VendorFinder·AI

THE METHOD

Five moves between the route and the partner.

The lane arrives unserved; the choice that leaves must be defensible. The atlas method closes that gap — evidence end to end, with your judgment on every decision that matters.

01 · DECOMPOSE

The route becomes a job description.

A lane you can’t serve alone arrives as a city pair; the desk turns it into requirements. It splits the route into legs and names exactly what the missing partner must do — destination customs clearance, DO issuance, deconsolidation, last-mile delivery, the consol handled without drama. The search starts from what the job needs, not from a directory scrolled at random.

route → legs · customs / DO / delivery / consol named
02 · SHORTLIST

Ranked on evidence, not recall.

Candidates come from your network memberships and your own shipment history — the partners you already share a floor with, plus the ones your files have quietly measured. Each is ranked on what actually happened: past jobs together, response record, quote competitiveness, with payment-protection status noted per network. And the desk shows its reasoning line by line, so you can disagree with a specific claim instead of arguing with a black box.

memberships + history · protection noted · reasoning shown
03 · ASK

One structured question, two channels.

Structured RFQs go out where agents actually answer — email and WhatsApp — with the lane, cargo, service level, validity requested and your deadline stated plainly. The structure is the point: identical questions produce comparable answers. Non-responders get one polite chase before the deadline. Nobody gets hounded, and silence itself becomes a datapoint.

email + whatsapp · deadline stated · 1 polite chase
04 · COMPARE

Five formats become one table.

Replies land the way partners actually reply — a rate-sheet PDF, three lines in an email body, a WhatsApp message at midnight. The desk parses every one into a single normalized comparison: all-in totals on a common basis, validity and surcharges noted, gaps flagged rather than forgotten. The collation hour your team used to spend in Excel simply doesn’t happen.

pdf / email / whatsapp → one comparison · all-in normalized
05 · CHOOSE

You pick. The ledger remembers.

The decision stays human — you weigh the numbers against the relationship, the reciprocity, the history only you know, and choose. What changes is what happens next: every response time, every quote, every won and lost comparison quietly updates that partner’s scorecard. Next quarter’s shortlist starts smarter than this one, and renewal season stops being a memory contest.

human decision · RSP / WIN / OPS / PAY updated

THE FIRST 14 DAYS

DAY 1–3

We scope your network: memberships, partner list and shipment history imported — scorecards seeded from what your files already know, not from a blank page.

DAY 4–10

The desk shadows live enquiries — building shortlists and drafting RFQs while your team checks every ranking and every reasoning line it shows.

DAY 11–14

Live under approval: real RFQs out on email and WhatsApp, comparisons landing, with every outbound send signed off by your team first.