QVCCS App Suite · Audit, compliance & governance

Permissions Auditor

A blanket, read-only audit of every user, role and permission in a Genesys Cloud CX organisation: which licence tier each user's permissions attribute them to, exactly which permission from which role escalates them, and what that means at your own contracted rates — reconciled against Genesys' own assignment and inference views, with a what-if simulator that previews role surgery before anyone touches production.

Continuously integrated and updated. This is a published sample of the user guide. The App Suite changes frequently, so this page may not reflect the latest features, screens and behaviour. The current guide is available in the application through the QVCCS apps portal, included with every Managed Professional Services tier.

At a glance

  • cheapest-covering attribution
  • driver provenance
  • what-if role surgery
  • your own rate card
  • 7-sheet xlsx
  • strictly read-only

Security at a glance

Permissions Auditor

Read-only

Sign-in
OAuth client credentials you supply; secret held encrypted on the server and never sent to the browser
Stores
Timestamped audit snapshots (users, roles, permissions, licence ids, last-login dates, aggregate media counts; no conversation content), scoped to your org credentials
AI
No AI
Exports
Seven-sheet Excel workbook
Genesys Cloud permissions
authorization:role:view, authorization:license:view, directory:user:view; optional analytics:conversationAggregate:view, billing:subscription:view

Compare every app

1What it is

Genesys Cloud CX licensing follows entitlement, not behaviour. In a multi-licence organisation the licence tier a user needs is set by the permissions their roles grant — whether or not they ever use those features. One quality:evaluation:view buried in a shared agent role silently turns an entire CX 1 voice population into CX 2 seats.

The Genesys admin UI shows each user's licence, but never why. Permissions Auditor answers the why: one read-only harvest of users, roles, the master permission catalogue and the org's own licence definitions; a deterministic attribution of every user to the cheapest covering licence; and full provenance — every escalation traced to the exact (permission ← role) pair responsible — which you can value at your own contracted rates.

The classic traps it is built to surface:

  • The shared-role trap — one "Agent" role reused everywhere gains a digital or QM permission for a pilot, and the whole population jumps a tier.
  • The default-role trap — a permission added to a default role escalates broad populations automatically.
  • Supervisor creep — evaluation, screen-recording, WFM or speech-analytics permissions granted "so people can see their own stuff".
  • Communicate drift — UC-only users who pick up a contact-centre role and start attributing (and billing) as full CX seats.
  • Zombie seats — active users with full role grants and no login for months.
  • Assignment drift — manual licence assignments that outlived a role cleanup (over-assigned) or lag behind a role expansion (under-assigned true-up exposure).
Nothing about tiers is hardcoded. The licence lattice — which SKUs exist, which permissions each covers, what comprises what — is read live from the org at every harvest. When Genesys reshuffles SKUs, a fresh snapshot reflects it. Each audit is a timestamped snapshot, kept in the app's warehouse, so you can re-run after a role cleanup and compare.

2Quick start

  1. Sign in — pick the org's region, paste an OAuth client-credentials Client ID and Secret. The permissions that client needs are in §8; the secret is held encrypted on the server and never sent to the browser.
  2. Choose the media-usage window — 90 days recommended; 60/30 also offered; skip makes the harvest faster but disables the usage-rightsizing lens.
  3. Run audit — the staged pipeline reports progress live: organisation → licence definitions → permission catalogue → roles → users → licence assignments → Genesys inference cross-check → billable-usage extract (best-effort) → media usage → attribution engine. Warnings (a missing optional permission, an inference cap) appear inline and are kept on the snapshot. You can navigate away — reloading the entry page reattaches to the stream.
  4. Explore — the finished snapshot opens on its Overview automatically. Start at the headline tiles, follow the Biggest levers into Findings, drill any user for the exact driver permissions, and simulate the fix in What-if before touching Admin → Roles.

No credentials to hand? The demo org — "Meridian Retail Group", a fictitious demo organisation of 657 synthetic users pushed through the very same engine — is fully explorable signed-out (View demo). Every §1 trap is planted, including the flagship storyline: a shared CC Agent Standard role carrying one QM permission that escalates 385 users — and a what-if strip of that single permission that recovers exactly that saving (385 seats × the CX 2 − CX 1 tier difference).

3How attribution works — the engine

The engine implements Genesys' documented multi-licence behaviour: a user is attributed the cheapest licence whose permission set covers every licence-relevant permission their roles grant. Attribution explains which licence a user's permissions require; how that is billed depends on the org's contract and billing model (see §6). Everything below exists to compute that honestly and explain it.

The licence model & canonical order

Licence definitions come from GET /api/v2/license/definitions — every SKU with its permission list and comprises lattice. Each definition is fuzzily bucketed (matching on id + description) into one of the six primary categories or left as an add-on:

  • communicate if the text mentions Communicate;
  • null (add-on) if it looks like an add-on, upgrade, WEM module, token/minute consumable, BYOC, Copilot/Assist or Analyzer SKU;
  • otherwise the digit 3/2/1 in the name picks CX 3 / CX 2 / CX 1, with a digital variant when the name says digital.

Where several definitions land in one category (rare), the one with the biggest expanded permission set keeps it; the rest demote to add-ons. The surviving primaries are ranked on the canonical order:

The licence model & canonical order
RankCategoryReading
0communicateUC only — no contact centre
1cx1Voice agent
2cx2digitalDigital-only, + QM
3cx2Voice + digital + QM
4cx3digitalDigital + WEM
5cx3Everything incl. full WEM
CX 4 licences are not yet modelled in this sample guide.
Prices never affect attribution. The ranking is this canonical order, independent of the rate card — so entering your rates can never flip which tier a user lands on. The rate card only values counts at view time (§6). The SKU→category mapping is shown openly on the Rate card page, so an oddly named SKU is visible rather than silently mis-bucketed.

Wildcard expansion & effective permissions

Both role grants and licence permission lists contain wildcards. Everything is expanded to concrete domain:entity:action strings against the master permission catalogue (GET /api/v2/authorization/permissions, paged):

  • Patterns arrive with 1–3 colon segments; missing trailing segments mean "all" (quality:* ≡ quality:*:*; bare * is everything).
  • A concrete 3-segment permission passes through even when the catalogue doesn't know it — the catalogue can lag brand-new platform features.
  • Roles are expanded from their flat permissions list plus their permissionPolicies (domain / entityName / actionSet triples).

Each user's effective permission set comes from GET /api/v2/users?expand=authorization,dateLastLogin&state=any — the authorization expand carries the user's roles and Genesys' own union permission set, so no per-user follow-up calls are needed. When the expand doesn't deliver the union, the engine unions the expanded role sets itself. state=any keeps inactive-but-still-granted users visible — they matter on named-user billing.

Effective permissions then split three ways:

  • Coverable — present in at least one primary licence: these (and only these) drive tier attribution.
  • Add-on-only — present only in category-null definitions: these flag that add-on SKU on the user by name (§3, Add-on licences) and never move the base tier.
  • Neutral — in no licence definition at all (platform/admin plumbing): ignored for licensing.

Attribution: the cheapest covering primary

Walk the primaries cheapest-first (canonical rank order); the first whose expanded permission set covers every coverable permission the user holds wins. If nothing covers everything (the user's permissions span disjoint SKUs), the user attributes to the top tier — matching how entitlement billing resolves the same situation. Users whose roles grant no coverable permission at all show as "No licence-relevant permissions" (typically integration or admin-only accounts).

Downgrade target, drivers & provenance

For every user above the floor, the engine picks the realistic downgrade target: among all primaries cheaper than the attributed one, the licence needing the fewest permission removals (ties break to the closest, i.e. highest, rank).

Not rank-minus-one. The tier ladder is not a chain: a voice agent one QM permission above CX 1 must be compared against CX 1 (1 permission away), not against CX 2 Digital — a "cheaper" tier their entire voice permission set would violate. Plain rank−1 produces nonsense driver lists (every voice permission "escalating"); fewest-missing produces the actionable one.

The permissions the target lacks are the user's drivers (capped at 60 per user), and each driver carries role provenance — the role(s) granting it, with sole source marked when exactly one role grants it: remove the permission from that one role and the user drops a tier. The engine also computes drivers against the org's floor primary, so "what keeps them above the cheapest tier at all" is answerable too.

Sole escalation rolls up to roles: a role solely escalates a user when it is the only granting role for at least one of that user's next-tier drivers. Those user sets power the Roles page's blast radius, the role-escalation findings and their savings.

The CC-floor rule

Users (and roles) attributed at the cheapest non-Communicate primary — the cheapest contact-centre tier the org owns, usually CX 1 — are reported as "at floor": no downgrade target, no drivers. Dropping a working agent to the UC-only Communicate licence is not a realistic remediation, so treating Communicate as the next step would flag every clean agent role as "escalated above Communicate" — pure noise. Because findings derive from drivers, this one rule kills that whole false-positive class. Genuine UC-only users who drift into contact-centre tiers are still caught — by the communicate-drift finding, which keys off their Communicate assignment, not off drivers.

Add-on licences are first-class

Some orgs never bought the next base tier — they upgraded the one they have. A real-world shape: CX 1 plus a pc1du2 digital-upgrade SKU plus a Predictive Engagement AI licence, with no CX 2 definition in the org at all. In such orgs the money lives in the add-on SKUs, so the auditor treats them as first-class:

  • Permissions found only in category-null definitions resolve to those definitions' ids — the user gets a named add-on chip (e.g. pc1du2, Predictive Engagement) next to their base tier.
  • Add-ons are filterable from the Users page's licence dropdown (own optgroup; the filter matches attributed tier or held add-on), listed in the Overview's add-on mix panel, and individually priceable on the rate card — a user's monthly figure is base tier + priced add-ons.
  • Validated against Genesys: on a production organisation with that shape, the engine's "CX 1 + these add-ons" answer matched POST /license/infer exactly.

Worked examples

A · The shared-role trap (the demo's flagship storyline)

  1. CC Agent Standard is a voice-agent role that silently carries quality:evaluation:view ("so agents can see their own evaluations").
  2. QM permissions live in CX 2 and above — CX 1 doesn't cover it — so all 385 holders attribute to CX 2 instead of CX 1.
  3. Downgrade target: CX 1 (exactly 1 missing permission — the QM perm; CX 2 Digital would need their whole voice set removed). Driver: quality:evaluation:view via CC Agent Standard, sole source.
  4. Finding: "Role 'CC Agent Standard' escalates 385 users", saving 385 seats × the CX 2 − CX 1 tier difference at whatever rates are on the rate card.
  5. What-if: strip that one permission → exactly 385 users move CX 2 → CX 1, the monthly delta is exactly the finding's saving. The audit's finding and the simulator's ground truth agree.

B · Fewest-missing beats rank−1

A voice agent holds 40 voice permissions + 1 QM permission → attributed CX 2. Cheaper primaries: CX 2 Digital (rank−1, missing all 40 voice perms), CX 1 (missing 1), Communicate (missing 45). Fewest-missing picks CX 1, and the driver list is one permission — the actual fix.

C · Add-on org, no tier movement

A digital agent in a CX 1-plus-upgrades org holds voice + digital perms. Digital perms exist in no primary (there is no CX 2 def) but do exist in the pc1du2 definition → attribution stays CX 1 and the user is chipped + pc1du2. Nothing pretends a tier exists that the org never bought.

4Reconciliation & the findings taxonomy

The three-way cross-check

This audit's attribution is never presented alone — it is reconciled against Genesys' own two views of the same question, plus the billing extract where permitted:

The three-way cross-check
ViewSourceWhat it says
Attributed (this audit)The engine, §3Cheapest primary covering the user's actual permissions — with the why (drivers + roles).
Assigned (Genesys)GET /license/usersThe licence Genesys currently has on each user. Mismatches vs attributed become findings.
Inferred (Genesys)POST /license/infer per distinct role set (capped at 300 sets)Genesys' own attribution for a set of roleIds — an independent check on the engine's maths, shown in each user's drill.
Billed quantitiesGET /billing/reports/billableusage (30-day window, best-effort)SKU quantities + counted users. No prices — see §6.

Disagreements become findings rather than being papered over — including a finding against the engine itself (engine-disagreement) whenever Genesys' inference names a different primary for a role set than the audit computed.

Assignment-vs-attribution mismatches

  • Over-assigned ("zombie seats" of the assignment kind) — Genesys has the user on a higher licence than the cheapest one covering their current permissions, typically a manual assignment that outlived a role cleanup. Saving per user = assignedMax − attributed: the price of the highest assigned primary minus the price of the attributed tier.
  • Under-assigned — the user's permissions imply a higher tier than every licence assigned. Because Genesys bills on entitlement this is a true-up exposure, not a saving — it is always valued at zero and flagged high severity.

The eight finding kinds

The eight finding kinds
KindSeverityWhat it meansSaving basis
role-escalationmedium (high at ≥25 users)A role is the sole source of permissions holding N users above the next-cheaper licence.Per user: attributed − downgrade-target price.
default-role-escalationhighSame, but the role is a DEFAULT role — it applies to broad populations automatically.As above.
assignment-overmediumAssigned above what permissions justify (stale manual assignment).assignedMax − attributed.
assignment-underhighPermissions imply a higher tier than assigned — billing true-up exposure.— (exposure)
engine-disagreementlowGenesys' license/infer named a different primary than this audit for the same role set. Treat Genesys as authoritative; use the drivers as the explanation to verify.—
zombie-seatshighActive, licensed users with no login for 90+ days (or ever).The whole seat: base tier + priced add-ons.
communicate-driftmediumUsers assigned UC-only Communicate whose roles now grant contact-centre permissions — they will attribute (and bill) as CX seats.Attributed − Communicate price.
usage-rightsizingmediumAttributed at a digital-capable tier (CX 2 Digital or above) but handled zero digital interactions and some voice across the usage window.Attributed − CX 1 price.

Findings sort by severity, then by affected-user count. Each carries affected users, a concrete remediation, and — where costable — the monthly saving at the current rate card.

Savings overlap. Fixing one role can empty the rightsizing list and shrink other findings too — the findings are independent lenses on the same population. Treat any summed figure as an upper bound and work top-down; the what-if simulator (§7) gives the exact combined effect of a specific set of edits.

The security lens (same harvest, zero extra calls)

  • Master admins — holders of a role named like "Master Admin".
  • Admins — holders of roles whose name starts with "Admin".
  • Role managers — holders of any role granting authorization:role:edit, authorization:role:add or authorization:grant:add — the people who can mint escalations.
  • Stale admins — any of the above with no login for 60+ days.
  • Wildcard-heavy roles — roles with ≥3 wildcard grants (top 20 by count).

5UI walkthrough

Entry & snapshots

The signed-in landing page: pick the usage window, Run audit, and watch the staged progress log stream live (stage lines, per-stage page counts, amber warnings). One harvest runs per org at a time — a second request is refused until the first finishes. Below, the snapshot list shows every past audit with org, timestamp, counts and status; completed snapshots open on click; failed ones keep their error. A reload mid-harvest reattaches to the stream automatically, and on completion the app navigates to the fresh snapshot itself.

Overview

  • Headline tiles — attributed monthly cost (with rate-card provenance), identified monthly savings across findings, users escalated by role config, zombie seats. Every tile drills through to the page that explains it.
  • Licence mix — attributed users per tier as proportional bars with users × price = total maths, plus the add-on mix beneath (holder counts, priced or "unpriced — set on the rate card"). Click any row to filter Users to it.
  • Reconciliation panel — Genesys-assigned licence counts, the billable-usage extract (or why it's absent), and direct links into any assignment-mismatch findings.
  • Biggest levers — the top findings by monthly saving, with the upper-bound caveat.
  • Security strip — the five §4 security counters, linking to the full lists.
  • Harvest warnings (or the demo banner) render at the top; Export workbook (.xlsx) is a one-click download.

Users

The full population, 50 per page. Filters: free-text search (name/email/department/title), the attributable-licence dropdown (base tiers and add-ons in separate optgroups — an add-on matches users holding it on top of any tier), assigned-vs-attributed mismatch (over / under / any), last-login age (30/90/180d), plus role- and finding-scoped views arriving via links from other pages. Sortable by name, attributed tier, last login, permission count or department.

Each row shows attributed tier + add-on chips, the monthly seat price, Genesys-assigned licences, the escalator summary ("3 perms above CX 1" or at floor), last login (red at 90d+) and voice/digital handled counts. Expanding a row is where the app earns its keep: every driver permission with its granting role badge(s), sole source highlighted where one role edit would demote the user, the add-on chips their permissions imply, all their roles, media handled by type, and Genesys' own inference for their role set.

Roles — licence blast radius

Every role scored by its licence impact, sorted by monthly saving: holders, the minimal licence the role's own permissions imply, how many users it solely escalates, the monthly saving if its escalating permissions were removed, and Genesys' own baseLicense tag for the role (another cross-check). Default roles are badged; permission and wildcard counts show under the name. Selecting a role opens the detail panel: escalating permissions highlighted amber against the full expanded permission list, with Holders (filtered Users view) and Simulate fix (pre-loaded What-if) buttons.

Findings

Every finding as a card: severity badge, kind, plain-English detail, a concrete Fix: remediation, the money at stake at the current rate card (with provenance), and jump links — the affected users, the offending role, and a pre-loaded what-if simulation. The security posture lists (master admins, stale admins, role managers, wildcard-heavy roles) live at the bottom of the same page.

Excel export — 7 sheets

Excel export — 7 sheets
SheetContents
SummaryOrg, snapshot time, counts, attributed monthly total, tier mix, pricing provenance, harvest warnings.
UsersOne row per user: department, division, state, attributed licence, add-on licences, monthly price, Genesys-assigned, mismatch (red under / amber over), roles, days since login (red at 90+), voice/digital handled.
Escalation driversOne row per (user, driver permission): attributed tier, downgrade target, granting roles, sole source flag.
RolesBlast radius per role: holders, minimal licence, solely-escalated count, monthly saving (red when >0), escalating permissions, Genesys base-licence tag, wildcard count.
FindingsThe full taxonomy with severity, user counts, savings, detail and remediation.
Security postureMaster admins, admins, role managers, stale admins, wildcard-heavy roles.
Rate cardThe six category prices + per-SKU add-on prices used, with provenance.

Frozen header rows, autofilters, and red/amber highlights on the rows that cost money — read-back verified.

6Rate card & costing

Why the rate card exists

Genesys exposes no tenant prices in the customer API. The org's own billable-usage extract (/api/v2/billing/reports/billableusage) carries SKU quantities and counted users — no money. The only priced surface on the platform is the partner-side trustee endpoint (/api/v2/billing/trusteebillingoverview/{trustorOrgId}, whose SubscriptionOverviewUsage rows carry prepayPrice) — usable only by the selling org's trustee access. QVCCS is not the selling org for its customers, so that route was deliberately dropped. Every monetary figure in the app is therefore your rate card × audited counts, and honest about it.

  • Defaults: indicative published list rates are pre-loaded as a starting point. The currency shown is display-only.
  • Provenance is always shown: every money figure is badged list prices until the card is edited, then your rates — so an estimate is never mistaken for invoice truth. Real contracts are discounted; enter yours before quoting savings.
  • Per-snapshot storage: each snapshot keeps its own card; saving re-prices every view instantly and — by construction — never changes attribution.
  • Add-on prices (keyed by licence id): a price row appears for every add-on SKU actually observed on users. Unpriced add-ons count as zero until set.

How money is computed (view time only)

  • User monthly = price(attributed category) + Σ price(held add-ons).
  • Attributed monthly total (Overview tile) = Σ user monthly across the org.
  • Finding savings — per §4's table: tier findings price the per-user delta to the target; assignment-over prices assignedMax − attributed; zombies price the whole seat including add-ons; exposures are valued at zero.
  • Role blast radius = Σ (attributed − next-cheaper) across the users the role solely escalates.

Attribution is not billing. Attribution answers which licence each user's permissions require — the cheapest licence covering every licence-relevant permission they hold. What the org is actually billed depends on the licences assigned and its billing model: under the concurrent licensing model, Genesys charges reflect the highest licence level users were assigned during the billing period. That is why the audit sets attribution beside Genesys' assigned and billed views (§4) rather than treating it as the bill; treat money figures as an estimate to reconcile against the invoice.

The page also shows the licence-definition mapping — how every SKU in this org was bucketed (category or "add-on / other") with its expanded permission count — so the fuzzy §3 categorisation is inspectable after the first harvest of any new org.

7What-if simulator

Role surgery, previewed: pick a role (the dropdown lists escalating roles first, with their savings), then either remove the role from every holder or untick permissions to strip from it — escalating permissions are pre-highlighted amber with a one-click Strip all escalating permissions. Edit as many roles as you like in one scenario, then Recompute cost.

The result panel shows the monthly saving (before → after at your rate card, with provenance), the licence-mix diff per tier, and the moved users (first 500 listed by name with from → to).

Simulation semantics

  • Both sides are recomputed from the simulator's role→permission sets (stored in the snapshot) — never mixing Genesys' harvested per-user union with our own union — so the delta reflects your edits alone, with no phantom differences.
  • Strip patterns accept wildcards (quality:*) and are expanded the same way as role grants; up to 50 roles removed and 200 patterns per role per scenario.
  • Moved total is computed from the tier histograms (half the L1 distance), so it is exact even beyond the 500-row detail list.
  • Nothing is written to Genesys. The output is a change plan for a human to apply in Admin → Roles — remediation is a report, not a button.

8Genesys endpoints & permissions

Everything the app ever calls against Genesys, in harvest order:

Genesys endpoints & permissions
EndpointUsed for
POSTlogin.<region>/oauth/tokenClient-credentials sign-in
GET/api/v2/organizations/meOrg identity (at sign-in and per harvest)
GET/api/v2/license/definitionsThe licence lattice: every SKU's permission ids + comprises
GET/api/v2/authorization/permissionsMaster permission catalogue (wildcard expansion), paged
GET/api/v2/authorization/rolesRoles + permissionPolicies + default/base flags + Genesys' own baseLicense tags
GET/api/v2/users?state=any&expand=authorization,dateLastLoginThe population with per-user roles AND Genesys' effective permission union, plus last login
GET/api/v2/license/usersGenesys' per-user licence assignments (the "assigned" view)
POST/api/v2/license/inferGenesys' own attribution for a roleIds[] set — the cross-check (≤300 distinct role sets per harvest)
GET/api/v2/billing/reports/billableusage30-day billed quantities (best-effort; quantities only, no prices)
POST/api/v2/analytics/conversations/aggregates/queryPer-user handled counts, groupBy [userId, mediaType], metric tHandle, in ≤30-day chunks (rightsizing lens)

OAuth client permissions

OAuth client permissions
PermissionUsed forRequired?
authorization:role:viewRoles + permission policiesYes
authorization:license:viewLicence definitions, per-user assignments, inferenceYes (assignments/inference degrade to warnings without it)
directory:user:view (+ authorization visibility for the expand)The user population, departments, last login, effective permsYes
analytics:conversationAggregate:viewPer-user media usageOptional — usage lens skipped with a warning
billing:subscription:viewBillable-usage extractOptional — typically admin-restricted; skipped with a warning
Read-only by construction: the HTTP client refuses every non-GET request before it reaches the network, except a two-endpoint POST allowlist of read-only-by-effect queries — license/infer and the conversation-aggregates query. There is no code path that can modify a role, a grant, a user or a licence. Politeness is built in too: requests are rate-limited, the app backs off automatically when Genesys asks it to slow down, and tokens are refreshed when they expire.
Propagation lag: Genesys can take a few minutes to propagate fresh role/permission grants into tokens. If a just-granted permission still 403s, wait 2–5 minutes, sign out/in, and re-run the audit.

9Storage, privacy & security

  • Snapshot warehouse — each audit is kept as a compressed snapshot in the app's own private store, so past audits can be reopened and compared.
  • Owner scoping — snapshots belong to the org credentials that created them. Two orgs using the same service can never see each other's snapshots; the same org signed in again sees its own.
  • Data held — user names, emails, departments, titles, divisions, role ids, permission strings, licence ids, last-login dates and aggregate media counts. No conversation content, no recordings, no participant data.
  • Credentials — the Client Secret is held encrypted on the server and is never sent to the browser or written in plain text. Tokens refresh automatically on the server, and sessions end after a period of inactivity.
  • Read-only posture — the §8 request guard means the auditor cannot change the Genesys org even if a bug asked it to. Nothing leaves the service except calls to the selected Genesys regional API; there is no AI component.
  • Demo isolation — the demo data is synthetic; its rate card is not stored and is shared between demo viewers.

10Technology

  • Web application — a server-side Genesys Cloud client with built-in rate limiting, live progress reporting during the harvest, and an Excel workbook builder.
  • Browser interface — snapshot pages load on demand, and filter state lives in the address bar so views are shareable.
  • Audit engine — attribution, reconciliation, costing and the workbook are all derived from each snapshot's stored audit result, so every view — and every what-if — re-derives from the same audited facts.

11Troubleshooting

The harvest completed with a billing-extract warning.
Expected for most OAuth clients — billing:subscription:view is typically admin-restricted. The audit is complete; only the billed-quantities reconciliation panel is empty.
403 warnings right after granting a permission.
Token scopes are minted at sign-in, and Genesys propagation can lag by minutes. Wait 2–5 minutes, sign out and back in, and re-run the audit.
A user's attribution looks too high.
Expand their row on Users: the drivers list shows exactly which permissions force the tier and which roles grant them. If Genesys' inference (shown in the same drill) disagrees, that's also an engine-disagreement finding — treat Genesys as authoritative and use the drivers as the explanation to verify.
Tier mix shows users under "No licence-relevant permissions".
Their roles grant only permissions that appear in no primary licence — typically integration or admin-only accounts. They price at zero.
A what-if strip moved fewer users than the finding promised.
The finding counts users the role solely escalates; a user whose driver is also granted by a second role needs both roles edited. The simulator is the ground truth — it recomputes real attribution per user.
An add-on SKU I expected isn't in the Users licence dropdown.
The dropdown lists add-ons actually observed on users in this snapshot. If nobody's permissions resolve to the SKU, it only appears in the Rate card page's mapping table.
Opening a snapshot says "Snapshot still harvesting".
It's mid-run — watch the entry page's progress stream. Failed snapshots keep their error on the list; just run a fresh audit.
"Org has N distinct role sets; inference cross-checked for the first 300."
Deliberate cap on the license/infer cross-check in very role-diverse orgs. The audit's own attribution covers everyone; only the disagreement lens is sampled.
Everything 401s mid-session.
The session ended after a period of inactivity. Sign in again — snapshots belong to the org credentials, not the session, so nothing is lost.
Savings quoted to a customer don't match their invoice.
Check the provenance badge: list prices means nobody has entered the contracted rates yet (§6). And remember the findings' totals are an upper bound — use What-if for the exact combined effect of a concrete change set. Attribution shows the licence each user's permissions require; on concurrent (rather than named-user) billing, charges reflect the highest licence level users were assigned during the billing period, so treat figures as directional.

QVCCS Permissions Auditor — role, permission & licence audit for Genesys Cloud CX. Strictly read-only against the Genesys APIs; each org sees only its own snapshots.

Continuously integrated and updated. This is a published sample of the user guide. The App Suite changes frequently, so this page may not reflect the latest features, screens and behaviour. The current guide is available in the application through the QVCCS apps portal, included with every Managed Professional Services tier.

Included at every tier

Use Permissions Auditor with QVCCS Managed Professional Services.

The App Suite comes with every Bronze, Silver, Gold and Diamond subscription, used by our engineers and your administrators alike – backed by the certified QVCCS bench.

Explore the tiers

Last reviewed

Questions about what you have read?

Clients, partners and people introduced to us can reach the specialists behind our applications and articles directly.

Who to contact