How we vet every helper

This is the unglamorous part of running a marketplace. It's also the part that decides whether the marketplace is actually worth using. Here's the exact path every Helpward helper application takes from submission to either approval or a declined-with-reason.

Step 1: profile completion

A helper signs up at /signup?role=provider, completes their basic profile (full name, bio, languages, service radius), and selects the services they want to offer from the catalogue. None of this gates the next step — the substantive vetting is what's gated.

Step 2: Stripe Identity verification

Before any background check runs, the helper photographs a government-issued ID (US: driver's licence, state ID, or passport; Canada: provincial licence or passport) and submits a live selfie. Stripe's system matches the face on the ID against the selfie and validates the document against the issuing authority's records. If identity fails, the helper application halts here — we don't proceed to a criminal check using someone's unverified identity.

Step 3: criminal background check

US helpers run through Checkr; Canadian helpers run through Triton Canada. Both are FCRA-accredited (US) / PIPEDA-compliant (Canada) and used by major marketplaces. The check pulls: SSN/SIN trace, national sex-offender registry, federal criminal records, county-level criminal records for every county the helper has lived in over the past seven years, OFAC/sanctions screening, and a motor vehicle record if the helper is offering any transportation services.

Step 4: human adjudication

Every flagged record goes to a Helpward reviewer. Helpward operates a documented adjudication policy — at /safety/background-checks — that explains what triggers automatic disqualification (sexual offences, violent felonies in past 7 years, convictions involving minors, identity theft, open felony warrants) versus what gets reviewed individually (non-violent misdemeanours over 5 years old, minor traffic violations, expunged records). The policy is public so applicants know what to expect.

We deliberately don't auto-disqualify on every record. A 12-year-old non-violent misdemeanour with no recent activity is the kind of thing a human reviewer can weigh in context. A recent violent offence is not. The line is published; the calls behind the line are made by people, not by an algorithm.

Step 5: bank linking via Stripe Connect

Once the background check clears, helpers link a bank account through Stripe Connect's onboarding flow. This is the lightest step procedurally but the most important for the helper economically — we don't activate the account until payouts can land.

Step 6: final human approval

A senior reviewer reads the complete application — profile, ID-verification status, background-check adjudication result, bank-link status — and either approves or declines with a written reason the helper can read and (if applicable) appeal. Approval flips the helper's status to 'approved' and they can go online from /provider/active.

Re-screening

• Annual full re-check on the anniversary of approval.
• Continuous monitoring (US) — Checkr surfaces new records on existing helpers as they appear.
• On-report — any customer safety report triggers an immediate ad-hoc check.
• Status review on prolonged dormancy — helpers who haven't taken a task in 12 months go through identity re-verification before reactivation.

Why we're transparent about this

Helpward only works if customers trust the bar. The bar is meaningless if it's a black box. So we publish what the bar is: the categories of records that disqualify, the categories that get reviewed, the third-party services we use, the response times we commit to. If your application is declined, you get a written reason and an appeal path. If we make a mistake, we own it.