The partner chapter: the network stopped running on memory.

First, what the numbers are and aren’t

An honest note before the story: the quantified, published results from this deployment are quoting-side — the win rate that moved from 35% to 73%, documented at aiquotedesk.com and freigh.ai. The partner chapter has no metrics table, and we won’t invent one: there are no published partner-side figures from this deployment, and partner-RFQ benchmarks don’t exist anywhere in the industry yet. What this page tells is the operational story — what changed on the desk when finding, asking and remembering partners stopped being manual work.

The desk before

A growing forwarder’s classic partner picture: a customer asks for a lane nobody serves, and the search starts from memory — who do we know out there, who handled something like this, whose card is still in the drawer from the last conference. RFQs go out one email at a time; replies come back as a PDF here, a WhatsApp message there, two silences; the comparison lives in Excel, assembled while the enquiry cools. And the knowledge of which partner actually performs — who quotes fast, who executes, who pays — lives in the heads of whoever handled the last files. Nothing broken; everything bounded by what one desk can remember.

What the partner desk changed

The RFQ round, as it runs now

When a lane request lands, the desk decomposes the route and builds the shortlist before anyone opens a directory — candidates drawn from Infinity’s network memberships and its own shipment history, ranked on the record rather than on recall. The RFQs it dispatches are structured — lane, cargo, service, validity requested, deadline stated — so that the replies come back comparable, whatever format they arrive in. The comparison assembles itself as responses land, and every timestamp quietly becomes scorecard data. The team’s job in the round is the part that was always theirs: choosing.

Scorecards that accumulate instead of fade

The compounding change is the memory. Every RFQ round leaves a residue — who answered in hours, who answered in days, whose rate won, whose shipment ran clean, whose settlement needed chasing — and the desk keeps all of it, per partner, per lane. Twelve months in, that residue is a record no conference impression can argue with: renewal conversations start from the card, not from whoever remembers the partner most fondly. The method behind those cards is laid out in the scorecard guide.

The other chapters

Six agents run at Infinity — quotation, booking, tracking, partner discovery, reconciliation, collections. The full platform story is at freigh.ai; the quoting chapter with its measured 35%→73% arc is at aiquotedesk.com; the booking chapter is at bookingdesk-ai.com. This chapter is where the partners behind those quotes and bookings get found, asked and — for the first time — remembered.