What does Blawby's chat actually do?

It runs a mode-aware conversation: REQUEST_CONSULTATION drives a structured intake against your template, ASK_QUESTION handles light questions, and CONVERSATION drives post-acceptance two-way chat. The AI routes incoming messages to the right mode automatically.

Is the AI giving legal advice?

No. The AI is built to surface intake context and route to a real attorney; it does not provide jurisdiction-specific legal advice. Your firm's disclaimer and acknowledgment language are configurable per intake template.

How many languages does it support?

18 languages with full RTL for Arabic. The widget detects browser language; clients can switch.

AI Chat Client Acquisition for Law Firms

Most firm websites are static brochures. The lead either fills a contact form (low conversion, slow response) or doesn't bother. Blawby's chat turns that surface into a 24/7 intake worker — qualifying clients in their language, collecting the structured facts your team needs to triage, and gating on a consultation fee when your policy requires.

This article is the strategy version. For the operational setup, see Set Up Client Intake and Intake Templates.

Why structured AI beats a contact form

A contact form is one event: name, email, message, send. The lead either filled it correctly or didn't, and you find out hours later when someone checks the inbox. Several things go wrong by default:

Drop-off — long forms feel like work; conversion drops with each required field.
Sparse data — even when filled, you usually get a free-text "I have a question" with no urgency, jurisdiction, or matter type.
Slow first response — most leads expect a reply in minutes, not the next morning.
No language coverage — if your client base isn't English-only, the form excludes them.

Blawby chat solves all four with structure: the AI asks one thing at a time, validates as it goes, conducts the conversation in the client's language, and submits to a notification-routed queue.

How the chat actually works

The chat surface is mode-aware. The AI doesn't just have one personality — it routes the first message into one of four modes and behaves accordingly:

Mode	What it does	When it fires
`REQUEST_CONSULTATION`	Conducts a structured intake against your template. Required → enrichment → fee → submit.	First message that signals a real legal need.
`ASK_QUESTION`	Lightweight one-off response, no funnel.	Casual or informational queries (e.g. "what areas do you handle?").
`PRACTICE_ONBOARDING`	The conversational setup flow when you first sign up.	Internal — only at firm onboarding.
`CONVERSATION`	Two-way attorney-led thread, AI assists.	After the firm accepts an intake.

You don't pick the mode manually. The platform classifies the first message and routes it.

This matters for client acquisition because most generic chatbots run one mode — every visitor gets the same script regardless of intent. Blawby's classifier separates "I'm curious about your firm" (don't run a 10-question intake) from "I was in an accident yesterday" (run the intake right now).

Diagram placeholder

Diagram of the intent classifier — first message routes to REQUEST_CONSULTATION, ASK_QUESTION, or CONVERSATION based on signal.

/media/docs/ai-chat/intent-routing.svg

Reach: 18 languages, full RTL

The AI runs intake in 18 languages out of the box:

Americas: English, Spanish, Portuguese, French
Europe: English, Spanish, French, German, Russian, Italian, Dutch, Polish, Ukrainian
Asia: Chinese, Japanese, Vietnamese, Korean, Thai, Indonesian, Hindi
Middle East / Africa: Arabic (full right-to-left rendering), French, English

The widget detects browser language. Clients can switch. RTL languages render fully right-to-left across all surfaces — including the engagement letter the firm later sends.

For firms serving immigrant communities, multilingual practice areas, or international cases, this is the difference between capturing a lead and losing one.

Trust building: the AI's tone

The intake AI is built around two principles:

Help before asking. The opening turn provides value (clarifying what's covered, what to expect) before requesting personal info.
No legal advice. The AI is intake-and-route, not advise-and-resolve. Disclaimers are configurable per template; the AI honors them.

A typical first turn for a REQUEST_CONSULTATION intake:

Hi — I can help you understand whether our firm can help with your situation. Tell me what happened and I'll ask a few follow-ups.

That's it. No 12-field form on screen. The AI then collects the required fields one at a time, with conditional follow-ups, and submits to your queue.

Conversion lever: consultation fees in chat

For firms with a paid first consultation, the most common drop-off is "the website didn't tell me how to pay." Blawby's intake template can require payment before submission — the Stripe payment card renders inline in chat, between the contact-collection step and submission confirmation. The client pays without leaving the conversation.

This isn't optional friction; it's a qualifier. Clients who pay for a consult are more committed than those who fill a form, and the firm's first response is to a paying lead, not a free one.

After-hours: the queue keeps running

The chat runs 24/7. Submissions stack in your queue overnight; notifications fire to your team via in-app, email, and (optionally) push (OneSignal).

What this lets you optimize:

First response time — the most-correlated metric for legal-lead conversion. Push notifications mean a same-hour response is realistic even after hours.
Coverage gaps — define a backup intake owner so vacations and holidays don't sink the queue.
Triage policy — paralegals can do first-pass triage; attorneys review only the ones that pass.

Integration: where chat fits in the broader funnel

The chat doesn't operate alone. The end-to-end conversion path on Blawby:

Widget on your website → REQUEST_CONSULTATION mode runs your template.
Submission lands in the Intakes queue with AI-generated summary and urgency.
Triage — accept (creates matter), decline, or "ask AI for more info."
Conversation — CONVERSATION mode after acceptance; attorney-led with AI assist.
Engagement — structured letter sent from the matter, signed in-thread.
Payment — retainer or initial fee collected after acceptance.
Active matter — time, files, invoicing all from the matter detail.

Each step has its own page in the docs. The point: chat is the entry, not the entire workflow.

Practical performance metrics

Track these monthly:

Visit-to-conversation rate — % of widget visitors who send a first message.
Conversation-to-submission rate — % of started conversations that complete intake.
Submission-to-accept rate — your team's triage decisions.
Accept-to-engagement rate — how many accepts actually sign.
Engagement-to-paid rate — payment completion.
Median time to first response — your firm's SLA on triage.

If conversation-to-submission is low, your template likely has too many required fields. If submission-to-accept is low, the AI isn't qualifying tightly enough — adjust template or services list.

Common pitfalls

Over-automating the conversation. Once you've accepted, switch to attorney-led mode. Clients can tell when they're still talking to a bot post-acceptance, and trust drops.
Skipping the disclaimer. Configure your template's pre-collection disclaimer once; don't omit it because "the AI seems helpful enough."
Treating the widget as set-and-forget. Review submissions weekly for the first month; refine template fields where clients consistently leave answers blank or off-target.
No multilingual fallback. If your client base includes Spanish or other non-English speakers, leave the language switcher visible — it's enabled by default.

Next steps

Configure your widget: Install the Chat Widget.
Build your template: Intake Templates.
Tune your queue: Review and Triage Client Intakes.

Ready to put your firm on a 24/7 intake surface? Start your setup.