Choose the flow a caller's journey starts in. Flow Narrator reads that flow and everything it calls —
bots, common modules, tasks, menus, decisions and switches, data actions and the integrations behind them, flow
outcomes and milestones, variables, prompts, and the flows and queues it transfers to — follows the logic to every
place a call can end, and writes it up as a plain-English as-built document in Word or PDF, for business and
technical readers alike.
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
starting flow only
executive summary + detail
every integration traced
Word · PDF
UTC time-stamped
strictly read-only
Security at a glance
Flow Narrator
Read-only
Sign-in
OAuth client credentials you supply; secret held encrypted on the server and never sent to the browser
Stores
Saved flow readings and the Word and PDF files written from them, held per OAuth client until you delete them under Stored data (deletion is permanent)
AI
No AI features are described in the user guide
Exports
Word and PDF as-built documents
Genesys Cloud permissions
View permissions via architect, integrations, routing, users, groups, scripts and process-automation read scopes; organization:readonly
An Architect flow is a precise but technical thing: tasks, menus, switches with conditions written in Architect's
expression language, data actions that call external systems and map their answers into variables, bots with slots
and intents, outcomes and milestones for reporting, and transfers to queues and other flows. Flow Narrator turns
that configuration into a written account of exactly what the flow does — what the caller hears and can do at each
step, what the system checks and why, what each integration is asked and how its answer changes the call, and every
way the call can end.
Each part of the document opens with an executive management summary for business owners, operations,
managers and project managers, followed by the full detail for flow engineers. The document is titled
QVCCS Flow Narrator – [flow name] v[version], carries its generation date and time in UTC, cites
Quo Vadis CCS Ltd as the generator, and includes a disclaimer and the permissions an accurate rendering
depends on.
Read-only posture: Flow Narrator only ever reads from your Genesys Cloud organisation. Every call
it makes is a GET; anything else is refused inside the application before it reaches the network.
2Signing in
Flow Narrator signs in with a Genesys Cloud OAuth Client Credentials grant, like the rest of the QVCCS tools.
In Genesys Cloud, open Menu > IT and Integrations > OAuth and add a client with the grant type Client Credentials.
On its Roles tab, give it a role holding the permissions in section 3, in the divisions whose flows you will describe.
On its Scope tab, if your organisation restricts OAuth scopes, allow the scopes in section 3.
Save, and copy the Client ID and Client Secret.
On the Flow Narrator sign-in page choose your organisation's region, paste the Client ID and Secret, and choose Sign in.
The secret is held encrypted on the server, never sent to the browser and never shown again. Sessions end after a period of inactivity; Sign out ends yours at once.
3Permissions and scopes
The document is only as complete as what the client may read. When the client cannot read something a flow
refers to — a bot it cannot open, a data action whose definition is hidden — Flow Narrator names the gap in the
document instead of leaving it out silently, and the disclaimer says so. For a complete and accurate narrative the
client needs:
Permissions and scopes
What is read
Role permission
OAuth scope
The organisation's name (sign-in)
None needed
organization:readonly
Flows, their versions and configuration — inbound, in-queue, bots, common modules, workflows, every type
Users (transfer targets, and who published each version) and groups
None needed
users:readonly; groups:readonly
Scripts shown to agents (screen pops)
Scripter › Script › View
scripts:readonly
Workflows started after the call (process automation triggers)
Process Automation › Trigger › View
process-automation:readonly
Every endpoint above was read successfully against a live organisation during the build, except process automation
triggers, which Genesys refused for a client whose role lacked processautomation:trigger:view. Its refusal
names the permission exactly, and the inventory screen shows it the same way.
Verified. Every row matches the permissions and scopes Genesys publishes for each endpoint
in its own API definition (checked 24 September 2026). “None needed” means Genesys requires no role permission for
that read; where your organisation restricts OAuth scopes, the scope is still needed. The same list is printed in
every generated document.
4Choosing the starting flow
After signing in you see every flow in the organisation, of every type, with its division and published version.
Type narrows the list to one kind of flow; it opens on Inbound call, where most caller journeys start. Each type shows how many flows of that kind exist.
Search matches the flow's name, description and division.
Refresh reads the list from Genesys again (the list is otherwise kept for a minute).
Choose a flow to see its details on the right, and pick the version to describe. It defaults to the published version; any other saved version can be chosen, with who created it and when (in UTC). Debug versions are marked.
Continue moves to the next step for that flow and version.
5What is followed, and what is not
Flow Narrator follows only what the starting flow calls or refers to: the bots it calls, the common modules
it runs, its tasks and menus, the data actions and data tables it uses, the queues it transfers to and the in-queue
flows those queues play, and the flows it transfers the caller to — and, in turn, what those call. It never reads
other flows in the organisation just because they exist; a flow the starting flow cannot reach is not part of its
story.
Versions. The starting flow is read at the version you choose. Every flow it calls is read at the
version Genesys runs: the version a Call Common Module or Call Bot Flow block pins, where it pins one, otherwise
the published version. If two blocks want different versions of the same flow, the document says which one it
describes. If the version you chose is not the published one, the document says so: call routes and post-call
workflows reach only the published version.
Queues chosen while the call runs. A transfer to a queue held in a variable (found by name, looked up
in a data table, returned by a data action) cannot be resolved from the configuration, so the in-queue flow the
caller would then hear is not described. The document lists every such transfer in its “Scope” appendix.
Post-call workflows are included when a trigger's condition values are exactly the id or name of
something in this journey, or one of the numbers the call routes send to this flow, or when the trigger runs
after every conversation with no conditions. A partial match never counts: “Sales” does not match “Sales Europe”.
5aReading the flow: the inventory
After you choose Continue, Flow Narrator reads the chosen version and follows everything it calls. A
progress bar and a running log show each flow as it is read. When it finishes, the inventory lists:
How calls reach this flow — the call routes (Architect › Call Routing) whose open, closed or holiday
hours send numbers to the starting flow, with every number and the schedule group that decides which applies.
Flows, grouped by what they do for the caller: the starting flow; the bots it calls; the common modules
it runs; the flows it transfers callers to; the in-queue flows of the queues callers are sent to; and post-call
workflows that a trigger ties to this flow. Each shows its version, who published it and when, how many actions
it holds, and exactly where it is reached from: the flow, the task or menu, the block and its tracking id.
Everything else it refers to, by kind: data actions with the integration, method and host they call;
data tables with their columns and row count; queues with their in-queue flow; schedule groups with their time zone
and schedules; prompts with their languages and how many recordings have no transcript; flow outcomes with their
descriptions; skills, scripts, users and groups. Each lists every block that uses it.
Gaps — anything Genesys refused to show, in Genesys' own words, including the permission it names.
The document will say what each gap prevents it from describing.
Genesys endpoints read — every kind of call made, and whether the client was allowed to make it.
Referenced flows are read at the version Genesys runs them at (see section 5), and the starting flow
at the version you chose. The inventory is saved, so opening the same flow and version again shows it at once;
Read again refreshes it from Genesys. Post-call workflows are included when a process automation trigger's
condition values are exactly the id or name of something this flow uses (the flow itself, one of its queues, flows
or resources), or the dialled numbers the call routes send to this flow (the most common way a trigger is
tied to a flow), or when the trigger runs after every conversation with no conditions. The inventory says which trigger included each
workflow and why, for example “matches 14 of its 14 dialled numbers”.
5bReviewing the model
From the inventory, Review the model shows what the document will be written from, before any prose is
written, so it can be checked against Architect. It is built from the flow configurations the trawl saved, in a
fraction of a second, and it follows Architect's own structure exactly: each task, menu or bot state is a list of
steps; a branching block (a switch, a decision, a data action, a schedule check, a collect input) opens one branch
per outcome; and when a branch's steps run out the flow carries on after that block, as Architect does.
Jumps never return. After Jump to Reusable Task, Jump to Menu, End Task, Loop Next or
Exit Loop, nothing stored below it in the same sequence runs. Such blocks are listed as “can never run”, with
the block they follow. (Confirmed on real call traces: in one production flow, a block that ran on every call up to
version 69 has not run once since a jump was placed in front of it in version 70.) Call Reusable Task is
different: it runs the task and carries on.
After a transfer. A transfer that succeeds hands the caller on. The blocks after it run only when the
transfer fails, so they are shown under that transfer's “If the transfer fails” branch.
When every branch leaves. If every branch of a block ends the call or jumps away, the blocks after it can
never run, and are listed as such.
Tasks and menus nothing leads to. A task or menu that no reachable part of the flow points at (directly
or through other such tasks) never runs. It is marked never runs here, left out of the journey, its endings,
outcomes, data action calls and variable counts, and described in its own appendix.
Tasks and menus — choose a flow and a task to see its steps in plain English, every branch nested under
its condition, with each block's tracking id. Menus show what they play, and each option's key, spoken phrases and
destination. Each task says where it is reached from and what its exits are.
Data actions — for every data action: the integration, method and address it calls, what it takes and
returns; for every place it is called, what the flow sends, where each answer is stored and which steps in the flow read
it (and whether other steps also set the same variable), which answers are stored but never read, and what happens on failure and on timeout.
Where calls end — every transfer, disconnect, bot hand-back and workflow end, with every condition
that leads to it — including the ones implied by earlier branches that leave the task (a transfer after an
opening-hours check is reached only when the service is closed, on a holiday or in an emergency).
Variables — every variable with its type and starting value, and how many steps set and read it
(and which). “Never set” and “never read” are shown.
Checks — per flow: blocks that can run (all described), blocks that can never
run and why, tasks and menus nothing leads to, lines quoted rather than translated, and any block type not yet
described.
How conditions are translated. Architect stores each expression as a parsed tree as well
as text. Flow Narrator translates the tree: Flow.queue.name == "Sales - Spare Parts" reads
“the name of ‹queue› is “Sales - Spare Parts””, If(IsNotSetOrEmpty(Flow.x), false, true) reads
“‹x› has a value”. Where a condition combines “and” with “or”, or holds another condition inside it, the inner part
is bracketed exactly as Architect groups it. Lists and text positions count from 0 in Architect:
Flow.ids[0] reads “the first item of ‹ids›”, Substring(Flow.x, 6, 2) reads “the 2 characters of
‹x› starting at the seventh character”. Variables appear in ‹angle quotes› as words; recorded prompts appear with their transcript, or
marked “no transcript”; in a flow with more than one language the transcript is the
default language's, marked with that language, and every language's transcript is in the Prompts appendix. Anything it cannot translate is shown exactly as written, marked quoted, and counted
on the Checks tab.
5cThe document
From the model, Read the document writes the as-built narrative and shows it as it will read in Word or PDF.
It is written fresh each time, dated and time-stamped in UTC, titled QVCCS Flow Narrator – [flow] v[version], and
cites Quo Vadis CCS Ltd as its generator. Every chapter and every task opens with a summary written only from
facts in the model — counts, names and connections, never guesses — followed by the full detail.
c The document
Part
What it contains
Front matter
Flow, version, organisation, generation and read times, the generator, and the disclaimer (what the document can and cannot say, and how permissions affect it).
1 · Executive summary
The whole journey in a few paragraphs for managers and owners, and an at-a-glance table.
2 · How calls arrive
Call routes and numbers; every schedule group's open, closed and holiday schedules written out in words (for example “every Monday to Friday, 07:30 to 16:00”), past dates flagged; standalone schedule checks with the time zone their block uses; languages, voices, and what happens on a system error.
3 · The caller's journey
The starting flow, task by task in the order a caller meets them: a summary of each, then every step numbered, with each branch and its condition nested beneath and each block's tracking id.
One chapter per kind, each flow with its own summary and its tasks in full.
Integrations and data actions
Grouped by integration: what each data action calls and with which method, what it takes and returns, and for every place it is called, what is sent, where each answer is stored, which steps read it, what is never read, and what happens on failure or timeout.
Where calls end
Every ending in every flow with every condition that leads there.
Flow outcomes and milestones
Each outcome, where it is set as a success or a failure or started, and whether it can end as a failure or is only ever a success.
Appendices
Design observations (blocks that can never run and why, tasks and menus nothing leads to, answers stored but never read, variables read but never set, schedules already in the past); parts of the flows that never run, described in full apart from the journey; gaps; every prompt and every recording without a transcript; participant data written; data tables; variables; scope (every flow and why it is included, versions, and queues chosen while the call runs); permissions and scopes, with every Genesys call made; the block index (every tracking id and the section that describes it).
Coverage check. The preview shows how many of the blocks that can run the document describes, and how many can never run (those are listed in the appendices). Every block that can run must be described; the Word and PDF files are refused otherwise (section 5d).
5dWord and PDF files
On the document page, Download Word and Download PDF produce the finished files. Both are made together,
from the same document the preview shows, so they always agree; the second download is immediate. A large flow
(a few hundred pages) takes well under a minute; the page shows each stage while it works.
Refused when incomplete. The buttons work only when the coverage check shows every block that can run as
described. The server checks again before it writes anything.
Contents with page numbers. The contents list chapters and their sections with the page each starts
on, and every entry is a link. When Word opens the file it offers to update the fields in the document: choose
Yes and Word renumbers the contents to its own page layout. Choosing No keeps the numbers the file
was produced with, which match the PDF.
Cross-references. Every numbered heading is a bookmark. In the block index, each tracking id's section
number links to that section, in Word and in the PDF. The PDF also carries the headings as its bookmark panel.
Style. A4, Arial throughout, royal-blue chapter bars with white text, summary boxes, and tables with a
blue header row that repeats on every page a table runs onto. Every page carries the title, the organisation, the
UTC generation time, Quo Vadis CCS Ltd as the generator, and Page n of m. Steps keep the numbering the preview
shows (1., a., i.), written into the text so it can never renumber itself and stop matching the tracking ids.
Fonts in the PDF. The PDF is laid out on the server, which does not hold a licence for Arial. It uses
Liberation Sans, which has exactly Arial's character widths, so every line and page breaks where it would in Arial.
The Word file names Arial, and shows in Arial on any computer that has it.
When files are made again. The files are kept for the inventory they were written from. Reading the
flow again, a change to the application, or a different organisation name produces new files with a new
generation time.
5eStored data, and deleting it
Stored data, in the header, shows everything Flow Narrator holds on the server for the OAuth client you
signed in with (its region and client id). It is the same for everyone who signs in with that client id, and anyone
who can sign in with it can delete it. Nothing of any other client is shown or touched.
Saved flow readings: one for each flow and version you have read (step 2). Each holds the complete
configuration of the starting flow and every flow it calls, and everything they refer to, with the call routes
and triggers that lead to it. Deleting a reading also deletes the Word and PDF files written from it; reading the
flow again recreates it.
Word and PDF exports: the files produced in step 5. They can be deleted on their own.
Signed-in sessions: each session signed in with this client, which holds the client id, the client
secret (encrypted) and the current access token. Sign out the other sessions ends all but yours.
Delete everything held for this client: type the organisation name to confirm. Every reading and export
is deleted, every session signed in with the client is ended (including yours), the copies held in memory are
cleared, and you are signed out.
Deleting is refused while a flow is being read or a Word or PDF export is being produced for the client: let it
finish first. Deletion cannot be undone. Nothing else about the organisation is kept.
6How it was built
How it was built
Step
What it adds
Status
1 · App shell
Sign-in, the flow picker and versions
Available
2 · The trawl
Reading the chosen version and everything it calls, and an inventory of what was found and why
Available
3 · The model
Conditions in plain English, variables traced from the system that sets them to the decision that uses them, every path to every ending
Available
4 · The writer
The narrative, chapter by chapter, each with its executive summary
Available
5 · The document
Word and PDF output in the QVCCS style, contents, cross-references, the coverage check
Available
6 · Proof
Validation on three differently built flows, and the complete version of this guide
Complete
7Troubleshooting
Troubleshooting
What you see
What it means and what to do
Sign-in fails with a region error
The region chosen is not the one your organisation lives in. Your Genesys Cloud address (for example mypurecloud.ie) tells you which.
Sign-in fails with invalid client
The Client ID or Secret is wrong, or the client is not a Client Credentials grant.
Genesys could not list the flows, with a missing scope
The client's role lacks Architect › Flow › View, or its scopes exclude architect:readonly.
A flow you expect is not in the list
The client's role does not cover that flow's division. Add the division to the role assignment.
The version list is empty
The flow has never been saved or published as a version; there is nothing to describe yet.
QVCCS Flow Narrator — as-built narratives of Genesys Cloud CX flows. Generated documents cite Quo Vadis CCS Ltd. Strictly read-only against Genesys.
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 Flow Narrator 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.