As of 13 August 2026, AI can only partly automate your customer follow-up emails.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, power-user skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Open the CRM, helpdesk or sales system and write down the exact customer event, available fields, communication-preference field and status values that should start or stop the follow-up.
   2. Export a small set of fictional or properly anonymised test records containing each relevant status, including an opted-out customer and records with missing fields.
   3. Open the email platform or automation tool and record its available triggers, customer fields, suppression rules, sender settings, reply address and delivery reporting.
   4. Paste the business details, system details, fields, consent rules and timing requirements into the prompt, then ask the model to mark every missing setting as NEEDS INPUT.
   5. Build the workflow in the selected automation platform from the model's trigger, condition, delay and stop-rule list, without connecting live sending yet.
   6. Run the workflow against the test records and compare each recipient, timing decision, personalisation field, link and stop rule with the expected result before enabling delivery.
   7. Send approved test emails to internal addresses, compare the wording against the current product and privacy information, then have a colleague approve the recipient rules and email sequence.
   8. Enable the workflow with monitoring and rollback available, check the first delivery reports and pause it immediately if an unexpected recipient, duplicate or message appears.

   Prompt

   Design a customer follow-up email automation for a UK business using the information below. Do not send anything or claim that an integration has been completed.
   
   Business and product context:
   [describe the business, customer journey and product or service]
   
   Customer event that should start the workflow:
   [for example, a purchase, enquiry, abandoned booking or support resolution]
   
   Systems available:
   [CRM, helpdesk, ecommerce system, email platform and any integration tools]
   
   Fields available for each customer:
   [list only the fields that are actually available]
   
   Follow-up objective:
   [what the email should help the customer do]
   
   Timing and stopping rules:
   [when to send, when to stop, and what customer actions should cancel later emails]
   
   Brand and message rules:
   [tone, sender name, reply address, links, exclusions and required wording]
   
   Consent and privacy constraints:
   [what consent or communication preferences are recorded, and which customers must be excluded]
   
   Produce:
   1. A plain-English workflow with triggers, conditions, delays, branches and stop rules.
   2. A three-email sequence, with a subject line, body, call to action and personalisation fields for each email.
   3. A list of required fields and integration settings, marking anything missing as NEEDS INPUT rather than guessing.
   4. A test plan using fictional records, including duplicate events, missing fields, unsubscribed customers, changed customer status and failed delivery.
   5. A launch checklist that requires human approval of recipients, consent rules, links, sender details and sample emails.
   6. A monitoring and rollback plan.
   
   Use no invented customer facts, dates, prices, discounts, legal claims or system capabilities. Keep the messages suitable for UK customers. Flag any point that needs checking against the business's privacy notice, email policy or a qualified professional before launch.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.