lay pick, and which retailer lane absorbed the gap."
For a longer conversation

Questions we'd love to explore together.

What we'd want to understand about co-loading coordination if we had 60 minutes.

The dock coordination mechanism

  • How does your site signal dock-ready when both Frito-Lay and PBNA haven't completed pick? Is there a manual step in that handoff today?
  • When one BU's pick finishes before the other, who makes the "hold the truck or ship partial" call — the site director, the BU ops lead, or the WMS?

Scorecard accountability

  • When a co-loaded truck misses Walmart's OTIF window, how do you determine which BU's pick sequence was root cause?
  • Does the retailer EDI currently distinguish between Frito-Lay items and PBNA items on the same shipment, or does it hit as a single PO?

Scale and cadence

  • How many co-loaded outbound lanes does San Antonio run per shift today? Is it growing?
  • What's the escalation path if PBNA pick is 20 minutes behind Frito-Lay during a peak promo week — and how often does that actually happen?

Forward

  • If San Antonio's co-load coordination were running cleanly — both BUs at pick completion before every departure commit — which decision downstream gets sharper first?
Who's building this

Enmovil · the technology thought partner for autonomous supply chains.

One AI-native platform. Demand sensing, inventory, dispatch, execution, freight settlement, sustainability reporting. All on one intelligence layer that sits above the systems you already own.

Under orchestration
~100K
Trucks per day across the customer base
Daily ETA accuracy
99%
Road, rail, ocean, air
Forecast accuracy
97%
Demand sensing on deployed customers
Logistics cost savings
8–15%
Measured across deployments
Integration
3–4wks
On existing SAP, Oracle, TMS, WMS. No migration.
Deployed at
Dispatch Planning
Multimodal Logistics
Multimodal Orchestration
Dispatch Planning
Fleet Management
Logistics Orchestration
Logistics Resilience
Inventory Management
Freight Settlement
Dispatch Planning
Transport Management
Export Planning
Runs under GDPR and SOC 2. Data ingestion via API, EDI, or bulk upload. Enterprise SSO. Deploys over existing SAP ECC 6.0 and above.
Ambition

If this works, every co-loaded truck in Texas departs on the right side of the retailer's OTIF window — both BUs coordinated, one execution view, no shift spent absorbing the mismatch manually.

The ask: a 60 to 90 minute working session with your team.

01

A closer look at co-load coordination

A walk-through of how we've seen cross-system execution gaps close in peer CPG and retail deployments. What maps to your co-load reality, what doesn't, and where the dock coordination decision surfaces first.

02

A stack read

A shared read of where CADDIE AI could sit relative to Frito-Lay WMS, PBNA WMS, and the retailer EDI layer — and which site-director decisions sharpen when both pick queues are visible in one place.

03

A co-drafted hypothesis

A one-page fit hypothesis naming the co-load coordination gap we'd test first, what "fit" looks like versus "doesn't fit" — something you can take back to the PFNA program leads.

Three questions before the session

When a co-loaded truck misses its OTIF window, where does root cause most often live?

For co-loading to work at scale across PFNA sites, what does your site need most?

For a 60 to 90 minute session to land, who needs to be in the room?

Predict.Plan.Execute.
Not a pitch. A working session. globalsales@enmovil.in enmovil.ai

CONFIDENTIAL · Enmovil × PepsiCo Site Operations discovery. Prepared for the American Supply Chain Summit 2026, Dallas. Third-party trademarks belong to their respective owners. Not affiliated with or endorsed by PepsiCo.