How it works

The clever bit isn’t one piece of software. It’s how everything works together.

A useful customer conversation can involve WhatsApp, structured automation, AI, business knowledge, CRM or booking data and a human team.

We design the joins, so the customer gets a simple experience even when the process behind it is more sophisticated. We start with what the customer is trying to do. The technology comes afterwards.

Customer

Asks in their own words.

Entry point

Website, QR code, advert, registration.

Conversation

Designed around what they are trying to do.

Automation & AI

Predictable steps, and open questions.

Decision

What should happen next?

A systemA person

Next action

Booked, recorded, answered or progressed.

The same customer travels the whole way through. They only ever see the conversation.

Before any of the technology

First, we work out what should actually happen.

Before we think about buttons, AI Agents or integrations, we map the customer journey.

Where does someone start? What are they trying to achieve? What do they need to know, and what does your team need to know? Which parts happen over and over? Where does it slow down or break?

That gives us the structure for everything that follows.

A team working through a customer journey together

The map comes first. The tools are chosen to fit it, not the other way round.

Behind one conversation

One customer conversation. Several useful layers behind it.

1

Entry point

The customer starts from your website, a QR code, an advert, a campaign, a registration, a booking, or a WhatsApp conversation they already have with you.

2

Conversation design

The questions, choices, prompts and next steps are built around what the customer is actually trying to do.

3

Structured automation

Repeatable stages — confirmations, qualification, reminders, simple choices — happen reliably without someone doing them by hand every time.

4

AI knowledge

Open questions get answered naturally, using approved information about your organisation.

5

Decision & routing

The journey decides what happens next: answer automatically, update a system, trigger an action, or involve a person.

6

Connected systems

Useful information moves into CRM, booking, registration, spreadsheets or operational workflows, where those systems support it.

7

Human team

When judgement, empathy, expertise or an exception is needed, a person takes over with the context already there.

Choosing between them

Not every conversation needs AI.

If someone is choosing a ticket type, confirming an appointment or picking a membership option, structured automation is usually quicker and clearer than asking AI to interpret everything.

Use automation when the next step is known

  • Confirmations
  • Menus and choices
  • Qualification
  • Reminders
  • Simple rules
  • Known next steps

Use AI when the customer needs to ask naturally

  • Open questions
  • Finding answers across approved knowledge
  • Questions that do not fit a button
  • Natural follow-up questions
  • Guiding someone through complex information
  • Recognising when a human should step in

The best journeys often use both.

People using their phones

Approved knowledge

AI is only as useful as what it knows.

We do not connect an AI model and hope it knows your business. Capture 1 builds an approved knowledge layer around the information the Agent is allowed to use.

Website pagesFAQsPoliciesProduct and service informationEvent agendas and exhibitor informationHotel or venue informationPDFs and internal guides

A real question

“Are dogs allowed at the ceremony?”

Answered from

Venue websiteWedding FAQsHouse policyGuest guide

Well-behaved dogs are welcome for the ceremony and outdoor photographs, but not in the dining room. Most couples arrange for someone to take them home after the drinks reception.

Outside the boundary

We also decide what the AI should not answer, how uncertain questions are handled, and the point at which the conversation moves to a person — rather than letting it guess.

The handover

Good automation knows when to stop.

A handover is not a failure of the system. It is part of the system. If a question is unusual, commercially important, sensitive or simply needs judgement, the conversation moves to the right person.

Two colleagues picking up a customer conversation on a phone

The customer should not have to email again and explain everything from the beginning.

What the team can see

Where it came from

Website — “Ask us on WhatsApp”

What the customer asked

Pricing and availability, September

Already gathered

Date, 90 guests, accommodation needed

Already answered

Dogs at the ceremony

Why it was handed over

Specific pricing needs the weddings team

What happens next

Then connect the conversation to what happens next.

A WhatsApp conversation is more useful when it can support the systems and workflows already behind the business.

  • Create or update a CRM contact
  • Use registration or booking data to start a journey
  • Record qualification information
  • Update a status or a spreadsheet
  • Trigger a reminder or follow-up
  • Route a conversation to the right team

Not everything needs connecting just because it can be

The exact route depends on the systems involved. We use the simplest reliable one, whether that is a direct integration, Make, Zapier, an API or a lighter data workflow.

An outdoor wedding ceremony setting

One evening, one enquiry

What that looks like in a real journey.

A wedding enquiry arriving on a hotel website at 8.45pm. For the customer it feels like one conversation.

8:45pm

Interest

Hi — we’re looking at your barn for a wedding.

Behind it

The visitor taps “Ask us on WhatsApp” on the weddings page. The entry point tells the journey which campaign and page they came from.

8:46pm

Respond

Lovely. Roughly what date, how many guests, and would you need bedrooms?

September next year, around 90, and yes please.

Behind it

Structured automation. Three questions that change what happens next, asked one at a time.

8:49pm

AI question

Are dogs allowed at the ceremony?

Behind it

Not a question anyone put on a menu. The AI answers from the venue’s approved information.

8:52pm

Decision

What would that come to for a Saturday?

Behind it

Specific pricing and availability need the weddings team. The journey stops rather than guessing, and says so.

Next morning

Human

Morning — I can see you’re looking at September for 90 with bedrooms. Here’s what a Saturday looks like…

Behind it

The weddings team receives the conversation with the date, the numbers, the requirements and the questions already answered.

And then

Next action

Behind it

The enquiry is recorded in the CRM and progressed into a show-round or a sales conversation.

Behind one conversation, several different parts of the system have each done the right job.

01

Map

The customer journey as it is now, and where the gaps are.

02

Design

Where WhatsApp, automation, AI and people should each fit.

03

Prepare

Approved knowledge, content, rules and handover boundaries.

04

Build

The journeys themselves, and the systems they need to connect to.

05

Test

Realistic customer questions, edge cases and failure scenarios.

06

Launch

Into live use, with the team workflow that goes around it.

07

Improve

Review real behaviour, update knowledge, refine the journey. This one does not stop.

The managed service is a big part of the value, because customer communication does not stand still.

Before it goes live

We test the awkward questions too.

A perfect demo journey is easy. Real customers are less predictable.

The aim is not to make the journey clever. It is to make it dependable.

What testing has to include

Incomplete answersPeople changing their mindDuplicate recordsUnusual wordingQuestions the AI should refuse to guess atA system being unavailable

The people side

Your team still matters.

Automation should make your team more useful, not make customers feel trapped. So we design the technology around the real people who receive, progress and manage these conversations.

  • Who needs access?
  • Which enquiries go to which team?
  • What context should they see?
  • How do they know a conversation is waiting?
  • What status or notes are useful?
  • What should happen when the conversation is finished?

Where WhatsApp sits

WhatsApp is often the conversation layer, not the whole system.

We like WhatsApp because it is familiar, immediate and low friction for customers.

But it should work alongside your website, email, apps, registration systems, booking tools and human teams. The aim is not to move everything into WhatsApp. It is to use WhatsApp where it makes the journey easier.

What should change

What should improve when the system is working properly?

Faster answers

Customers get useful information at the moment they need it.

Less repetition

Routine questions and predictable actions stop consuming the same manual effort.

Better handover

People join conversations with more context than a name and a number.

More consistent communication

Reminders and follow-ups happen because the journey is designed to trigger them.

Better visibility

You learn what customers actually ask, and where the journey needs improving.

Where to begin

It does not have to start as a huge project.

Some of the best projects begin with one clear friction point.

Start there. Prove the journey. Then extend it where the value is obvious.

Good places to start

Too many repeated questionsWebsite enquiries waiting overnightA registration-to-attendance gapGuest information scattered aboutManual confirmations and remindersA poor handover between automation and people

Questions we get asked

How it works: common questions

Do we have to replace our existing systems?

Usually not. We normally design WhatsApp and automation around the systems you already use, connecting them where there is a useful reason to do so.

Does every project use AI?

No. Structured automation is often better for predictable stages. AI is added where open questions or complex knowledge make it useful.

How does the AI know our business?

We create an approved knowledge layer from the information you want the Agent to use: website pages, FAQs, PDFs, policies and service information.

Can the AI make things up?

Any AI system needs clear boundaries. We reduce the risk by grounding answers in approved knowledge, defining what should not be answered, and building a human handover route for uncertainty and exceptions.

Can our staff take over conversations?

Yes. Human handover is designed into the journey, with the preceding context available so the customer does not have to start again.

Can it connect to our CRM or booking system?

Potentially. The route depends on the system, its integration options and what information genuinely needs to move.

Who updates the AI knowledge?

Capture 1 can manage the knowledge and the journeys as part of the ongoing service, working with you when information, policies or services change.

How quickly can something like this be built?

It depends on complexity, integrations and the amount of knowledge involved. A focused first-stage journey is usually much quicker to launch than a large multi-system programme.

Start with a conversation

Show us the journey you have now.

You don’t need to know which software you need.

Show us where the customer starts, what they ask, what your team does next and where things become slow or disconnected. We’ll help you work out what should be automated, where AI can help, what should connect, and where a person should stay involved.