A home insurance quote form can run to forty questions. A renewal pack arrives as a PDF nobody opens. A claim starts with a phone queue. WhatsApp Flows for insurance replace each of those with a short, structured form inside the chat your customers already use, and the answers can go straight to the systems that price, renew and settle.
This guide is for insurers, MGAs and brokers who want more than a chatbot that sends PDFs. It covers five journeys step by step, shows how to connect a Flow to your policy and rating systems through an API, n8n or an AI assistant, and sets out the design rules that keep a regulated journey safe.
The short answer. A WhatsApp Flow collects quote, renewal or claim details as taps and short fields inside WhatsApp. The answers arrive as structured data, so your own systems can price the risk, update the policy or open the claim, and the result comes back in the same chat. Start with one journey, usually renewals, and connect it to the API you already have.
Insurance asks for more information than almost any other purchase, and most of it is now requested on a small screen. Zuko's 2025 benchmark of more than five million form sessions puts insurance form conversion at 47.8% on mobile, against 56.8% on desktop[1]. Roughly half the people who start on a phone do not finish.
WhatsApp is where those same people already are. Ofcom's Online Nation 2025 report found that WhatsApp reached 90% of UK online adults in May 2025, and 74% used it every day[2]. In a 22-market study for Meta, Kantar found that 73.3% of consumers prefer messaging when they communicate with a business[3].
We learned this first hand. Flowella grew out of our team's work on CoverKraft, an insurance platform where quote requests, claim details, onboarding forms and renewals are captured in WhatsApp and sent to the platform as clean, structured data. The tool we built to do that became Flowella.
WhatsApp Flows are Meta's native forms: screens that open inside the chat with dropdowns, date pickers, option lists, text fields and photo or document upload. They are different from a chatbot, which asks one free-text question at a time and leaves someone to interpret the replies. We covered that difference in the real problem with WhatsApp chatbots.
| What happens | Web form or chatbot | WhatsApp Flow |
|---|---|---|
| Where the customer answers | A browser tab, a login or a long chat | A form inside the chat they already have open |
| The data you receive | Free text someone reads and retypes | One field per answer, ready for your systems |
| Evidence for a claim | Emailed photos, matched by hand | Photos and documents uploaded in the form |
| The record | Scattered across inboxes and phones | Each answer on the customer record, with the time it was given |
Meta treats insurance as a core use of Flows. Its developer documentation includes a full "Insurance Quote" guide, with screens for applicants, cover level, excess, details, policy selection and a quote summary[4]. One detail in that guide matters for anyone planning a real build: the prices in Meta's sample are fixed values written into the demo code. Connecting the form to a real rating engine is left to you. The rest of this guide is about that part.
Each journey below follows the same shape: something starts it, a Flow collects the facts, your systems do the insurance work, and the customer gets the result in the chat.
This is the journey we suggest starting with. Your system already holds the risk details, so the Flow is short, and a change told to you before renewal is better for everyone than one discovered at a claim. Our renewals workflow guide shows the HubSpot build.
Always leave a route to a person. In a 2026 KPMG UK survey, 90% of adults said human interaction is important when a claim is handled[8], and WhatsApp's own policy requires a clear way to reach a human[6].
There are two ways to put live insurance data behind a Flow. They suit different systems, and most insurers should start with the second.
Meta lets a Flow call your own server between screens. The customer taps Continue, WhatsApp sends the answers so far to your endpoint, and your endpoint replies with the next screen, such as a list of plans and prices[4].
It is the smoothest experience, and it sets a high bar for your systems. You build and host an encrypted endpoint yourself. Meta asks for a response in under a second and gives up after ten. If your endpoint is slow or unavailable, Meta first limits the Flow to ten sends an hour, then blocks it[10]. A rating engine that calls a panel of insurers, or a policy system with a maintenance window, will struggle to meet that.
The second pattern keeps the Flow simple and does the slow work between messages. The customer is never waiting on a spinner, and a slow or failed call never puts the Flow itself at risk.
| Route | How it works | Best for |
|---|---|---|
| HubSpot workflow | The Form Completed event starts a workflow. A Send a webhook step passes the answers to your API, and Flowella's Send WhatsApp Template action returns the result. | Teams already on HubSpot Professional with Data Hub |
| n8n | A Webhook node receives the answers, an HTTP Request node calls your rating API, and a second HTTP Request node calls the Flowella API to send the quote. | Teams that want a visual, self-hosted layer between systems |
| Your own code | Your service verifies Flowella's signed webhook, calls your systems and uses the REST API to reply. | Insurers with an integration team and an API gateway |
The Flowella side of each route is the same handful of calls, documented in the API reference[11]:
POST /v1/flows takes a list of fields and where the answers should go: a HubSpot form, the Flowella contact, or your own web address.POST /v1/templates/send sends an approved message, filled with values such as the customer's name, the price and the renewal date. A button on it can open a Flow.X-Flowella-Signature header, so your system can check it came from Flowella.A request that sends a quote looks like this:
POST https://api.flowella.io/v1/templates/send
Authorization: Bearer flo_your_api_key
{
"templateId": "your_quote_ready_template_id",
"whatsappChannelId": "your_whatsapp_number_id",
"recipients": [
{
"phone": "+447700900417",
"bodyParameters": [
"Maya",
"buildings and contents",
"£284.16",
"30 November"
]
}
]
}On n8n: there is no native Flowella node yet. It is on our roadmap. Today the connection uses n8n's own Webhook and HTTP Request nodes, and the interactive API docs include an n8n quickstart.
Flowella also has an MCP server, which lets AI assistants such as Claude, ChatGPT and Cursor work with your account[12]. It is a building and operations tool, not the thing that prices a quote. A product owner can ask an assistant to draft the renewal messages, create the Flow, check which templates Meta has approved or report on delivery, and the assistant must confirm before it sends or changes anything. The pricing call still belongs to your rating engine.
Any system with an API you are licensed to use. Some publish their documentation openly: Socotra documents a quotes API that creates, prices and underwrites a quote[13], and the DVLA publishes its vehicle lookup. Many core and broker platforms keep their API documentation for customers and partners. Flowella does not ship ready-made connectors to those platforms, so the right question for your supplier is which quote, policy and claim calls your licence already includes.
This is not legal advice, and your compliance team will have its own view. These are the points we see matter most.
We would rather be straight about the evidence. There is no published UK insurer case study built on WhatsApp Flows yet. What exists points in one direction:
All of these are self-reported, and only the last two involve Flows. Treat them as reasons to run a measured pilot of your own, not as a forecast.
If you use HubSpot, connecting WhatsApp to HubSpot takes one sitting, and your first Flow can come from an existing form.
Yes. Either the Flow calls your rating engine between screens through an endpoint you host, or the Flow collects the answers and your system returns the price as the next message. The second route is easier to run reliably.
Yes. WhatsApp Flows include photo and document upload fields, so the evidence arrives with the answers it belongs to[7].
No. HubSpot is Flowella's native integration, and Flows can be built from HubSpot forms on any plan. Other systems connect through the REST API and signed webhooks.
Yes. Insurance is not on WhatsApp's list of restricted industries. The limits are on what you ask for in the chat, such as card, account and ID numbers, and on how you obtain consent[6].
Not out of the box. Flowella provides the Flow, the messaging and an open API. The connection to your policy, rating or claims system is built with that API, directly or through a tool such as n8n.
Insurance runs on accurate information, and WhatsApp Flows are a better way to collect it than a forty-question web page or a phone queue. Pick one journey, connect it to the system that already does the insurance work, and measure what changes. We are at Insurtech Insights Europe in London in March 2027, and happy to walk through a working quote or renewal Flow before then.
Flowella turns your insurance forms into WhatsApp Flows, with an open API for your policy and rating systems.