Answers from approved information
It answers from your help pages and policies, plus a short document of answers for questions they do not cover. When something is not in them, it says so instead of guessing.
A customer service chatbot on your website answers repeated questions from information you approve. When it does not know, it says so and passes the conversation to your team.
Book a free consultation
In a shop it is "where is my order" and "how do returns work". For a property manager it is "who fixes the heating" and "when is rent due". Your team answers them by pasting the same replies, while a customer with a real problem waits further down the queue. Out of hours, nobody answers at all.
Answers repeated questions and passes the rest to a person.
It answers from your help pages and policies, plus a short document of answers for questions they do not cover. When something is not in them, it says so instead of guessing.
Where your software's API allows it, it looks up the status of an order or a repair request. It checks a reference number and an email address before it shares anything.
Complaints, refunds, damaged items and anything it cannot answer go to a person by email, or into your helpdesk where its API allows, with what was said so far. The customer is told what happens next.
You can read every conversation. Questions it could not answer are marked, so you can add the missing answer where it belongs.
Pick the questions your team answers most and the pages they copy the answers from. Tell us which cases must always reach a person, and which helpdesk you use, if any.
In this order, and nothing goes live before you have tried it.
We go through recent support emails and phone notes and list the questions that come up again and again, in the words customers use.
Some answers will be missing from your help pages or out of date. We point out the gaps and you approve the new wording. The chatbot answers from that.
We agree what always goes to a person, such as complaints and refunds, and how it reaches them: by email, or into your helpdesk where its API allows.
We test it with your real questions and some awkward ones before launch. Afterwards you read the conversations and fix answers at their source. Support after launch is agreed at handover.
What it looks like in a few kinds of business.
Order, delivery and returns questions, answered from your store data.
Guest questions before arrival, with booking enquiries sent to reception.
Repair requests logged and emergencies sent to the on-call line.
Turns trip ideas into briefs your consultants can quote from.
It says so and offers a person. It does not guess or invent a policy. You can read those conversations and add the missing answer.
It explains your refund policy as written. The refund itself, and every complaint, goes to a person with the conversation attached.
Yes. It introduces itself as an automated assistant, and a customer can ask for a person at any point.
Where your helpdesk's API allows it, handovers arrive as tickets with the conversation attached. Otherwise they go to your support inbox.
Yes. It replies in the language the customer writes in, from the same approved information. We test the languages that matter to you before launch.
Add the answers your team gives today, and we can show you how the chatbot would handle them. The consultation is free, over Google Meet, a voice call or text.