Since 2025
Support requests sorted, drafted and prioritised by AI
Requests arrive by email, by WhatsApp, by web form. They are qualified, attached to the right project, and a reply is drafted — a human approves before it is sent.
incoming requests qualified and attached to their file
accounts followed in the same stream, whatever channel they arrive on
email, WhatsApp and web forms, feeding a single inbox
The context
A small firm looking after several clients receives requests everywhere: multiple mailboxes, WhatsApp messages, website forms. The cost is not in the reply, it is in the sorting — remembering which client this is, finding the history, judging the urgency, and above all letting nothing drop. That is exactly the kind of task with measurable volume, a stable rule, and recoverable errors.
The workflow, in motion
One queue: tickets, emails, SMS, WhatsApp
Tickets, emails, SMS and WhatsApp land in a single queue: AI absorbs the request, prepares the answer, and hands the call back to a human.
Incoming request
Email, ticket or message — no format imposed on your teams.
Triage
Topic, priority and scope identified within seconds.
Context gathered
History, business data and the relevant code are pulled together on their own.
Answer prepared
Fix, tests and a summary of what changes, ready to review.
Human validation
One call, then it ships. No exceptions.
The human keeps the decision: nothing ships without validation.
What we did
- A single inbox for every channel: email, WhatsApp, forms. A request arriving anywhere arrives in the same place.
- Automatic routing to the right client from the sending address, phone number or keywords — with written rules, not guesswork.
- A reply drafted from the file's history, submitted for approval. Nothing goes out without a human having read it.
- Scheduled follow-ups on anything awaiting a client reply, so tracking does not depend on somebody's memory.
- Technical alerts from every project poured into the same stream: an outage caught by a probe lands where a client request lands.
What changed
- No more requests lost between two mailboxes: there is one entrance, and anything unanswered stays visible.
- The time spent sorting has been absorbed. The time spent replying stays human — that is the part with value in it.
- Human validation is systematic on anything going to a client. The automation drafts, it does not send.
- It is the tool we use ourselves every day. We do not sell automation we do not live with.
The services involved
This engagement drew on the following expertise. Each has its own page, with its method and its deliverables.
Other case studies
FlatBay
Since 2015Agencies drowning in incomplete files. An assistant that chases missing documents, computes the affordability ratio, and only passes on complete applications.
Read the case study
Clinalliance — réservation
Since 2016A centre opens its slots, clients book, pay and get their invoice. The service has been running since 2016 and has followed every version of our stack.
Read the case study
Référentiels de données
Since 2024Addresses, public transport, property sales, land registry: four services answering the same questions as paid APIs, built from open public data, running on our servers.
Read the case studyLet's talk about your project
Thirty minutes is enough to tell whether we are the right fit. We reply within 48 hours, and we say no when it is not for us.