Two scenarios, built straight from your intro call. Pick one, enter the phone you want the AI to dial, and the agent places a real call with the Ofload allocator's context already loaded.
Before we jump into what to build, here's the landscape you walked us through on 7 April.
An asset-light model: Tier-1 shippers like Mondelez, Kellogg's, Bunnings and Allied Pinnacle on one side, a Long Tail of 6,000 independent carriers on the other. Six years old, on track for $340–360M this year. Growing 40% YoY, heading toward half a billion by end of next year. Demand is outstripping supply — the answer can't just be to keep hiring.
Allocator and transport manager teams in Sydney and Melbourne make hundreds of calls every afternoon — chasing ETAs when the platform goes stale, confirming PODs, reconciling pallets, and sourcing capacity on new lanes. Drivers often don't answer unknown numbers. Language barriers are common. Telematics on trucks isn't universal, so the platform isn't always a live source of truth.
JustCall for some of the team, Front for inbox, HubSpot Enterprise in parallel — a stack that's fragmented while consolidation decisions are still being made. 40+ staff moving off personal mobiles onto IVR now. Foundations being laid before the AI layer sits on top.
Growing a customer every week for six years has meant documenting processes is always the next thing, never this thing. Allocators have their own patterns, their own relationships. A CPTO-led AI strategy runs across the business — the CX piece is the puzzle that hasn't been prosecuted yet.
Four outbound calling scenarios came up on the call. Two of them are transactional — AI handles them end-to-end. Two are relationship- and negotiation-heavy — AI stays top-of-funnel and humans close.
Driver hasn't updated the platform. Need current location and ETA. Same question every call, data write at the end, shipper auto-notified if running late.
Fan out to 5–10 carriers on a new lane. Ask availability, rate, fuel levy. Capture responses. Hand the yeses to an allocator for the negotiation.
Relationship- and rate-heavy. AI stays top-of-funnel on discovery; humans handle the rate conversation and the booking.
That's the split we're proposing. AI picks up the 80% of dialling that frees your allocators to do the 20% that matters.
Pick one, enter a mobile, fire. The AI will call that number and speak first — Ofload context already loaded.
AI dials the driver, confirms they're on the right job, gets current location and ETA, writes it back to the platform. Auto-notifies Mondelez if it's more than 30 min late.
AI dials the top-ranked carrier on this lane, describes the load in one sentence, captures rate + fuel levy. Does not commit. Moves to the next carrier. Posts a summary to Bec when done.
Every action is logged. Every tool call is visible to Ofload ops after the call.
Calls getShipmentToChase or getCarriersForLane before the phone rings. AI knows the load, the allocator's name, and the history.
"Hi [name], it's Ofload about your [load] to [destination]." References Bec by name if the driver is suspicious.
Writes the ETA via updateShipmentETA, or logs carrier availability + rate via recordCarrierResponse. No "please update the app".
Late shipments auto-notify Mondelez. Capacity yeses get posted to Bec's Slack for negotiation. Escalates to a human if the driver raises safety, fatigue, or CoR.