Behind the scenes: how Vera autonomously resolves freight disputes
Vera has autonomously resolved over $25M in carrier and vendor financial disputes for Heineken across 70 countries. She does it without a human writing claim emails, without Excel, and without a freight-audit BPO — by running a closed-loop pipeline of invoice audit, evidence assembly, carrier negotiation, and settlement.
This is a deep-dive for heads of freight, logistics finance, and shared-services leaders in FMCG and 3PL businesses drowning in invoice dispute backlog.
Why we built this
Freight disputes are an industry-wide tax. Across global FMCG and 3PL operations, a meaningful share of freight spend is typically billed incorrectly — accessorials misapplied, fuel surcharges out of date, weight-break tiers miscoded, detention charges duplicated. Most shippers recover a fraction of that because dispute work is slow, painful, and requires a human to reconcile a line item to proof-of-delivery evidence across disconnected systems.
Freight-audit BPOs exist for this reason, but they are human-scaled, take months per dispute, and rarely touch more than a top slice of spend. Vera was built to close the entire loop autonomously, at full-spend coverage, in days not months.
How it works
Vera sits inside AgentFleet and connects to Nexa (settlement) upstream and to Atlas (control tower) for context. Her pipeline has four stages.
1. Invoice audit against rate card and execution reality
Nexa first normalizes the carrier invoice against the contracted rate card. Vera inherits those flags and goes further: she compares invoiced charges to actual shipment execution data inside Shipsy — the scans, the weights, the dwell times, the proof of delivery. If a detention charge is billed but the dwell-time data shows the truck was released on time, the line is flagged.
2. Evidence assembly
For every flagged line, Vera assembles the evidence package that a human auditor would otherwise spend hours compiling — timestamped POD, weight captured at origin and destination, geo-fenced arrival/departure events, matched rate-card clause, and the delta in currency terms. The package is machine-readable and human-reviewable.
3. Drafted claim and carrier negotiation
Vera drafts the claim in the carrier’s preferred format and submits it via the carrier’s portal, email, or EDI channel. When the carrier responds, Vera parses the response, matches it against her own evidence, and either accepts the counter-position, re-submits with stronger evidence, or escalates to a human if the dispute enters a judgement zone. Every exchange is logged and attributable.
4. Settlement and write-back
Once a dispute resolves, Vera writes the outcome back into Nexa, which updates the payable, closes the invoice, and reconciles against the expected cash position. The same resolution trains her for the next similar dispute on the same carrier.
Crucially, Vera operates under tolerance guardrails: the maximum single-dispute value she can close autonomously, the maximum recovery percentage she can accept, and the carrier-specific rules of engagement. Everything above the threshold routes to a human with the evidence pre-assembled.
Here’s the sequence at a glance:
Early results
Heineken: $25M+ in carrier and vendor financial disputes autonomously resolved across 70 countries via Vera — the anchor proof point for autonomous freight dispute resolution in FMCG. Alongside the recovery number, Heineken logged a 28% reduction in LSP excessive-stay payments, a 50% reduction in failed deliveries, and 70%→90% route settlement automation. Dispute cycle time compressed materially, and the human backlog that typically builds up in global shared-service centres shrank with it.
The same pattern plays out in 3PL businesses, where Vera disputes are carrier-outbound rather than shipper-inbound but mechanically identical.
What’s next
Vera is being extended into vendor-side disputes beyond freight — warehousing, detention, and accessorial pools — and agent-to-agent negotiation design partnerships are underway where Vera negotiates directly with a carrier’s own AI system, closing low-value disputes in minutes.