A briefing from Enmovil for
Wedgewood Pharmacy

Forty six thousand formulations. Lot size of one.
Demand planning when the forecast unit is a patient.

What does S and OP look like when the forecast unit is one patient's prescription, not a SKU[1]? When 503A work is patient specific[2], 503B work is office use and regulated as cGMP[2], USP 795 default BUDs cap aqueous preparations at 35 days preserved and stability extensions at 180 days[6], FDA pulled injectable semaglutide off the shortage list in February 2025 and whipsawed compounder demand within weeks[3], and 20 percent of US API sits at 467 China registered sites with 21 percent more from India under routine Import Alert 66-40 scrutiny[8]? Caddie, Enmovil's agentic AI, sits above your ERP, your PIMS, your compounding lab schedule, and Blue Rabbit[16]. Closes the loop from API shipment to the patient's refill, with demand planning built for intermittent demand[15].

The situation

What does 2026 actually feel like inside the largest compounding pharmacy in the US, four years into a Partners Group era, scaling human compounding on top of a 46,000 formulation veterinary base?

Wedgewood operates five PCAB accredited state licensed pharmacies and two FDA registered cGMP 503B outsourcing facilities. A 46,000 plus formulary. More than 50,000 prescribing veterinarians. 70,000 plus clinics. Over one million animals served a year[1]. The US compounding pharmacy market sits at roughly 6.45 billion dollars in 2025 and 6.82 billion dollars in 2026[4]. Animal drug compounding alone is 1.68 billion dollars in 2025 and tracks toward 3.25 billion dollars by 2033[5]. The FDA removed injectable semaglutide from the shortage list on February 21, 2025. 503A compounders had until April 22, 2025 and 503B until May 22, 2025 to cease copies[3]. The China India API base is 20 percent plus 21 percent of US supply, with Import Alert 66-40 routinely detaining shipments on data integrity grounds[8]. Chandler AZ opened a 90,000 square foot facility in 2025 with 250 jobs planned by 2027[7]. NJMEP value stream mapping in 2024 targeted a 30 percent production time reduction[11]. So what does S and OP look like when most SKUs ship rarely, a few ship daily, and the forecast is really a lot size of one distribution?

46,000+formulations
Veterinary + human compounding. Species, weight, dosage, and palatability permutations. Most SKUs ship rarely. A few ship daily.[1]
5 + 2PCAB + 503B
Five state licensed pharmacies plus two FDA registered cGMP 503B outsourcing facilities. NJ, AZ, CA, CO footprint.[9]
14 / 35 / 180day BUD
USP 795 non preserved aqueous 14 days. Preserved 35 days. Stability extended caps at 180 days.[6]
Blue RabbitPIMS platform
PIMS integrated home delivery of compounded plus commercial drugs plus diets plus preventatives to vet practices. Launched 2024.[10]
$6.45BUS market 2025
US compounding pharmacy market. Animal drug compounding sub segment 1.68 billion 2025 to 3.25 billion 2033.[4]
6agents
Demand Sensing, API Sourcing, Plant and Line, QC and Release, Blue Rabbit Dispatch, Settlement. One loop.
Caddie
Enmovil's agentic AI. Sits above your ERP, your PIMS, your compounding lab schedule, and Blue Rabbit[16]. Closes the loop from API shipment to the patient refill. Demand planning built for intermittent demand[15], not rolling average. Typical go live is three to four weeks. ISO 27001 and AICPA SOC 2 Type II certified.
Three questions an informed observer would ask

Pick the one costing the most working capital or patient trust this fiscal. We will scope a two week experiment on it.

01 · The forecast unit question
How do you forecast when the forecast unit is one patient's prescription, not a SKU? Does Croston grade intermittent demand math run behind S and OP, or is the plan still driven by rolling average history on API families?
Croston's method and the SBA and TSB variants are the standard forecasting approach for intermittent demand with many zero periods[15]. Standard exponential smoothing fails at lot size of one shapes. Across 46,000 formulations, the demand by formulation is mostly zero across any given week, with a long tail of rare but critical fills. When Harry runs the plan, does the model see that shape, or is it averaging zeros into a mean that does not help Chris schedule the compounding lab.
Caddie answer: Demand Sensing agent reads the full distribution across 503A patient specific and 503B office use. Intermittent demand shapes get the math they need, not a rolling average.[16]
02 · The BUD vs batch question
When the preserved aqueous default BUD is 35 days and stability extended caps at 180 days, how do NJ and AZ teams decide batch size without locking working capital into inventory that will expire?
USP 795 default BUDs set a hard clock on compounded inventory[6]. Stability extension to 180 days requires a stability indicating assay, container closure testing, and antimicrobial effectiveness testing per preparation. The math is different for 503A make to order and 503B make to stock. Where does the line drift today, between batch size for lab efficiency and BUD exposure on the balance sheet, and who owns the call when the lab at Chandler is running hot and Swedesboro is not.
Caddie answer: Plant and Line agent joins 503A and 503B demand to NJ and AZ lab capacity under USP BUD constraints. Batch sizing reads expected demand against remaining shelf life.[16]
03 · The API to clinic signal question
A China API hit on Import Alert 66-40 moves through to an AZ production plan how fast? Does the same signal reach Harry's demand plan, Chris's line schedule, and the vet clinic PIMS order on Blue Rabbit in hours, or in the next weekly S and OP?
The FDA has 467 registered API sites in China, 20 percent of US supply, with another 21 percent in India. Import Alert 66-40 detained Global Calcium, Wuhu Nuowei, and Chengdu Innovation in 2025 alone[8]. A single detention ripples through to a 503A or 503B schedule inside days. If the signal lives in one buyer's inbox and not in the demand plan, the clinic learns last. Blue Rabbit is the place where the clinic feels it[10].
Caddie answer: API Sourcing agent reads detention signals, lead time drift, and alternate source coverage into one plan. Surfaces the switch or substitute to demand planning, production planning, and Blue Rabbit dispatch together.[16]
The operating layer · Caddie, Enmovil's AI co-pilot

Six agents. One continuous loop. Over your existing ERP, PIMS, and compounding lab schedule. No rip and replace.

A demand signal enters at Demand Sensing. It exits at Settlement. Autonomously. In minutes. Then the next one enters. The loop runs 24/7.[16]

Demand Sensing
Intermittent demand math across 46,000 formulations. 503A patient specific and 503B office use separated.
API Sourcing
China + India lead time drift. Import Alert 66-40 detentions. Alternate source coverage continuously, not quarterly.
Plant & Line
NJ, AZ, CA, CO compounding lab sequencing under USP 795 BUD constraints. Batch size reads expected demand.
QC & Release
PCAB plus USP 795, 797, 800 compliance gates. Lot genealogy back to API source. BUD clock starts on release.
Blue Rabbit Dispatch
PIMS integrated home delivery signal. Joins vet practice order to compounding schedule and BUD remaining.
Settlement & Audit
Parcel audit, accessorial drift, shelf life writedowns, freight settlement. Continuously, not quarterly.
Caddie, in Enmovil's own words: “Your AI co-pilot for supply chain orchestration. Unifies planning, logistics, and execution into one autonomous intelligence layer.”[16]
Operational impact · what changes in the first two quarters

No dollar sign promises. Six operating numbers you can measure.

Every number below comes from deployed customer outcomes on the Enmovil platform. Not projections. We show you the deployments. You pick the one closest to Wedgewood's shape.

Daily ETA accuracy
97%
Road, rail, ocean, air on the deployed customer base.
Source: Enmovil Caddie
Demand sensing accuracy
97%
Deployed customer benchmark. Long tail and intermittent demand shapes included.
Source: Enmovil deployed customer benchmarks
Exception response
4xfaster
4 to 6 hours compressed to under 30 minutes per incident on live deployments.
Source: Enmovil deployed customer benchmarks
Transportation spend
8 to 15% down
Measured across deployments. Parcel audit plus network rebalancing compounded.
Source: Enmovil deployed customer benchmarks
Integration
3 to 4weeks
Over existing ERP, PIMS, WMS, compounding lab schedule. API, EDI, or CSV. No migration.
Source: Enmovil deployment playbook
Planner minutes reclaimed
30 to 50%
Auto routed exceptions take clerking off planner plate. Validated on your data at the booth.
Baseline: Enmovil deployed customer benchmarks

What we are not promising on this page: a specific BUD writedown delta, a forecast accuracy number, or a fiscal year figure. Those are the things we think are true. We want to scope them with you in the room, not pre print them on a slide.

Where Enmovil has already solved this shape

Three customer stories that map onto Wedgewood's chain.

We do not have a US compounding pharmacy to name yet. We have a high SKU, low velocity aftermarket network, an FMCG raw material plus finished goods integrated plan, and a Fortune 500 SAP native manufacturer where the operating shape is close enough that the pattern transfers. Named references available under NDA at the booth.

Aftermarket spares network
Thousands of low velocity, high criticality SKUs with forecast shapes that are mostly zero.
The shape of a long tail where every SKU is critical to someone.
Shape
Commercial vehicle aftermarket network running thousands of low velocity, high criticality SKUs across FTL, LTL, and courier on one control tower with intermittent demand math underneath.
Outcome
Single orchestration layer. Forecast error on the long tail compressed. Planner minutes reclaimed at the branch level.
Relevance
Closest analog to Wedgewood's 46,000 formulation tail. Most SKUs ship rarely. A few daily. The forecast is a distribution, not a mean.
“Transfers to Wedgewood as: Demand Sensing reads the full distribution. Harry's plan sees the tail. Chris's schedule gets the shape of what is actually coming.”
FMCG one plan loop
Finished goods and raw material on one integrated planning loop.
The shape of API, WIP, and finished Rx on one cadence.
Shape
FMCG brand running finished goods plus raw material on one integrated plan. Two cadences closed into one.
Outcome
Planners write one plan, not three. Working capital compressed. Forecast error on the tail drops.
Relevance
Matches the Wedgewood problem where API inbound, lab WIP, and patient facing Rx should sit on one forecast, not three plans per facility.
“Transfers to Wedgewood as: API Sourcing, Plant and Line, and Blue Rabbit Dispatch see one forecast. NJ and AZ stop running parallel plans.”
Fortune 500 SAP-native
Manufacturer with SAP as ERP. 90 minute extract to decision to writeback cycle.
The shape of Caddie living on top of the ERP without migrating the ERP.
Shape
SAP native manufacturer ran a 16 hour overnight forecast job and a 40 minute planner report cycle before Caddie.
Outcome
Overnight job retired. Reports now 5 minutes. 90 minute extract to decision to writeback. Selected over SAP, Kinaxis, EY, and Blue Yonder.
Relevance
Wedgewood's transformation is people, processes, and systems. The ERP does not get torn out to chase demand sensing. Caddie sits on top.
“Transfers to Wedgewood as: the daily S and OP screen becomes one pane of glass. The overnight job goes away. No rip and replace.”
Scale
~100,000 trucks per day under orchestration across the deployed customer base. 97 percent daily ETA accuracy. 97 percent demand sensing forecast accuracy. Selected over Blue Yonder, Manhattan, Kinaxis, o9, Oracle, and EY in tier 1 competitive evaluations.
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 to 15%
Measured across deployments
Integration
3 to 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.
The ask

A thirty minute working session. Not a demo.

A thirty minute working session. Not a demo. We leave the room with three scoped experiments and a shared view of where next fiscal's operating margin actually sits, post restructuring. Three questions we want to ask before the session. Your answers tell us what experiments to bring.

Question 1 of 3
Across 46,000 formulations, what does the current forecast model actually do with the long tail?
Rolling average on API family, smoothed across zeros
Manual override by experienced planners on the known movers
Intermittent demand math on a subset of the tail
The tail is not forecast. Production reacts to orders.
Mix of the four. Depends on the therapeutic category.
Question 2 of 3
Where does BUD exposure on the balance sheet sit largest today?
503A patient specific, drifted into shelf for future refills
503B office use, made to stock ahead of soft demand
Duplicated buffer between NJ and AZ because cross visibility is partial
Stability extended preps that did not ship inside 180 days
Depends on therapeutic category and season
Question 3 of 3
If you could automate one workflow across demand planning, production planning, API sourcing, and Blue Rabbit in the next two quarters, what would you pick?
Intermittent demand forecast on the 46,000 formulation tail
Cross site NJ plus AZ compounding lab sequencing
API detention and lead time drift surfacing into demand and production together
Continuous BUD and shelf life risk scoring
Blue Rabbit delivery promise joined to compounding lab capacity

If you have already solved these, that is the conversation we most want to have. Harry is the spine. Bring one demand planner and one production planner on each coast, and one API buyer.

Book 30 minutes →
At the American Supply Chain Summit · Dallas · Booth [BOOTH_NUMBER]