Field ops and dispatch, run for you · Freight, last-mile, and field logistics
We run dispatch and field operations. Exceptions resolved before the call.
Antino builds an operations brain — your routes, carriers, and SLAs as an executable graph — and runs dispatch, multi-carrier workflow, and exception handling on it. The smooth shipment moves itself. The stuck one finds an owner.
Below: the whole pipeline — playable
01 · The problem
Operations is a control tower that runs on phone calls.
Dispatch is a dispatcher's intuition. Carrier selection is whoever picked up. When a shipment is late, lost, or rejected, the team finds out from an angry customer and starts working the phones. The knowledge of how to clear an exception lives with the senior ops lead — and walks out the door at 6pm.
02 · How we'd run it
The ops brain: every case in, two ways out.
Intelligent decisions. Zero busywork. Every exception, owned.
Shipments in
Every order. Every carrier.
Orders & AWBs
From every sales channel
Carrier rules & rates
Cutoffs, lanes, SLAs
Live tracking events
Scans, delays, attempts
Exception playbooks
What to do when it breaks
The ops brain
Dispatch. Track. Resolve.
- Carrier selection
- Route & cutoff rules
- SLA watch
- Delay prediction
- Exception triage
- RTO prevention
every decision logged · every outcome written back
Straight-through
~75%Dispatched, tracked, delivered — nobody touched it.
Exception
~25%Routed to a ops owner
Complex cases. Human judgment — with the case pre-packaged:
Executable playbooks
Tribal fixes become rules
Auditable by design
Every dispatch decision logged
Continuous learning
Each exception improves routing
Carrier-agnostic
Rules survive carrier churn
Scales by default
Peak season without panic
Shares are illustrative of a typical mature lane mix.
03 · The pipeline, end to end
The science of shipping, stage by stage.
Nine stages from order to write-back — the two halves of the promise: the smooth shipment moves itself, the stuck one finds its owner first. Drag the agentic-depth dial to watch the control tower go quiet stage by stage, and hover any stage for a worked example. The customer-impact call never automates.
Agentic depth — drag it
0%Dispatch
Ops brain · dispatch0%stage 01 · dossier
Order intake
Orders and AWBs from every sales channel are normalized into one stream with SLA and service level attached.
- Ingests and normalizes orders from every channel into one queue
- Attaches SLA, service level and special-handling flags to each shipment
Marketplace, D2C, B2B — one queue
Three channels with three formats land as one normalized stream. The 6 am batch doesn't wait for the person who knows which spreadsheet maps to which carrier portal.
Orders re-keyed from five channelsOne normalized stream, any source
click the stage to collapse
stage 02 · dossier
Carrier selection
Each shipment is matched to a carrier by coverage, cost, capacity and real performance on that specific lane.
- Scores every eligible carrier on the lane: cost, capacity, cutoffs, performance
- Books the best fit and logs the reasoning behind the choice
- Ops sets the rules of the game — SLA priorities, carrier mix floors.
The dispatcher's intuition, written down
The pattern from running multi-carrier field operations: the senior dispatcher's lane knowledge is encoded into selection rules — so it runs at 6 am, on his day off, exactly as he would have.
Whoever answered the phoneScored on the lane's actual history
stage 03 · dossier
Dispatch & booking
The booking is placed, labels and manifests generate, and pickup is scheduled inside the carrier's cutoff.
- Places bookings and generates labels and manifests automatically
- Sequences pickups against each carrier's cutoffs and capacity
The 4:55 pm cutoff, made every day
The same-day lane's cutoff stops being a daily sprint: bookings sequence themselves against it, and the shipment that can't make it is rebooked before anyone had to notice.
Portals, labels and cutoff panicBooked and labelled before the cutoff
stage 04 · dossier
Pickup watch
First-scan is watched per pickup window; a failed pickup triggers its playbook immediately, not tomorrow.
- Watches for the first scan against the promised pickup window
- Fires the reattempt or alternate-carrier playbook on a miss
The pickup that failed quietly — almost
No scan by window close. The reattempt books itself for the next slot and the shipper is notified with the new pickup time — before the warehouse closed, not after.
Discovered missing the next morningReattempt fired within the hour
stage 01 · dossier
Order intake
agentOrders and AWBs from every sales channel are normalized into one stream with SLA and service level attached.
- Ingests and normalizes orders from every channel into one queue
- Attaches SLA, service level and special-handling flags to each shipment
Marketplace, D2C, B2B — one queue
Three channels with three formats land as one normalized stream. The 6 am batch doesn't wait for the person who knows which spreadsheet maps to which carrier portal.
Orders re-keyed from five channelsOne normalized stream, any source
click the stage to collapse
stage 02 · dossier
Carrier selection
agenthumanEach shipment is matched to a carrier by coverage, cost, capacity and real performance on that specific lane.
- Scores every eligible carrier on the lane: cost, capacity, cutoffs, performance
- Books the best fit and logs the reasoning behind the choice
- Ops sets the rules of the game — SLA priorities, carrier mix floors.
The dispatcher's intuition, written down
The pattern from running multi-carrier field operations: the senior dispatcher's lane knowledge is encoded into selection rules — so it runs at 6 am, on his day off, exactly as he would have.
Whoever answered the phoneScored on the lane's actual history
stage 03 · dossier
Dispatch & booking
agentThe booking is placed, labels and manifests generate, and pickup is scheduled inside the carrier's cutoff.
- Places bookings and generates labels and manifests automatically
- Sequences pickups against each carrier's cutoffs and capacity
The 4:55 pm cutoff, made every day
The same-day lane's cutoff stops being a daily sprint: bookings sequence themselves against it, and the shipment that can't make it is rebooked before anyone had to notice.
Portals, labels and cutoff panicBooked and labelled before the cutoff
stage 04 · dossier
Pickup watch
agentFirst-scan is watched per pickup window; a failed pickup triggers its playbook immediately, not tomorrow.
- Watches for the first scan against the promised pickup window
- Fires the reattempt or alternate-carrier playbook on a miss
The pickup that failed quietly — almost
No scan by window close. The reattempt books itself for the next slot and the shipper is notified with the new pickup time — before the warehouse closed, not after.
Discovered missing the next morningReattempt fired within the hour
Watch & resolve
Ops brain · resolve & learn0%stage 05 · dossier
In-transit watch
Events stream in and match the lane's expected pattern. The on-track shipment consumes zero human attention.
- Matches every scan against the lane's expected event pattern
- Keeps the on-track shipment operationally silent — no alerts, no noise
The shipment nobody watched, correctly
Pickup scan, hub scan, line-haul — all on pattern. No WhatsApp group, no dashboard vigil. The absence of attention is the feature, and it's verified, not hoped for.
Dashboards watched by tired eyesSilence, verified scan by scan
stage 06 · dossier
Delay prediction
Scan gaps and hub dwell beyond lane thresholds predict the delay before anyone outside the system knows.
- Detects scan gaps and dwell beyond the lane's learned thresholds
- Opens the exception with impact assessed: SLA at risk, customer promise, options
Nine minutes of head start
Hub dwell crosses the lane threshold at 14:55. By the time a person could have noticed, the exception is already scoped — SLA exposure, reroute options, customer promise — and the customer is none the wiser.
Reported by an angry customerPredicted from a scan gap
stage 07 · dossier
Exception resolution
The encoded playbook fires: reattempt window, alternate carrier, rebooking, customer notification with the new ETA.
- Evaluates the playbook: reattempt, reroute, alternate carrier with lane capacity
- Rebooks and notifies the customer with the revised ETA
- Ops reviews playbook outcomes weekly and edits the plays.
Rerouted before the first phone call
An alternate carrier has capacity on the lane. The brain rebooks, the customer gets the new ETA, and what used to be a phone marathon is one decision with a log line.
Forty minutes of phone callsOne logged reroute decision
stage 08 · dossier
Customer-impact escalation
Address disputes, damage claims, angry-customer moments — the cases no rule should resolve go to a named owner, case built.
- Builds the case: full history, the customer's messages, the playbook's viable options
- A named ops owner makes the customer-impact call. This gate is permanent.
Priya decides in ninety seconds
An address dispute lands on her desk with the timeline, the customer's messages and two viable plays attached. The decision takes her ninety seconds; the case-building took her none.
The mess finds whoever's aroundA named owner, options attached
stage 09 · dossier
Carrier write-back
Every delay, reroute and resolution writes back to lane-level carrier performance and reshapes allocation.
- Updates lane-level carrier performance with every outcome
- Shifts next week's allocation away from the carrier that slipped
- Carrier relationships and contract calls stay with the ops lead.
The lane the carrier quietly lost
Today's failure becomes tonight's data. Next week's dispatch shifts share on that lane — no review meeting, no escalation deck, just a network getting better at being a network.
A quarterly review deckNext week's dispatch already knows
stage 05 · dossier
In-transit watch
agentEvents stream in and match the lane's expected pattern. The on-track shipment consumes zero human attention.
- Matches every scan against the lane's expected event pattern
- Keeps the on-track shipment operationally silent — no alerts, no noise
The shipment nobody watched, correctly
Pickup scan, hub scan, line-haul — all on pattern. No WhatsApp group, no dashboard vigil. The absence of attention is the feature, and it's verified, not hoped for.
Dashboards watched by tired eyesSilence, verified scan by scan
stage 06 · dossier
Delay prediction
agentScan gaps and hub dwell beyond lane thresholds predict the delay before anyone outside the system knows.
- Detects scan gaps and dwell beyond the lane's learned thresholds
- Opens the exception with impact assessed: SLA at risk, customer promise, options
Nine minutes of head start
Hub dwell crosses the lane threshold at 14:55. By the time a person could have noticed, the exception is already scoped — SLA exposure, reroute options, customer promise — and the customer is none the wiser.
Reported by an angry customerPredicted from a scan gap
stage 07 · dossier
Exception resolution
agenthumanThe encoded playbook fires: reattempt window, alternate carrier, rebooking, customer notification with the new ETA.
- Evaluates the playbook: reattempt, reroute, alternate carrier with lane capacity
- Rebooks and notifies the customer with the revised ETA
- Ops reviews playbook outcomes weekly and edits the plays.
Rerouted before the first phone call
An alternate carrier has capacity on the lane. The brain rebooks, the customer gets the new ETA, and what used to be a phone marathon is one decision with a log line.
Forty minutes of phone callsOne logged reroute decision
stage 08 · dossier
Customer-impact escalation
agenthuman ownsAddress disputes, damage claims, angry-customer moments — the cases no rule should resolve go to a named owner, case built.
- Builds the case: full history, the customer's messages, the playbook's viable options
- A named ops owner makes the customer-impact call. This gate is permanent.
Priya decides in ninety seconds
An address dispute lands on her desk with the timeline, the customer's messages and two viable plays attached. The decision takes her ninety seconds; the case-building took her none.
The mess finds whoever's aroundA named owner, options attached
stage 09 · dossier
Carrier write-back
agenthumanEvery delay, reroute and resolution writes back to lane-level carrier performance and reshapes allocation.
- Updates lane-level carrier performance with every outcome
- Shifts next week's allocation away from the carrier that slipped
- Carrier relationships and contract calls stay with the ops lead.
The lane the carrier quietly lost
Today's failure becomes tonight's data. Next week's dispatch shifts share on that lane — no review meeting, no escalation deck, just a network getting better at being a network.
A quarterly review deckNext week's dispatch already knows
Depth bars are illustrative of a mature lane mix. The customer-impact call is a permanent human gate — the brain builds the case and holds the options; a named owner decides.
04 · One brain, two engines
How the function runs on top of the brain.
Revenue OS
Dispatch and allocateWe run dispatch on the operations brain. Each shipment is matched to the right carrier and route by cost, capacity, and SLA — automatically, and consistently, instead of by whoever's on the phone.
- Carrier and route selection by cost, capacity, and lane performance
- Multi-carrier workflow orchestrated as one executable allocation engine
- Dispatch decisions made consistently, with the reasoning logged
- Capacity and cost trade-offs surfaced before commit, not after
Delivery OS
Track and resolveWe run exception handling on the live operations brain. The on-track shipment flows untouched; a failed pickup, delay, or rejection triggers its playbook and lands on a named owner with the next action ready.
- Exceptions detected from field and carrier signals before the customer calls
- Failed pickup, reattempt, and RTO workflows triggered automatically
- Each stuck shipment routed to an owner with the playbook step attached
- Status and reasons answered from the graph, not from a dispatcher's memory
05 · What actually changes
The same function, before and after the brain.
06 · The outcome
Figures are illustrative unless tied to a named proof engagement.
07 · Live reference · Porter / SpeedBox
Field operations and multi-carrier logistics workflows built and run with Antino.
More use cases
The same brain, a different function.
09 · Run a function
Stop renting hours. Start running functions.
Pick the function you want off your plate. We'll map the brain and name the outcome we'd commit to — before you do.