Messaging assistant

An immediate answer for your customers, even when you are busy.

I set up an assistant that answers frequent questions on your website or your messaging channel, using information you have approved. When it does not know, it passes the request to the right person, with its context.

Does this sound familiar?

  1. The same questions every day

    Opening hours, prices, lead times, service area: you answer the same messages again and again.

  2. Enquiries arrive out of hours

    A customer writes in the evening or at the weekend. By the time you reply, they have contacted someone else.

  3. Messages get lost

    Between the website, messaging and the phone, nobody knows who has answered what.

What happens if nothing changes

  • An enquiry without a quick reply often goes to a competitor.
  • Your team spends part of the day on simple questions.
  • Important enquiries get buried among the others.

These are common situations, not a measurement of your business.

How I set it up

  1. 01

    Gather the answers

    The frequent questions and the information the assistant can use, approved by you.

  2. 02

    Configure and test

    The answer rules, the limits and the handover to your team. I test with real questions, including those it must not answer.

  3. 03

    Monitor and adjust

    You see the conversations and the requests passed on. The answers improve with your corrections.

What you receive

  • An assistant set up on your approved information, on the agreed channel.
  • A handover to a person for requests it must not handle.
  • Documented tests, with the planned questions and the expected answers.
  • A clear notice for your customers: they know they are talking to an AI assistant.

Where it can answer

  • On your website, in a chat window
  • On a business messaging app, if its provider allows it
  • By email, with a suggested reply to review

The assistant on this website, “Any questions?”, works on this principle: it answers questions about the services and hands over to a person.

Limits

  • The assistant only answers from approved information; it can make mistakes, hence the handover to a person.
  • No sensitive data should be entrusted to it: the rules are written before it goes live.
  • Each messaging app has its own access conditions; I check what it allows before committing.

What I commit to on every project.

These rules apply from the first call to the handover.

  1. The price before you decide

    Scope, deliverables and price are written in the quote before any work starts.

  2. Accounts in your name

    Tools, access and subscriptions stay yours. You can take back control at any time.

  3. A person approves

    Important actions, such as sending or publishing, are approved by someone on your team.

  4. No invented figures

    Fictional examples are labelled as such. A client project is only shown with the client’s consent.

  5. I tell you when not to build

    If a simpler tool is enough, or if nothing should be built, I tell you.

Frequently asked questions

Will my customers know it is an AI?

Yes. The assistant introduces itself as such, and a person can always take over the conversation.

What happens if it does not know the answer?

It says so, and passes the request to the right person with the context of the conversation.

How much does it cost?

The price depends on the channel, the number of questions and the tools to connect. You receive a written quote before deciding.

Let’s talk about your messages.

Tell me which questions come up most often and through which channel they arrive.

Free 15-minute first conversation, remotely or in Lyon. Every project is quoted individually.