After-Hours Enquiries: How to Set Up an AI-Assisted First Response That Feels Human

BOTrento Team2 November 20268 min read

Why one automatic reply is not enough

An enquiry that arrives late on a Friday evening has two needs. The sender wants to know it landed somewhere real. And later they want a reply that shows you read what they wrote: the event, the equipment, the dates, the problem.

A generic "Thank you for your message" covers the first need and fails the second. A fully automated, detailed answer sent without anyone looking can state the wrong price, promise availability you have not checked or misread an unusual request. For quotes and dates, that is a real risk.

So treat them as two layers with different rules:

• First layer, acknowledgement: short, factual, sent by your mailbox. No prices, no commitments.
• Second layer, the real reply: drafted by AI from the actual email text, reviewed and sent by a person.

First layer: the acknowledgement that does not sound like a robot

Set this up in the tool you already use. Gmail and Outlook both support automatic replies, and CRMs such as HubSpot or Pipedrive can send an acknowledgement from a form or sequence. Keep it plain.

What a good acknowledgement contains

• A named sender (a person or the team), not "noreply".
• Confirmation that the enquiry was received and who will read it.
• A realistic statement of when they will hear back, in terms you can keep, such as "on the next working day".
• A couple of things that would help you answer faster, phrased as an invitation, not a form.
• A way to reach you if it is genuinely urgent.

What to leave out

• Prices, availability or delivery dates.
• Phrases like "our AI will answer you". Say what happens, not which tool does it.
• Long lists of questions. Three at most; more feels like a barrier.

Template

Subject: We have your enquiry – [short reference to what they asked]

Hello [First name],

Thank you for your message about [event / equipment / service]. It reached us outside office hours, so I will read it properly on [next working day] and come back to you with a concrete answer.

To make that reply as useful as possible, it would help to know:

• the date or timeframe you have in mind,
• the rough scope or quantity,
• whether you have a budget range in mind (a rough idea is fine).

If you can reply with a line or two on any of these, great. If not, no problem, I will ask when we speak.

If this is urgent, call [number] on [day] from [time].

Best regards,
[Name], [Company]

Notice the budget question is optional and soft. Asking for specs, timeline and budget is sensible, but a demand reads as a gatekeeper. An invitation reads as a colleague.

Second layer: an AI draft ready for the morning

This is where AI earns its keep. Overnight, the assistant reads the enquiry text, works out what is being asked and what is missing, and prepares a draft reply. A person then reads it, corrects it and sends it.

A good draft does three things the acknowledgement could not:

1. Refers to specifics. It mentions the venue, the product or the problem the sender described.
2. Asks only what is still missing. If they already gave a date, it does not ask again.
3. Proposes a next step. A call, a site visit or a short quote, with a time frame.

The draft stays a draft. Pricing, availability and anything contractual are checked by someone who knows the business before it goes out. For the full logic of which tasks to give to an assistant and which to keep, see which sales tasks to hand over first.

Step-by-step set-up

1. Decide your response promise. "Next working day" is better than a number of hours you will miss at weekends.
2. Write the acknowledgement once. Use the template above and adapt the questions to your trade. An event producer asks about date and venue; an equipment rental firm asks about dates and quantity.
3. Switch it on in your mailbox or CRM. Limit it to new enquiries so existing customers do not get it for every reply.
4. Define what counts as urgent. A same-week event or a breakdown is different from a general price question. Write the rules down; see rules for categorising and prioritising enquiries.
5. Give the assistant your real material. Past replies, standard wording, price logic and your do-not-promise list.
6. Name the person who approves. Include a deputy for holidays and sick days.
7. Review the first batch closely. Edit every draft for the first week or two and note which corrections repeat. Those become rules.

If you want a pre-flight check, the readiness checklist for your inbox covers mailbox access, templates and approval rules.

How BOTrento handles this

BOTrento Chief of Staff is built for the second layer. It watches enquiries and quotes in your mailbox and in uploaded data, and prepares draft replies, quotes and follow-ups in four languages (Czech, English, German and French). On every working day morning it shows an overview called Today 1+3, in the app and by email, so the person on duty sees what arrived and what needs a decision first. Weekend enquiries therefore appear on Monday morning.

Some limits are worth stating plainly:

• BOTrento sends nothing by itself. Every email and quote draft is approved by a person.
• It is not an instant auto-responder. If you need an immediate acknowledgement, keep the first layer in Gmail, Outlook or your CRM.
• It does not confirm dates or bookings, and it does not read attachments; it works from the email text.
• It captures promises such as "I will get back to you by Friday" and reminds you of them.

AI is included in the plan, with a monthly usage limit. You get a warning at 80%, and at 100% new work pauses until the next month, with no automatic extra charge. Using your own API key is optional; today that means Anthropic (Claude). You can see the flow in the process demo, and the plans are on the pricing page.

You can also test it on your own enquiries. The free trial runs for 14 days, needs no card and no API key of your own, and starts when you first upload data.

On privacy: data is stored on servers in Germany (EU). The text needed for a task is processed by a commercial model provider through its API, so an EU server with an external AI call is not a closed environment. For a practical review, see the GDPR checklist for AI sales tools, and ask a lawyer about your specific case.

Checklist: does it feel human?

• Is the sender a real name or team, not "noreply"?
• Does the acknowledgement promise only what you can deliver?
• Does it ask three questions or fewer, and are they optional?
• Does the morning reply mention something specific from the original email?
• Has a person checked every price, date and commitment?
• Is there a route for urgent cases that does not wait for morning?
• Does the reply sound like your best salesperson on a good day?

For speeding up the human part afterwards, see how to reply to sales enquiries faster. If you work in event production or equipment rental, where evening and weekend requests are common, the pages for events and production and equipment rental show how this applies to your trade.