A lead follow-up flow, end to end.

Now we put it all together into the one flow that pays for itself fastest in almost any business: enquiry follow-up. Around half of enquiries go to whoever replies first, and most that hit silence for a few hours go cold. This automation makes sure every enquiry gets read, recorded, acknowledged in seconds, and flagged if it's hot, while you sleep. If you want the business case before the build, our guide on how to automate lead follow-up lays out the why. Here's the whole thing, step by step.

The flow, in one sentence

When a new enquiry comes in, an AI step reads and tags it, it gets saved to the CRM, the customer gets an instant acknowledgement, and anything urgent or high-value gets flagged for a person to ring. That's the blueprint from lesson one, just with more steps. Let's build it.

Step by step

• 1. Trigger: new enquiry. Your contact form or a new email is the start. The scenario fires the moment it lands and pulls in the name, email and message.
• 2. AI: summarise and classify. Map the message into an OpenAI or Claude step. Ask it for two things: a one-line summary, and a category from a fixed list, say sales, support or other, plus a simple urgency read like hot or normal. Use the tight prompt recipe from lesson three so the output is clean.
• 3. Save to the CRM. Create or update the contact, and write the AI's one-line summary into the notes. Now whoever picks it up sees the gist at a glance instead of wading through the raw message.
• 4. Acknowledge the customer. Send an instant, plain email: "Thanks for getting in touch, we've got your enquiry and someone will be back to you shortly." Short, warm, honest. This is the step that holds the lead.
• 5. Flag the hot ones. Add a filter: if the AI tagged it sales and hot, post an alert to Slack or fire a text to the owner so a human can call quickly. Everything else just sits tidily in the CRM for the next working pass.

Five steps, and you've turned a leaky manual process into a reliable one. Notice what each part is doing: the AI does the reading and sorting, the CRM keeps the record, the email buys you time, and the flag points your attention at the leads worth a fast call.

Why the instant acknowledgement matters so much

This one step does the heavy lifting on revenue. When someone enquires and hears nothing for hours, they assume you're not interested or not around, and they message the next business on their list. A reply within seconds, even an automated one that simply says "got it, we'll be in touch," keeps you in the running. It sets the expectation, it reassures them, and it holds the lead steady until a person does the real follow-up. You're not faking a personal reply; you're acknowledging quickly and being upfront that a human will follow.

Keep it honest, keep it safe

A couple of guardrails so this stays a help and not a liability:

• The auto-reply only acknowledges. Don't let AI send a full personalised sales pitch on its own. The instant email is a short, fixed acknowledgement; the real follow-up is a person, who can use the AI summary to move fast.
• Only message people who enquired. This flow contacts someone because they just got in touch with you, which keeps it useful and on the right side of the Spam Act. Don't bolt a bought list onto it.
• Test it like lesson two. Run one real-looking enquiry through and watch all five steps: summary sensible, contact saved, acknowledgement received, hot ones flagged. Fix and rerun until it's clean.

The same shape, reused

Once this clicks, you'll see the pattern everywhere. The exact same shape, trigger, AI read, record, respond, flag, becomes an AI-sorted support inbox (classify and route each email to the right person), or a weekly digest (a scheduled trigger that gathers the week's rows and has AI summarise them into one tidy email). Build this one well and you've built the template for the next three. Last lesson: making it reliable, what it costs, and knowing when you've outgrown the no-code tools.