KIN · The AI front desk
Ask in plain language. Get the real answer.
KIN is the human-facing half of an AI back-office: a bespoke assistant your team reaches by voice, Telegram, or WhatsApp — the tools they already have open. It's wired into your actual systems, so “what's on for Thursday” and “quote this driveway” return your numbers, not a plausible invention.
- Reachable on
- VoiceTelegram + WhatsApp
- App to install
- NoneNothing to learn
- Knows
- Your businessLive records
- Setup
- Done for youEnd to end
- Guardrails
- ScopedPer person, per action
Why it exists
Frontier AI is useless to the person holding a hose.
The people who lose the most hours to admin are usually the least likely to open a dashboard. They are on a site, in a truck, or under a machine. Handing them a login and a settings panel doesn't help — it just moves the work.
So KIN meets them where they already are. A voice note on the drive home becomes a logged job note. A photo in a chat becomes a filed receipt. A question about Thursday gets Thursday's actual schedule. No prompt engineering, no configuration, no new habits.
What it does
An assistant that can actually do the thing.
-
01
Answers from your records
Schedule, customer history, quote status, outstanding invoices — read live out of the systems that run the business, not summarized from a stale export.
-
02
Starts the automations
The same pipelines that run on a schedule can be triggered by asking: file this, price that, mark the job complete, send the follow-up.
-
03
Takes photos and voice
A picture of a receipt or a two-sentence voice note is a complete input. Nothing needs to be typed into a form later.
-
04
Knows its limits
Money-moving and customer-facing actions are gated: KIN drafts, a human confirms. Every action it takes is written to the same audit trail as the rest of the system.
How we set it up
Configured for your business, then handed to your crew.
- Bespoke, not generic. KIN is configured against your operation: your rates, your job types, your vocabulary, your escalation rules.
- Per-person scope. Crew, office, and owner get different permissions — who can read what, who can trigger what.
- On the channels you use. Voice line, Telegram, or WhatsApp, whichever your team already answers.
- Backed by JARVIS. Answers come from the same context engine every other agent uses, so nothing contradicts.
- Yours at the end. The configuration and the source we write for you is handed over with the rest of the build.