Cue Flows
Customer service automation that hands over properly
Flows are the automation layer in Cue. They answer the enquiries that repeat, collect what your team needs before anyone picks up, and reach into the systems that hold your real answers.
Anything that needs a person goes to the right person, with the whole conversation attached. Automation carries the volume. Your team keeps the conversations that need them.
Message received
Service enquiry
Reply and create ticket
Immediate, with a reference
If condition
Booking, order or something else
Needs a lookup
HTTP request
Open slots from the system
Knowledge answer
Transfer to AI Agent
Returns a reason
Transfer to team
With everything collected
Automation running for customer teams including
The category
What is customer service automation?
Customer service automation is software handling parts of a customer enquiry without a person doing the work: replying to common questions, collecting details, checking information in another system, routing the enquiry to the right team, and closing it off. Most customer service automation is not one tool. It is a set of decisions about which enquiries a person needs to see and which ones they do not.
In Cue, those decisions live in a Flow: a customer journey you design once and publish to a channel. Flows run across WhatsApp, web chat, email and Messenger, and they are what turn Cue from a shared inbox into an operation that scales.
The problem
Your team is answering the same twelve questions all day
Look at a week of your queue and the shape is always the same. Where is my order. Are you open on Sunday. Can I move my appointment. How do I return this. Your team answers them competently, hundreds of times, and every one of those answers is time not spent on the enquiry that actually needed a person.
Then there is the second cost. The enquiries that do need a person arrive with nothing attached. No reference, no order, no idea what the customer already tried. So the first thing your agent does is ask, the customer repeats themselves, and a five-minute job becomes fifteen.
£7.3bn
a month is what failures to get service right first time cost UK businesses, along with about four lost productivity days per full-time employee a year.
Institute of Customer Service, UKCSI, January 2025
This morning’s queue
Live
Where is my order
Needs a lookup
x31
Opening hours
Knowledge answer
x18
Complaint, needs judgement
Arrived with nothing attached
x4
The mechanism
What a Cue Flow actually does
A Flow is built in a visual builder as a sequence of steps. You can see the whole journey, follow any branch, and change it without waiting on a release.
Send messages, ask questions, validate the answers, and end the conversation cleanly when it is finished. On WhatsApp a Flow can request the customer’s phone number directly.
Step family
What it gives you
Conversation
Message, request phone number, end conversation
Logic and navigation
If condition, go to
Data
Set value, delete value, update contact, add or remove from list
Integration
HTTP request, function
Human service
Create ticket, transfer, close ticket, out of office
AI
Transfer to AI Agent
The difference
Where the automation stops and a person takes over
This is the part most automation pages skip, and it is the part that decides whether automation works in your operation. Automation should not try to resolve everything, because it cannot.
In Cue the Flow owns that decision. When an AI Agent has answered what it can, it returns control to the Flow with a reason. The Flow then decides: try a different branch, look something up, create a ticket, or transfer to the right team.
Because the handover is a designed step rather than an accident, the person who picks it up gets the conversation as it happened plus everything the Flow collected on the way.
AI Agent returned control
Reason
confirmed_unresolved
Outside what the AI Agent knows
step 4
Ticket created
Transcript and collected details attached
step 5
Transferred to Service team
Kirsty M picked up
step 6
14%
of service issues are fully resolved in self-service, rising to 36% for issues customers themselves consider very simple. Any platform promising near-total automation of a real queue is describing a queue that does not exist.
Gartner, August 2024
Integrations
A Flow is only as good as the systems it can reach
This is the difference between a chatbot and a workflow layer. A chatbot can answer what it has been told. A Flow can go and find out.
Most enquiries that clog a queue are not knowledge questions. Where is my order. Is the part in. Can I move Thursday to Friday. Nobody answers those from a help centre article, because the answer lives in your DMS, CRM or booking system.
At any point in a journey a Flow can make an HTTP request to your system’s API, take what comes back, and use it in the rest of the conversation. It can fetch a record so the customer gets the real status, create or update a contact or ticket as the conversation happens, write the transcript back to where your team already looks, and push an event so another tool carries the work on.
Lookup inside the conversation
Resolved
Flow collects the reference
Registration or order number
step 1
HTTP request out
Open slots returned from the system
step 2
Answer used in the reply
Contact updated, ticket closed
step 3
A real answer, not a promise to look into it
Getting started
Four Flows most teams build first
1
Acknowledge and triage, so nothing sits unanswered
Every enquiry gets an immediate reply, a ticket and a route. The Flow asks what the enquiry is about, branches on the answer, and either resolves it or transfers it to the team that owns it. Out-of-office behaviour handles anything arriving outside your hours.
Message
If condition
Create ticket
Transfer
2
Collect what your team needs before they pick it up
Instead of an agent opening a conversation and asking for the order number, the Flow asks first. Answers are stored against the conversation, and the ones worth keeping are stored against the contact so nobody asks twice next time. The agent starts with the facts.
Message
Set value
Update contact
Transfer
3
Look something up in the system that holds the answer
A large share of enquiries are one lookup away from resolved. A Flow can call your CRM, DMS or booking tool with an HTTP request, take the response, and use it in the conversation: confirm the booking, give the status, quote the balance, offer the available slots.
Message
HTTP request
Function
Close ticket
This is the highest-value Flow most teams build, and the one that needs real design work. See who builds a Flow below.
4
Close the loop after resolution
When a ticket is closed, a resolution Flow runs. Thank the customer, ask the one question you actually want answered, and push the outcome and a transcript to the system where you keep your customer record.
Close ticket
Message
HTTP request
Verified customer result
Stellantis dealerships lifted survey response rates from 26% to 77% by following up over WhatsApp through Cue, and customers reached that way scored up to 12 NPS points higher.
Channels
The same Flow logic, on every channel your customers use
The logic you learn is the same everywhere: ask, branch, look up, decide, hand over. What changes is which steps are available and what the channel allows.
The deepest Flow capability of any Cue channel, and the one most customers reply on. In the UK, 74% of online adults use WhatsApp daily (Ofcom, 2026). A WhatsApp Flow can request the customer’s phone number as a step, update their contact record, and run the whole journey from first reply to handover.
It is also the only channel where a Flow can pick up the replies to an outbound broadcast, so a campaign and a conversation are the same thread rather than two systems.
Example Flow, service booking
Customer messages, immediate reply, Flow asks for registration, HTTP request checks open slots in the DMS, customer picks one, contact record updated, ticket closed. Nobody touches it.
Message
Request phone number
HTTP request
Close ticket
What changes between channels
Most vendors leave this vague. It is more useful stated plainly, because it changes what you design.
Web chat
Messenger
Run a Flow
Yes
Yes
Yes
Yes
Branch, look up data, route to a team
Yes
Yes
Yes
Yes
Request phone number as a step
Yes
No
No
No
Update the contact record
Yes
Yes
Yes
Not documented
Hand over to an AI Agent
Yes
Yes
Yes
Yes
Proactive outbound
Broadcasts and approved templates
Not evidenced
Not evidenced
Individual emails only
If a journey depends on a step that is not available on your channel, that is a design decision to make before you build, not a surprise to find afterwards.
Teams
Which teams get the most out of Flows
Flows are built for customer service, and that is where they do the most work. But a Flow is really a way of handling an inbound conversation properly, and more than one team is doing that.
Customer service
The primary case, and the deepest. Flows carry the enquiries that repeat, collect what an agent needs before they open the conversation, look up the answer in the system that holds it, route what is left to the right team, and follow up after the ticket closes. Out-of-office behaviour covers the hours you are not open.
What it changes: the same team covers more customers, response times stop depending on how busy the morning was, and the conversations that reach a person arrive with the facts already attached.
Example Flow
Acknowledge, triage by enquiry type, AI Agent answers what it can, unresolved goes to the right team with the transcript and collected details, resolution Flow closes the loop.
Message
If condition
Transfer to AI Agent
Transfer
Verified customer results
Affinity Health reduced customer service costs by 73% while managing more than 17,000 customers a month. YourBusinessNumber cut manual customer service work by more than 80%.
Impact
What changes when Flows carry the volume
The same team covers far more customers
When the repeatable enquiries are handled automatically, headcount stops being the limit on volume.
Proof
Affinity Health reduced customer service costs by 73% while managing more than 17,000 customers a month. Baroka Funerals cut costs by 40%.
Manual work drops, and what is left is worth doing
Flows take the repetitive collection and routing off people.
Proof
YourBusinessNumber cut manual customer service work by more than 80%. In Gartner’s October 2025 survey, 55% of service leaders kept staffing stable while handling higher volumes, against 20% who reduced staffing.
Customers get answered immediately, at any hour
Automation replies while your team is asleep, and holds the conversation until they are back.
Proof
Pace Car Rental brought average response times down to 5 to 10 minutes. Dream Drive’s automation resolved 96% of the queries it handled, from more than 50 enquiries a day.
Customer results are specific to each business, its channel mix and its enquiry types. They show what is achievable, not a guaranteed outcome.
Effort
Who builds a Flow, and how much work is it?
Worth answering plainly, because the answer depends on the Flow.
Your team builds these
Journeys inside Cue
Messages, questions, branching, data capture, routing, ticket creation, handover, out-of-office, resolution and feedback. Built visually by the person who owns the queue, and changed the same way when your process changes. No developer needed.
This is a build
Journeys that reach into other systems
An HTTP request needs an endpoint, credentials and a JSON payload. A Function step is JavaScript. Error handling is designed, not assumed. Someone technical is involved, whether that is your team, your systems provider or Cue.
Cue is honest about this rather than calling it no-code, because a buyer who is promised no-code and then handed a JSON payload has been sold something. The visual builder is genuinely visual. Deep integration is genuine engineering work. Knowing which is which before you start is what keeps an automation project on schedule.
Flows and AI
How Flows work with Cue AI Agents
People assume these are the same thing. They are not, and the difference is worth understanding before you design your automation.
A Cue AI Agent answers questions from knowledge you give it: your documents, help content and website material. It composes an answer, clarifies where it needs to, and then returns control with a reason for what happened.
A Flow is the journey around it. The Flow decides whether the AI Agent is involved at all, what happens with the reason it returns, when data gets looked up, which team receives the conversation, and when the ticket closes. The Flow is in control.
Use an AI Agent for the questions where the answer is written down somewhere, and a Flow step for anything that has to look something up, write something back or reach a person.
Cue AI Agent
Cue Flow
Job
Answer from your knowledge
Run the journey and decide what happens
Good at
Open questions with a documented answer
Structured journeys, data, routing, handover
Not for
Checking a live order or updating a record
Composing an answer to an unpredictable question
Controls the outcome
No, it returns a reason
Yes
Flow FAQs
Common questions about Cue Flows
Automation and Flows
Customer service automation is software handling parts of a customer enquiry without a person doing the work, such as replying to common questions, collecting details, checking information in another system, routing the enquiry to the right team and closing it off. In Cue this is done with Flows, which are customer journeys you design in a visual builder and publish to a channel.
Integrations and systems
Yes, through an HTTP request step that calls your system’s API inside the customer journey, and through webhooks that push data out to other tools. Cue has working integration patterns with CRM platforms including HubSpot, Salesforce and Zoho, helpdesks including Zendesk and Freshdesk, and for automotive with dealer management systems including Keyloop.
Got a question that is specific to your queue? That is a five-minute conversation.
See a Flow built for your queue
Bring us your actual enquiry mix and we will show you which parts Flows can carry, where a person still needs to take over, and what that means for your response times and your team’s day.