No. The system prompt declines settlement-value estimates, fault percentages, and comparative-negligence analysis. Those depend on medical records, liability investigation, and jurisdiction-specific rules that the bot does not have. The bot consistently refers the visitor to a free contingency consultation. Audit the refusal quarterly with adversarial prompts ("hypothetically what's the range", "a friend's case was worth") to confirm the guardrail holds. Settlement-value claims are the most common compliance complaint against PI chatbots, so the prompt is strict.

Configure multibot to run a separate intake per state if your firm operates across jurisdictions. No-fault states need PIP-limit and threshold-injury questions; at-fault states need different liability and insurance-limits posture. The bot reads which state the crash happened in and routes accordingly. For firms covering only one state the configuration is simpler, but the intake fields still reflect that state's rules so the case-screening attorney walks into the call with the right context.

It declines to advise on whether to give a recorded statement, accept a settlement offer, or sign a medical-records release for the adjuster. Those are attorney calls. The bot tells the visitor that many people prefer to wait until they have spoken to an attorney, and offers to book a consultation as soon as possible, often same-day. The system prompt lists adjuster-communication strategy as a declined category explicitly so the refusal language is consistent.

The bot captures the structural information your case-screening team uses, injury severity, treatment status, time since crash (statute of limitations posture), insurance limits visible. It does not make the viability decision; that is for the screening attorney. The intake form lets the attorney look at the structured fields before the consultation and route low-viability matters to a quick decline letter rather than a 30-minute call. The bot makes the funnel faster without making the legal judgment.

As fast as your scheduler allows. Most PI firms expose same-day slots through a calendar tool and the bot books directly into the next open slot, often within an hour of the chat. For high-priority matters, recent ER visit, time-sensitive insurance contact, the bot escalates the booking posture and surfaces the soonest available slot. Confirmation by text follows automatically so the visitor does not drop off between the chat and the call. Speed is the entire competitive game in PI intake.

Yes. Many car-accident practices serve significant Spanish-speaking communities and SleekAI replies in the language the visitor writes in. The booked consultation is flagged with preferred language so the attorney prepares accordingly. Legal-advice refusal language remains intact across translation because the prompt rules apply in any language. For markets with Mandarin, Vietnamese, Tagalog, or Korean speakers, the same multilingual capability extends naturally without changing the guardrails or routing logic.

BYO API key (OpenAI, Anthropic, Google, OpenRouter) with the standard terms means data is not used for training by default. Conversation logs live in your WordPress database under your control. For PI firms the typical posture is medium-length 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 provides the building blocks; the policy posture is the firm's.

Configure the bot to recognise statute-of-limitations urgency, crashes approaching the two-year mark in most states, or longer-tail discovery issues, and prioritise immediate booking. The system prompt should not give a specific statute-of-limitations answer (that's legal advice) but should escalate the routing posture when the visitor indicates the crash was over a year ago. The case-screening attorney handles the actual statute analysis on the consultation, on a compressed timeline if needed.