For enterprises burned by vendors before

Vendor problems fixed on a clock. Vendor risk watched for you.

Escalate an issue, and a cure clock starts: 48 hours for a no-show, with a Trust Score consequence waiting behind it. Watch your bench, and receive alerts before problems become projects. All of it is underwritten by proctored AI-interview evidence, never self-declared claims.

800+empanelled vendor partners
70+cities
12,400+proctored assessments
0–100Trust Score, five pillars
Vendor trust cardLive
Nimbus Infoservices
GSTIN · 24AACxxxxx1ZP  ·  empanelled 2024
74 / 100
Trusted
PillarScore
Profile Authenticity86
Interview Integrity74
Candidate Calibre68
Fulfilment Reliability49
Compliance & Standing88
Weights 30 · 25 · 20 · 15 · 10Vendor card
The problem

Three vendors walk into your RFP. Their decks look identical.

Archetype 01

The polished low-calibre shop

Great sales motion, weak bench. You find out three sprints in, when velocity collapses and the “senior” engineers can’t clear a code review.

What the score seeslow Candidate Calibre, sliding Fulfilment.
Archetype 02

The fake-profile mill

Stand-in interviewees, borrowed CVs, engineers moonlighting across three payrolls. The person who interviewed is not the person who logs in on day one.

What the score seesfailed identity checks, proxy-interview evidence, and overlapping employment records. These floor events cap the band at High Risk.
Archetype 03

The genuinely good partner

Strong engineers, honest sourcing, and support when it’s hard. Today, they’re indistinguishable from the other two, which is unfair to them and dangerous for you.

What the score seesa high, stable Trust Score they’ve earned and can finally show.

A pitch deck can’t tell them apart. Twelve months of measured behaviour can.

How scoring works

Five pillars. One number. Zero self-declared claims.

Every candidate a vendor submits passes through proctored AI interviews that include identity verification, environment scanning, gaze and lip-sync analysis, typing dynamics, and cross-channel fusion. The vendor’s score is the aggregate of what we measured.

Weighted by what fraud costsAuthenticity outweighs everything
30%

Profile Authenticity

Identity, credential and employment verification outcomes. Fake profiles cost the most.

25%

Interview Integrity

Human-confirmed cheating incidents per session. Raw flags never count.

20%

Candidate Calibre

Calibrated technical assessment outcomes and L1 pass rates.

15%

Fulfilment Reliability

Joining ratio, no-shows, replacement rate, and the escalation loop below.

10%

Compliance & Standing

GSTIN status, documentation, MSA standing, tenure in network.

Trust score bands74Trusted
this vendor74
040557085100
  • High Risk0–40
  • Watch40–55
  • Established55–70
  • Trusted70–85
  • Assured85–100

Trust Scores summarize human-reviewed evidence from proctored AI interviews. Weights are indicative pending outcome calibration. Aggregates shown for cohorts of 10+ candidates.

The escalation loop · new

Flag it. The clock starts. It gets fixed, or their score pays for it.

A formal escalation beats three weeks of email. The vendor gets a fixed cure window: 48 hours for no-shows and 14 days otherwise. If the issue is verified and left uncured, it hits their Trust Score in front of every client on the network. Pick a path and watch how it plays out.

Pick a path
1

Client flags an issue

A quality collapse, no-show pattern, or misrepresented profile is filed on the network with evidence attached, not buried in email threads.

logged on the shared timeline
2

Vendor notified · cure clock starts

There are 48 hours for no-shows and abandonment, and 14 days otherwise. The vendor sees exactly what was flagged and gets a window to replace, correct, or resolve the issue with you directly.

no score impact yet
3

Sharedpro verifies

A human reviewer checks the evidence and the vendor’s response. One-sided claims from either party don’t survive this step.

human review, always
4a

Cured in window → logged, no score impact

Resolution is recorded on the evidence timeline. Vendors who fix problems fast build a visible remediation record.

score unchanged
4b

Verified & uncured → the score takes the hit

The escalation lands on Fulfilment Reliability, the band can drop, and every client on the network sees it before their next empanelment call.

consequence, network-wide

Filing and adjudication are always free. Panel fees apply only if a decision is contested, and if your escalation is upheld, your fees are credited. Being right never costs you money.

Vendor trust scoreLive
Cured in window · escalation closed
74/ 100
Trusted
± 0 · no score impact
Cure window
14 days · 48h for no-shows
Adjudication
Human reviewer, both sides
Filing cost
Free, always

Cured within the window. Score unchanged at 74; resolution logged on the timeline.

Monitoring · watchlists

Empanel once. We watch the bench so you don’t have to.

Star any vendor in the registry, and alerts come to you by email or Slack. Defaults are deliberately quiet: band changes and floor events only, with nothing else unless you ask. Zero-event weeks mean zero messages.

Floor event · criticalAlways on

Zenith Manpower: identity fraud confirmed

Band capped at High Risk pending resolution. Critical alerts can’t be muted, don’t wait for digests, and ignore quiet hours because safety is never a preference.

Band change · standardDefault on

Nimbus Infoservices: Trusted → Established

Fires only on genuine band transitions, never on score flicker. One email per vendor per day, coalesced, delivered in your hours.

Correction · standardAlways follows

Acme Talent: escalation overturned, score restored

If you were told about an upheld escalation, you will always be told about its correction. That rule has no off switch; accusations and corrections travel together.

You also receive a quarterly bench-health digest covering your watched vendors versus network medians, incident and cure rates, and open appeals. It is built to be forwarded to your CFO or auditors as-is.

Why a network beats a blacklist

One verified escalation. Every client warned.

Your internal vendor blacklist protects one company. A trust network makes every client’s experience part of every vendor’s record, just as a credit bureau turns one bank’s bad loan into a warning for every bank.

1 → 800+One verified escalation informs every client evaluating that vendor
24 moConfirmed incidents decay from the score; the evidence log is permanent
14 daysStandard cure window before any escalation can touch a score
100%Of score-affecting events pass human verification first
Client registry · manufacturing“Zenith”
Zenith ManpowerHigh Risk
Identity fraud confirmed · floor event
Client registry · BFSI“Zenith”
Zenith ManpowerHigh Risk
Identity fraud confirmed · floor event
Client registry · GCC“Zenith”
Zenith ManpowerHigh Risk
Identity fraud confirmed · floor event

Filed once. Visible in every client’s registry search before they sign.

Fair by design

A score clients believe because vendors can’t be ambushed by it.

A trust score is only worth using if neither vendors nor clients can game it. These rules are structural, not policy pages.

Structural rule

No unilateral score hits

A client flag alone never moves a score. Verification and the cure window stand between every complaint and every consequence.

Structural rule

Cure window before consequence

Vendors always get the chance to fix it first. Fast remediation is recorded, allowing good vendors to build a visible recovery record.

Human-confirmed incidents only

Behavioural flags from proctoring are evidence for reviewers, never automatic verdicts. Same rule for escalations.

Dispute rights on every item

Vendors can dispute any event on their record. Disputed items are flagged on the timeline pending resolution.

Thin files aren’t punished

For fewer than 15 assessed candidates, we show “insufficient history,” never a low score caused by a lack of data.

Candidate privacy is absolute

Only vendor-level aggregates with a 10-candidate display floor. Individual candidate data never surfaces to anyone.

Empanelment

Move your vendor bench onto the network in four steps.

01

Bring your bench

Your existing vendors, or new ones from our 800+ partner network, are empanelled under a single Sharedpro MSA. Missing one? A one-field assessment request onboards them.

02

We assess under proctoring

Every submitted candidate goes through proctored AI interviews. Scores begin accruing from the first batch.

03

Scores go live

Search any empanelled vendor by name or GSTIN. Export any trust card as a due-diligence report for your procurement and audit files.

04

You watch & hold escalation rights

Watchlist alerts and the escalation loop give you a standing lever against vendor issues, backed by score consequences.

Questions enterprises ask

The fine print, out loud.

Can a client tank a vendor’s score with one complaint?
No, not structurally. A flag opens a case; it doesn’t move a number. The vendor gets a 14-day cure window, after which a Sharedpro reviewer verifies the evidence and the response. Only verified, uncured escalations reach the score. One-sided claims from either party don’t survive review.
Does the vendor see who flagged them?
The vendor sees the issue, the evidence, and the deployment it relates to, giving them enough information to fix it. Escalations appear on the public evidence timeline in anonymized, aggregate form.
What if we disagree with the resolution?
Either side can contest once. The case goes to an independent three-member panel consisting of one legal, one senior technical, and one industry expert. Members are randomly drawn from published pools, with one strike allowed for each party, and decide by majority vote through a contractually binding expert determination. Contested items carry an under-appeal flag until decided; overturned outcomes restore the score with a public correction.
Is the score based on AI judgements?
The evidence is machine-collected through proctored interviews, verification pipelines, typing dynamics, and cross-channel fusion. The judgements are human. No score-affecting event is published without a reviewer’s sign-off, and score weights are published as indicative pending outcome calibration.
We already have a vendor blacklist. Why this?
A blacklist is a memory of your pain. The network is a memory of everyone’s. A vendor who burned another client last quarter shows up in your registry search today, before you sign. It also works in the vendor’s favour: good partners carry proof they’ve earned.
Get started

Stop underwriting vendor risk on gut feel.

  • Single MSAyour existing bench, empanelled once
  • Proctored evidencescores from measured behaviour, not claims
  • Escalation rightsa cure clock with a score behind it
  • Watchlist alertsband changes and floor events, nothing else
Trust Scores summarize human-reviewed evidence from proctored AI interviews.
© 2026 Sharedpro · sharedpro.co · Vadodara, India · [email protected]Making employment portable since 2020
STEP 1 OF 5 · ABOUT YOU

Let’s start with you.

A few details so we can make the conversation relevant.