The address and the machine at it
Properties and equipment live on the customer, not in a message. A returning customer never re-types their address, and the technician sees what was found there last time before leaving the yard.
An AI front office for service businesses: it qualifies every enquiry, offers only real slots, and turns each confirmed visit into a work order.

Intake, the customer record, the service template, the action log, delivery and the funnel — the things a service business usually keeps in a notebook, a spreadsheet and somebody's memory, kept once here where the next request can use them.
Website chat, email, WhatsApp, Telegram and Instagram land in the same queue. A customer who writes on one and sends a photo on another does not become two jobs.
Properties and equipment live on the customer, not in a message. A returning customer never re-types their address, and the technician sees what was found there last time before leaving the yard.
Each service carries its duration, required skills, buffers and the questions that must be answered first. The AI never decides that a request is complete — your template does.
Every write is an action run: the tool, the input, the actor, the result — including the ones the policy layer refused. It is the record you want when a customer asks why they were told that.
Confirmations are durable runs with attempts and a visible state, not a best-effort send. A booking the customer was never told about is a no-show, and this is the layer that makes it a bug you can see.
Conversations in, requests qualified, visits confirmed, work orders out. Not first response time. The report answers whether the enquiries turned into work.
This is the loop the front office runs on every enquiry, and your team already does all four steps. What changes is who does the typing, the checking and the chasing — and that it stops short of anything that should be your call.
Ven turns a vague message into a structured request: address, equipment, fault, access. Every answer carries where it came from and how sure the model was — so you can always tell what the customer said apart from what the AI inferred.
Ven proposes, your operator accepts or rejects, field by field. Accepted changes apply atomically. Nothing is sent, promised or booked on a proposal alone.
Availability comes from real working hours, skills, service zones and travel buffers — never from a model guessing. The slot is held with an expiry.
Confirming creates the work order in the same transaction, with a brief frozen at that moment: service, address, access, equipment, answers. Editing the request later never rewrites the instructions somebody already drove out with.
This is not a chatbot with a calendar plugin. The model proposes a structured decision; the application checks state, permissions and policy, and only then acts — which is why a repeated confirmation returns the same booking instead of a second van.
Every workspace starts in Observe, where Ven writes down what it would have done and changes nothing. You climb from there, one mode and one service at a time. The last three cards are on the roadmap and say so.
Observe records decisions without touching the request. Assist proposes for review. Qualify lets Ven collect the answers on its own. Book stays switched off until the scheduling tools are in Ven's hands, and the console will not let you pretend otherwise.
Working hours with a timezone, the service zones you actually drive to, and what happens to an emergency at 23:40. The server enforces all three before a slot is ever offered.
Coming soon
Two-way sync with Google and Microsoft, so external busy time blocks a slot and a confirmed visit lands in the calendar they already look at. The native week is live today; the connectors are next.
Mon
Tue
Wed
Thu
Fri
Booked visitsExternal busy
Coming soon
Ven drafts the quote from the service template and what it learned; a named person approves it before it goes anywhere. Nothing financially binding is ever sent autonomously.
Needs A. Weber before it can be sent
Coming soon
Every confirmed job already publishes a domain event — the seam the field-service and accounting connectors plug into. The event is live; the connectors are being built on it.
The event is live; connectors read it
Every plan starts with a shared inbox and live chat. Add Ven AI's autonomous replies as your volume grows.
$27/month+ tax
For a service business that takes every enquiry and runs every visit itself.
$71/month+ tax
When Ven should do the qualifying instead of your dispatcher.
$224/month+ tax
For multi-crew operations running several sites and channels.
Custom pricing
Custom capacity, security review, integrations, onboarding and SLA.
Lavenity is an AI front office for service businesses. Every enquiry — website chat, email, WhatsApp, Telegram, Instagram — lands in one inbox and becomes a structured service request. Ven AI asks the qualifying questions, your team approves what matters, and a confirmed visit becomes a work order with a brief the technician can leave with.
A helpdesk optimises for closing conversations; a front office optimises for turning them into work. Lavenity keeps the shared inbox and grounded AI replies you would expect from a helpdesk, and adds the operational layer underneath: service requests, qualification templates, availability, holds, confirmed visits and work orders.
Not today. Ven qualifies the request and proposes the next step; a person confirms the slot. Autonomous booking exists as a mode and is deliberately switched off until the scheduling tools are in Ven's hands — we would rather tell you that here than have you find out in the console.
Three ways, all on by default. Ven starts in Observe and changes nothing. Answers stay grounded in the knowledge you connect. And every write passes through a policy layer that checks state, permissions and version first — the model proposes, the application decides.
No. A slot is held with an expiry and confirmed against a database constraint rather than a check in application code. Repeated and concurrent confirmations return the same appointment and the same work order, which is proven by an integration suite that runs them in parallel against Postgres.
Website chat, email, WhatsApp, Telegram, Instagram, Facebook Messenger, Discord, Slack and SMS all arrive in the same queue, and your team replies in whichever channel the customer already used.
One service with its duration and required questions, your working hours and service zones, and one resource with availability. Most teams are live in an afternoon, and start in Observe so nothing is at risk while you read what Ven would have done.
Not yet. The native calendar, availability, confirmations and work orders are live today; Google and Microsoft sync, quotes with approval, and field-service connectors are on the roadmap. Every confirmed job already publishes an event, which is the seam those connectors will use.
Service businesses where an enquiry ends with somebody driving out: heating, plumbing and HVAC, electrical, appliance repair, cleaning and maintenance contractors. It was built around German SHK workflows first, and works anywhere the loop is the same.
Lavenity's privacy practices are built in line with the GDPR, and the front office treats addresses, photos, serial numbers and message content as personal and operational data. Data in transit is protected with TLS, access is role-based, the chat widget runs in an isolated iframe, and every write is auditable. Enterprise plans add custom data retention and a dedicated security review.
Yes, every plan starts with a 14-day free trial. No credit card is required, and you can stay in Observe mode for as long as you want to watch before you switch anything on.
A 14-day free trial and an afternoon of setup. Start in Observe — Ven changes nothing until you say so.