Picks up every call — during jobs, after hours, on weekends, and when the office line is already busy — using your configured business information and call flows.
Managed AI employees
Turn missed calls into booked jobs.
LScale builds and manages AI employees for service businesses — starting with an AI Receptionist built to handle calls 24/7, qualify leads, book jobs, and update your calendar and CRM.
Start free. In 3 minutes, you'll be talking to your AI.
-
Incoming call+1 (503) ••• 4182 · after hours
-
AnsweredAI Receptionist picked up on the second ring
-
Request understood“No heat — furnace is blowing cold air”
-
Lead qualifiedHomeowner · address in service area · not an emergency
-
Appointment selectedFirst open diagnostic window · Tue, 8–10 AM
-
Job bookedConfirmed with the caller before ending the call
-
Calendar and CRM updatedJob #4182 created · assigned to Tuesday route
-
Summary sentOwner and on-call tech notified
Illustrative call · sample data
The company
We build and run AI employees for service businesses.
Not software you configure. LScale designs the workflow, configures it, tests it against real scenarios, launches it, and then monitors, manages, maintains and improves it for as long as you use it.
The packaged AI Receptionist is configured for these trades today.
The operational problem
Every missed call can become a missed job.
You already pay to make the phone ring. The money is lost after that — while crews are on jobs, driving between calls, or standing in a customer's kitchen, and every evening and weekend after the office closes.
A homeowner with no heat finds you and dials. This is the lead you already paid for.
InboundThe office closed hours ago. The on-call tech is on another job and can't pick up.
UnansweredMost callers with an urgent problem don't leave a message. They keep dialing.
No messageThe job is booked — by someone else. You never learn the call happened.
LostYour two numbers
Everything below is derived from these two, with each rate shown where it comes from.
A range, not a promise. The three conversion rates are published figures from Invoca's analysis of 70M+ inbound calls. The 50–75% recovery band is our own assumption — we have not measured it, so we show it as a band and label it.
Stop the leak- 500 calls a month your number
- 240 never reach a person 48% · Invoca, 70M+ calls
- 91 of those are real leads 38% · Invoca, 70M+ calls
- 41 would book on the call 45% · Invoca, 70M+ calls
- 246 – 369 jobs a year we could answer 50–75% of 492 · LScale assumption
Move any rate and its source line changes to your estimate — we will not put your number under someone else's name. The 50–75% recovery band stays fixed: it is a band precisely because we have not measured it.
What the AI Receptionist does
Four jobs, done the same way every time.
Collects what your team actually needs before anything is scheduled: the problem, the property, the address, urgency, and whether it falls inside your service area.
Checks available times and books qualified appointments into your configured calendar workflow while the customer is still on the line.
Creates or updates the relevant records and sends your team a concise summary of what happened and what to do next.
The part most AI gets wrong
When it shouldn't answer, it doesn't guess.
The AI Receptionist is governed by the rules configured for your business. When a call falls outside them — an unusual request, a question with no approved answer, a safety situation — it stops, says so plainly, collects what a person will need, and follows your escalation path.
-
“Water is coming out from under the heater”
-
Outside configured booking rulesActive leak · shutoff unconfirmed
-
Declines to book a windowNo invented answer, no guessed availability
-
Routed to the on-call team, flagged urgentAddress and callback number captured first
Where it writes
The booked job lands where your crew will see it.
Standard connections are part of the managed setup. Unusually complex implementations fall under Custom / Advanced Setup.
The managed process
We build it, test it, launch it, and keep running it.
You are not handed a workspace and a setup guide. LScale does the operational work — and stays responsible for it after launch.
-
We learn your business
Services, service area, booking rules, hours, escalation paths, languages, and the questions your customers actually ask.
-
We configure your AI Receptionist
Call flows, qualification logic, appointment booking, calendar and CRM connections, language configuration, summaries, and routing rules.
-
We test it before launch
Common calls, difficult questions, booking scenarios, urgent requests, and escalation behavior — run and reviewed before a single customer reaches it.
-
Go live in 3–5 business days
Your number is connected and the AI Receptionist starts taking real calls.
Standard setups · custom or complex builds take longer -
We manage and improve it
We review real interactions, tune responses and workflows, and stay accountable for how it performs.
- Services
- Repair · Maintenance · Install · Emergency no-heat
- Area
- Portland metro · 5 ZIP groups · 35-minute drive limit
- Hours
- Office 7 AM–5 PM · after-hours on-call rotation
- Booking
- 2-hour arrival windows · 1 diagnostic slot held daily
- Escalate
- Gas smell, no heat under 40°F, commercial accounts
- Languages
- English · Spanish
Thanks for calling — are you calling about a repair, or scheduling maintenance?
My furnace is blowing cold air.
Intent → no-heat · ask temperature + occupancy before offering a slot
I can get a technician out. Is anyone in the home who needs heat urgently tonight?
Every question, threshold, and fallback is written down and reviewed with you before it goes anywhere near a customer.
- Standard repair booking — books the correct window
- Caller outside the service area — declines and captures the lead
- Pricing question with no configured answer — does not guess
- Caller switches to Spanish mid-call — continues in Spanish
- Gas smell reported — stops booking, follows the safety escalation path
- Existing customer rescheduling — finds the job, moves the window
We run these against your configuration and fix what fails. You review the results before launch day.
- Discovery call and business profile
- Configuration and integration setup
- Test cycle and your review
- Number connected · live on real calls
This timeline covers standard setups. Custom workflows, unusual integrations, or multi-location businesses take longer, and we tell you the real timeline before you commit.
- Reviewed
- Real call recordings and transcripts, on a set cadence
- Tuned
- Questions, thresholds, and routing adjusted from what we hear
- Updated
- New services, price changes, seasonal hours, staff changes
- Reported
- What came in, what booked, what needed a person
- Owner
- LScale — not your office manager
Why this isn't a tool
Not another AI tool your team has to manage.
There are three realistic ways to cover the phone. Two of them leave the operational work with you.
- You Design the call workflow
- You Write and tune prompts and rules
- You Connect the calendar and CRM
- You Test the edge cases
- You Monitor conversations
- You Catch and correct failures
- You Keep business information current
Works — if someone on your team has the hours and the appetite to run it every week.
- Them Answers and writes down the request
- Them Sends the message to your team
- You Call the customer back
- You Ask the qualifying questions
- You Check the schedule
- You Book the job
- You Update the record
The call gets answered. The job still waits on someone in your office picking up the thread.
- LScale Designs the call workflow
- LScale Writes and tunes the rules
- LScale Connects the calendar and CRM
- LScale Tests the edge cases
- LScale Reviews real conversations
- LScale Corrects what goes wrong
- AI Books the job on the call
LScale owns the setup, testing, launch, tuning, and ongoing management. You review what came in and get back to running the business.
Hear it
Listen to it handle a real conversation.
What you see
One screen that answers: what came in, and what needs me?
Not an analytics product. A short list of what happened on your phone line, with the two or three things that actually need a person marked clearly.
Interface concept. Sample data shown for illustration.
Pricing
One monthly fee. Setup, management, and tuning included.
No per-seat licensing for software your team then has to operate. You are buying a managed AI employee and the people who keep it working.
Custom workflows and additional AI Employees are scoped by consultation.
Questions
The things owners ask first.
It answers inbound calls, identifies what the caller needs, asks the qualification questions you've defined, checks whether the job fits your service area and rules, offers real appointment windows, books the job, updates your calendar and CRM, and sends your team a short summary. Anything outside those configured rules gets escalated rather than improvised.
One credit covers up to three minutes of an AI Receptionist call. Start Small includes 50 credits and additional credits are $1.75 each. The Managed AI Receptionist plan includes 200 credits, with additional credits at $1.50. Most routine bookings finish inside a single credit, and each call shows the credits it used alongside its summary.
When your included credits run out, nothing is cut off: calls keep being answered at your plan's additional-credit rate, billed automatically to your card. There is no hard cap — your phone line never goes dark mid-month — and per-call usage stays visible so the bill is never a surprise.
Standard setups go live in 3–5 business days: discovery, configuration, a test cycle you review, then connection to your number. Custom workflows, unusual integrations, or multi-location businesses take longer — we give you the real timeline before you commit rather than after.
Yes. It checks availability in your configured calendar workflow and books qualified jobs into it while the caller is still on the line — using your arrival windows, buffer rules, and any slots you hold back for diagnostics or emergencies.
Standard connections to systems like Housecall Pro, Jobber, ServiceTitan, Google Calendar and Calendly are part of the managed setup, along with CRM record updates and webhook or API connections. Unusually complex implementations — deep multi-location routing, non-standard data models, or systems with no API — are scoped under Custom / Advanced Setup. We confirm exactly what your business uses during onboarding.
It follows the escalation path you defined during setup. When a call falls outside its configured rules, it recognizes that, tells the caller plainly instead of guessing, collects the details a person will need, and routes or summarizes the request to the right team member. Safety situations and commercial accounts are common escalation triggers, and you decide what else belongs on that list.
Calls are recorded, transcribed and summarized — that is how the AI books the job, updates your records, and hands off cleanly when a person takes over. A recording and consent disclosure can be configured into the opening of every call flow, which matters because several US states require all parties to consent. We agree the exact disclosure wording with you during onboarding.
Yes. Language handling is part of configuration — including which languages to support and whether it should switch automatically when a caller starts speaking another one. The call summary tells you which language was used.
No. A phone bot answers and reads a script. An AI employee completes the operational workflow — qualifying the lead, booking the job, updating your records, and telling your team what happened. And it's managed: LScale designs it, tests it, launches it, and keeps tuning it based on real calls. Voice is where we start, not the whole model.
Yes, by consultation. The packaged AI Receptionist covers the common home service call patterns. Beyond that — unusual booking logic, additional systems, or AI employees working in chat, email, or your CRM — we scope it with you first and tell you honestly whether it's a good fit.
The part that matters to your business is the managed layer around it. LScale handles setup, configuration, testing, launch, tuning, reporting, and escalation rules — and stays accountable when something needs to change. We select and maintain the underlying technology so you don't evaluate, integrate, or babysit it, and we update it as better options appear without you having to rebuild anything.
Get started
Let your business answer while you are busy running it.
Your next customer is going to call at the worst possible moment. Someone will answer — the only question is whether it's you, or your AI employee.
Start free. In 3 minutes, you'll be talking to your AI.