Which tool to choose
Two pieces, two tools.
- To write the sequence: ChatGPT or Claude. Claude better maintains a consistent, non-pushy tone across several messages.
- To keep track of who needs to be contacted again and when: a Google Sheets sheet or a table in Notion are enough to start. If you manage many contacts, a CRM (Customer Relationship Management, the customer management system) automates the reminders, but you don't need it from day one.
How to do it
From computer or phone, only where you keep the table changes; the logic is the same.
- Define the starting point. Are you following up with someone who received a quote, someone who bought once, or a cold contact? The follow-up changes tone depending on the moment.
- Have the sequence written. The operative syntax:
Act as a sales consultant. Write me a follow-up sequence of 4 messages for a customer I sent a quote to and who hasn't replied yet. For each message indicate: after how many days to send it, the recommended channel (email or message), and the ready text. Each message must add a new reason to reply — a useful question, a benefit, a real deadline — without ever repeating "I'm following up to check in". The last one closes gracefully, leaving the door open. Friendly tone, never nagging. I sell bathroom renovations to private clients.
- Build the tracking table. Ask the AI: "make me a table with columns customer, date of last contact, next step, date of next step, status". You copy it into the sheet.
- Work the table every morning. You look at who has today's date, send the already-prepared message, update the status. Five minutes, not an afternoon.
A concrete example
Elena has a graphic design studio and loses jobs because she sends the quote and then "doesn't want to bother". With Claude she generates a sequence of four messages: the first, after three days, asks whether the quote is clear, the second, after a week, adds an example of a similar job done, the third offers a fifteen-minute call, the last closes leaving the door open.
She puts the customers in a Google sheet with the date of the next contact. In the first month she follows up on nine stalled quotes: three turn into jobs, two had gone to the competition, the others stay warm. Before, those nine would have stayed without even a second message. The system didn't sell for her: it took away her "I'm not bothering" excuse.
When it does NOT work (and how to fix it)
If you seem nagging
If the messages arrive too close together or sound like reminders, the customer shuts down. Widen the intervals (three, seven, ten days) and ask the AI to rewrite each message "giving value before asking": a useful tip gets more opens than an "I'm writing again".
If all the messages seem the same
It's the sign that different reasons are missing. Impose in the prompt that each message brings something new — a real case, a question, a deadline — and that none repeats the previous one. A sequence is a conversation, not the same message sent four times.
If you don't keep track and forget
Without the table, the system collapses after a week. If you really can't keep it updated, ask the AI to turn the sequence into calendar reminders with precise dates, so it's the phone reminding you instead of your memory.
A tip from someone who really uses it
Write a single well-made sequence and reuse it as a template for everyone. When a customer replies mid-sequence, you paste their reply into the AI and ask for the next message, tailored: you start from a solid base and personalize only where needed, instead of inventing from scratch every time.
Frequently asked questions
How many follow-ups are too many?
As a reference, a sequence of four or five contacts covers most cases: the bulk of replies arrive between the first and the fourth. Beyond that, the risk of annoying outweighs the benefit. Better to close gracefully and reopen later with a new reason.
Can AI send the follow-ups on its own?
To send them automatically you need a CRM or connected automation tools. To start, the manual system is better: you see the reactions, understand which messages work and fix the texts before automating something you haven't yet tested.
Doesn't an automatic follow-up system ruin the human relationship with the customer?
It's the right fear, and it debunks a myth: the follow-up doesn't replace the relationship, it keeps it from fizzling out through forgetfulness. AI gives you the structure so you don't lose anyone; you build the relationship when the customer replies. A timely, useful message is attention, not cold automation — as long as there's a person behind it ready to really reply.