AI can help interpret information or prepare a draft. Automation can move information or start a task according to agreed rules. A process can use either or both.
An example
A sales enquiry arrives through a website form. AI could help sort it and draft a reply. Automation could add it to a customer relationship management system (CRM), where customer and sales records are kept, and alert a colleague. A person can check the draft before sending.
Where else it may help
Similar steps may help with invoices, meeting notes, customer updates or support enquiries. A simple rule may be enough when the task follows a predictable pattern.
Where to start
Look for repeated work that causes delays or mistakes. Agree what information can be used and where a person should check the result. AI can misunderstand or invent details, and an automated step can fail.
Start with a decision, not a tool
A rule can route a form when the customer selects an enquiry type from a fixed list. If the same request arrives as an email, AI may help suggest a category from the free text. The team can then check the suggestion before acting.
Neither approach needs to decide what happens to every request on its own. Leave unclear or incomplete enquiries for a person, and make sure the team can see when a connection has failed.
A first workflow to try
Start with an enquiry that already arrives in a shared inbox. A rule can create a follow-up task; AI might prepare a short summary for the colleague who picks it up. The colleague checks the original enquiry, corrects the summary and decides what to send.
Before testing, agree who can see the enquiry, what information the tools need and how the team will spot a missed task. Compare a few examples with the existing process. If checking takes longer than doing the work directly, change the design or stop the trial.
For a closer look at choosing the work, read how to find repetitive work to improve.