How nexPro actually
screens, scores, and negotiates
Not a black box. Every stage below is the real pipeline running behind your comparisons — what it checks, what it refuses to guess, and where you stay in control. Through to the order, the invoice, and what nexAI flags along the way.
Five stages, one RFQ
Every bid on your RFQ passes through all five before you ever see a recommendation. Hover a stage to pause it; click to jump to the detail below.
Reduced-motion? The ring holds still — every stage is still clickable below.
The pipeline, screen by screen
An animated walkthrough of the actual interface and its real language — not a recreation of the idea, the screens themselves. The five evaluation stages, then what nexAI flags and what happens after the award.
Quotes are in 2 currencies (INR, USD)check first
Gamma is 62% below the medianverify scope
Dispatched 12 days ago; nobody has bidchase
Every invited supplier has answeredno wait
What each stage actually does — and what happens after
The plain-language version of the pipeline, plus the sixth step the other five exist for: the order the supplier actually delivers against. No step here is decorative; each one exists because skipping it caused a real problem somewhere.
Screening
Every bid is checked against your hard limits — certifications, minimum specs, mandatory terms — before a single score is calculated. A bid that breaches one is set aside here, permanently. It never reaches the scoring stage, which means it can never buy its way back in on price.
Evaluation brief
Before any score is visible, you write down what actually matters for this purchase — in your own words, plus the real cost parameters (operating hours, energy tariff, expected life) that turn a quoted price into what the equipment will actually cost you to run. Recorded now, so the basis of the decision is defensible later — and never edited retroactively to fit a preferred outcome.
Comparison
Every surviving bid is scored against your weighted criteria — and every number is shown, not hidden inside a single black-box total. A second, independently-computed ranking method cross-checks the result and flags it when the two disagree. Statistically unusual bids (suspiciously cheap, implausibly fast) are flagged for a second look, not silently trusted.
Negotiations to consider
For every runner-up, nexPro calculates the exact number that would let them overtake the leader — a real figure to open a conversation with, not a guess. It tells you plainly whether the leader's win is genuinely contestable or effectively uncontested, and drafts the actual outgoing email — reviewable and editable — before you ever send anything.
Award
nexPro recommends; it never decides. The recommendation, the full score breakdown, and every negotiation exchange sit together as one auditable record — so if a supplier ever asks "why them and not us," the answer already exists.
Order, delivery and invoice
An award is not the end of the job. The moment you award, the winning supplier gets an order in their portal — they move it through Processing, Shipped and Delivered, and raise their invoice against it. You see the invoice, approve it or reject it with a reason, and mark it paid. No separate system, no re-keying, and no supplier left wondering whether the work they won actually exists anywhere.
It reads your own records and tells you what needs you today
Not a chat box bolted on the side. nexAI works out the things a buyer would otherwise have to notice by hand — and the findings are computed in code, not guessed by a language model, because a confident wrong answer about whose quote is cheapest is worse than no answer at all.
- ✓It refuses to compare the incomparableQuotes in two currencies, or covering two different scopes, are not a close race. You are told before a "cheapest" is named, not after.
- ✓It names the outlier instead of averaging it awayA quote far below the pack is usually a different quantity, not a bargain — so it says which, and by how much.
- ✓It notices silenceA tender dispatched days ago with no reply from anyone usually means the invitation never landed. That is a delivery problem, and it reads as one.
- ✓It recommends; it cannot actnexAI creates nothing, sends nothing and awards nothing. An award nobody can explain is one nobody can defend.
The terms you'll actually see on screen
Short, honest definitions of everything the product does — no jargon left unexplained.
Mandatory gate
A hard limit checked before scoring, not during. Breach it and the bid is set aside — price can never compensate for it.
Evaluation brief
Your priorities and real cost parameters, written down before any bid scores are visible — so the decision basis predates the decision.
Weighted score
Every criterion's raw value, normalized value, weight, and contribution — shown in full. Nothing folds into an unexplainable single number.
Cross-check ranking
A second, differently-built ranking method run on the same data. When it disagrees with the primary score, that disagreement is surfaced, not buried.
Anomaly flag
A bid that's a statistical outlier against its peers — often a suspiciously low price. Advisory only: it asks you to verify, never disqualifies automatically.
Negotiation lever
The exact concession — a price, a delivery date, a warranty term — that would close the gap to the current leader. Computed, not estimated.
Landed cost
What you actually pay, not what was quoted — tooling amortised, freight, expedite fees and delivery history folded in. It is routinely the number that reverses a ranking.
Comparability warning
Two quotes in different currencies, or covering different scopes, are not a close race. nexPro says so before naming a cheapest, and never converts at a rate nobody chose.
Order & invoice
An award becomes an order the supplier can see, move through delivery and invoice against — which you approve, reject with a reason, or mark paid.
Send invites
Invite suppliers by email from the account menu, several at once. They register themselves, fill in their own details, and their bids arrive in your workspace.
CAD preview
STEP, STL and IGES attached to the RFQ and converted for the browser, so a supplier can rotate the actual part before quoting on it.
Procurement memory
Every RFQ you have run stays searchable by part and revision — what it cost last time, what tooling you already paid for, what the lead time actually turned out to be.
Sealed audit trail
Criteria seal when invitations go out and the ranking seals at award. Who changed what, and when — so nobody can retro-fit the reasoning after the result is known.
Every negotiation opens with a number, not a hunch
This is the stage that turns a ranking into an actual outcome. It doesn't just tell you who's winning — it tells you exactly what it would take to change that, and drafts the email to go find out.
- ✓Contested vs. dominant, stated plainlyA clear read on whether the leader's win is genuinely contestable, so you know when a negotiation is worth your time.
- ✓The exact ask, not a guess"Ask for 23,000 off" or "ask for one more year of warranty" — the precise number that closes the gap.
- ✓Sustainability checks built inIf a vendor's already an outlier on the number you're about to push, you're told before you ask.
- ✓Review before send, alwaysThe email is drafted and fully editable. Nothing reaches a supplier on the strength of one click.
See it run on your own RFQ
Fifteen minutes, your real bids — walk through screening, scoring, and negotiation end to end.
Request a demo