Home · Business · IT, Data & Security · Automation & integrations

PARTLY

As of 13 August 2026, AI can only partly send automated customer follow-ups.

This still needs a person who signs their name to it.

Can you do it?

15 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0/month

Skill neededpower-user

Who has to check ita colleague

What the alternative costsA purpose-built alternative is Botpress, an open platform for building LLM chatbots and agents.

If this goes wrong, customers may receive repeated, mistimed or inappropriate messages and your team may have to investigate the workflow and repair the customer relationship.

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 30 minutes until you can act on the result.

    How to actually do it

    1. Open the customer relationship, helpdesk or messaging system that should trigger the follow-up, and write down the exact trigger event, available customer fields and contact permissions.
    2. Gather your current follow-up policy, approved wording, sending channel, delay rules, opt-out process and the person responsible for approving customer communications.
    3. Paste those details into the prompt, replacing every bracketed slot, and ask the chatbot to produce the workflow, drafts, failure handling and test plan without activating anything.
    4. Build the proposed workflow in Botpress or your existing automation platform, connecting only the required systems and granting the minimum permissions needed to read and send data.
    5. Create dummy customer records covering a normal case, missing data, an opt-out, a failed delivery, a duplicate trigger and a customer who should not be contacted.
    6. Run the workflow in test mode and compare every message, variable, delay, opt-out action and audit entry against your policy and approved wording.
    7. Ask a colleague responsible for customer communications or data protection to approve the tested workflow, then activate it for a small monitored group and check the send logs before expanding it.

    Prompt

    Design a safe automated customer follow-up workflow for [BUSINESS TYPE]. The workflow should start when [TRIGGER EVENT] occurs, wait [TIMING OR DELAY], and send [NUMBER] follow-up messages through [CHANNEL OR BUSINESS APP]. Use only these customer fields: [LIST OF FIELDS]. The purpose is [PURPOSE]. The messages must use this tone: [TONE]. Include an unsubscribe or stop-contact route, avoid sensitive information, and do not contact anyone unless our existing records show that the contact is permitted under our organisation's rules. Do not invent system capabilities, field names, permissions, customer facts or compliance conclusions. First return: 1. the workflow steps; 2. the exact message drafts with variables clearly marked; 3. the required data fields and permissions; 4. duplicate-send, missing-data, failed-delivery and opt-out handling; 5. a test plan using dummy records only; 6. a monitoring and rollback plan. Ask questions where the information is missing. Do not publish, activate or send anything.

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

What it gets wrong

  • AI cannot establish that your customer-contact permissions and opt-out records are complete.
  • AI cannot see undocumented rules in your CRM, helpdesk or messaging account unless you provide them and connect the systems.
  • AI cannot reliably predict every edge case involving duplicate triggers, changed customer details, failed delivery or unusual timing.
  • AI cannot take responsibility for the customer relationship or the consequences of sending an incorrect message.
  • AI cannot replace live monitoring of delivery logs, complaints and opt-outs after activation.

What caps this at PARTLY: consent and privacy, private data access and verification cost.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs1
Verification1
Liability1
Effort delta2
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT send automated follow-up emails?
It can draft the messages and help you design the workflow, but a chat interface does not by itself have permission to send from your business systems. You need an approved automation or agent platform, connected accounts, tested rules and ongoing monitoring.
Can AI automatically follow up with customers?
Partly. AI can handle message drafting, timing logic and much of the configuration, but you must provide the customer data, contact rules and system access, then test the workflow before it sends anything.
Is it safe to automate customer follow-up messages?
Only with controls around permission, opt-out handling, personal data, duplicate sends and failed delivery. Test with dummy records first and have a colleague check the workflow and logs before using it with real customers.
What is the best AI tool for automated customer follow-ups?
Botpress is a suitable purpose-built option because it provides an open platform for building LLM chatbots and agents. It still needs your systems, rules, permissions, testing and monitoring, so the product does not remove your responsibility.

Nearby answers

All Automation & integrations rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.