No. The system prompt enumerates declined categories explicitly: sentencing-guideline analysis, plea exposure, statute interpretation, case-strength assessments. The bot consistently refers the visitor to a qualified attorney and offers to book a consultation. Run quarterly audit tests with adversarial phrasing ("hypothetically what's the range", "what would you do in my position") to confirm the refusal behaviour holds across rewordings. Bar associations treat unauthorised practice of law strictly in white-collar contexts, so the guardrail is non-negotiable.

It captures parties, opposing parties, related entities, parent companies, and known counsel into a structured intake form. The conflicts team reviews before any consultation is confirmed. The bot itself does not make conflict decisions, that's a professional-responsibility judgment for humans. For corporate matters the bot also asks whether the inquiry is on behalf of the company, an officer, or an outside director, because those routes diverge quickly and need different conflicts treatment.

Pre-engagement, conversations are not yet attorney-client privileged. Communicate this clearly in the widget header or initial message so visitors understand the posture. Privilege begins at formal engagement and depends on jurisdictional rules. Storing logs in your WordPress database keeps control with the firm, which makes it easier to align retention with your engagement-letter terms and the records-retention policy applicable to white-collar matters in your jurisdiction.

Yes. Multibot lets securities-enforcement, tax-controversy, healthcare-fraud, FCPA, and general fraud each have a focused assistant. A securities bot can ask about Wells notices and SEC stage; a healthcare-fraud bot can ask about False Claims Act posture and government payer mix. Scoping the bot per practice line keeps intake quality and routing accuracy higher than a single firm-wide assistant. Display conditions in SleekAI gate each bot to the right URL patterns.

For firms with cross-border practices, configure the bot to ask jurisdiction up front and route DOJ, FCPA, and SFO matters to the right partners. The bot replies in the visitor's language using GPT-4o or Claude. Legal-advice refusal language remains as configured, and translation does not soften the guardrails. For matters that involve coordinated US and foreign proceedings, the bot can flag the multi-jurisdiction posture so the case team prepares for parallel-investigation issues at the first meeting.

BYO API key (OpenAI, Anthropic, Google, OpenRouter) under the standard API terms means data is not used for model training by default. Conversation logs live in your WordPress database and retention is your choice. For white-collar firms, common posture is short retention with structured handoff to the case-management system once an engagement starts. Document the bot in your privacy notice and align retention with engagement-letter terms. The plugin gives you the building blocks; the policy posture is yours.

Configure the bot to recognise cooperator and whistleblower signals (qui tam language, SEC whistleblower program references, prior contact with prosecutors) and route those carefully. Most firms route them to a senior partner directly rather than through routine intake because the conflicts and strategic posture differ from a defense-only inquiry. The bot should not turn the prospect away brusquely; it should gather enough context for the senior partner to make the call about consulting.

Configure the bot to recognise time-sensitive language (target letter received, subpoena returnable this week, grand-jury appearance scheduled, voluntary interview requested) and prioritise same-day or next-day slots. Speed matters more than the standard intake flow. The system prompt should escalate the routing posture explicitly and shorten the intake to the minimum needed for conflicts clearance, so the partner can pick up the matter on a compressed timeline without losing the lead to another firm.