Connecting a chat assistant to your CRM and ERP: one contract, many systems
How Kav is being built to answer deal, case and invoice questions from your CRM or ERP through common contracts. The contracts exist; vendor connectors are planned.
Articles
How to build an AI assistant people can trust.
Does Kav support WhatsApp? What's the status?
Yes — WhatsApp is live on Kav, alongside the website widget and Telegram. A new organization connects its own WhatsApp Business number; Meta approves that account.
Is the Kav widget accessible under Israeli Standard 5568 and WCAG?
The widget and Studio are built to WCAG 2.2 AA and IS 5568, checked by an automated gate in CI — here is what that gate covers and what it can't prove.
Which channels does Kav support — and which are live today?
Live today: the website widget, Telegram and WhatsApp. Messenger, Instagram, SMS and voice are not live yet. Here is every channel's status.
Which languages does the assistant answer in?
Kav's own assistant answers in Hebrew, Arabic, Russian and English. The platform also supports Amharic; each organization switches on the languages it chooses.
How Kav works: one engine behind every channel
From a message arriving to an answer on screen, the same fourteen checks run every time — never a guess the model made alone.
A real table, not a paragraph of numbers
When an answer is genuinely tabular, Kav renders a real HTML table a screen reader can navigate — never a wall of numbers in a sentence.
Fail closed: the one rule that runs through every layer
One rule runs through every layer of Kav: when something can't be evaluated cleanly, it denies — never guesses yes.
MCP names the problem. Here's how Kav solves it without a tool server to ship.
MCP names a real problem: how a model calls real systems. Here's the mechanism Kav actually runs to solve it — compiled fresh, never a tool server you deploy.
A capability has three faces, and only one changes per vendor
A capability has three separate faces, and only one of them is different for every organization that installs it.
When a list is too long for a card, the widget decides — not the model
Past four results, Kav's widget turns a card list into a swipeable rail on its own, with no auto-advance, so it can't fail the rule a carousel usually breaks.
Why not just a generic chatbot
Four guarantees a prompt bolted onto your website can't make — and why each one is a real constraint Kav is built around, not a feature to point at.
Your own AI tool, from your own API — configured, never coded
Point at your API, map what it returns, tell the AI when to use it, publish. The whole build is Studio configuration — nothing to deploy.
One queue for every channel: why a channel isn't a department
When every channel gets its own inbox, residents wait in the wrong one. Here's how Kav's agent desk queues conversations by department, not by the app they came from.