Skip to content

Automation · Customer service and scheduling

Answer on WhatsApp with the help articles your team publishes in Contentful

An AI agent searches Contentful for the help article or product page that answers the customer's question and sums it up on WhatsApp.

How do I make my WhatsApp answer with the help articles I already have in Contentful?

When a customer asks on WhatsApp how to do something, how a warranty works or what a product includes, the Met searches Contentful for the help article or page that answers that question, first by meaning and, if nothing comes up, by text. It opens the entry, gets the entries it links to, such as the steps or the related product, and replies with what your team wrote. If the article has an image or a PDF, it shares that file. When no article answers, it does not make one up: it hands the conversation to a person. It is all reading: the procedure does not change or publish anything in Contentful.

What starts it

When the customer writes

It starts when a customer writes on WhatsApp with a question about a product or a process. There is no trigger to configure: the Met assigned to the line answers it in that conversation.

Nothing to set up. The Met assigned to that channel picks up the conversation and runs the procedure inside it.

Who runs it

Support Met with your help base in Contentful

Customer service team

This is not a named, seeded Met. It is the Met that already answers the business's WhatsApp, with Contentful connected: its work ends when the customer gets the answer that is in your help articles, or when someone on the team takes the conversation.

Run by the customer support Met

Channels: WhatsApp

What you need connected

6 steps · 7 tools

The procedure, step by step

Each step shows the tool that runs. They belong to the real connector: if one of them stopped existing, this page would not build.

  1. 1

    Find where the help lives

    The Met lists the environment's content types to find the one your team uses for help, such as «Help article» or «Product», as the skill says.

  2. 2

    Search the article by what the customer meant

    The Met searches entries by meaning, so it finds the article even when the customer uses other words. If nothing comes up, it searches and filters by text inside the help content type.

  3. 3

    Read the full article

    The Met opens the entry it found and reads its full content, so it answers with what the article says and not with a search summary.

  4. 4

    Get what the article links to

    If the article links other entries, such as the steps of a process or the product it covers, the Met resolves those references and adds what the customer needs to the answer.

  5. 5

    Share the image or the PDF

    If the entry has a file, such as the PDF manual or the product photo, the Met gets that file and shares it in the conversation along with the answer.

  6. 6

    Hand off to a person when there is no article

    If no entry answers the question, the Met does not improvise: it escalates to someone on the team with the reason and leaves the conversation.

Before you start

What you need ready

The Met solves none of these for you. If one is missing, the automation stops there.

The limits

What this automation does NOT solve

Better said here than discovered halfway through the rollout.

FAQ

Questions about this automation

The tool exists on the server, but this procedure assumes nothing about your space: if it returns no results, the Met moves on to text search.

If your space has the help in several languages (locales), the Met can read the version that fits the customer. Tell it in the skill which locale to use.

Not in this procedure, which only reads. If you want the Met to update entries, the admin leaves those actions on in Connections, where they arrive on, and the Contentful MCP app has to allow writing.

Keep going

Other automations built on the same tools

Want this automation running on your own data?

We will show it to you with your accounts connected, not with a canned demo.