Use Lumi conversations and first-party journey data to see what visitors wanted before they left

Begin with what the visitor actually said

The most reliable way to find out what a visitor was looking for is to read the question they asked. Lumi, the AI Concierge for your website provides chat and voice assistance using the business’s approved information. Its workspace receives conversations, messages, leads, voice-use data and AI analysis. When a visitor asks about a service, policy, product or next step, the conversation provides direct evidence of that need rather than requiring your team to infer intent from a page view alone.

This distinction matters. Traditional traffic counts can tell you that somebody arrived and viewed a page. A conversation tells you what remained unclear after they read it. That is why an AI Concierge is useful in the broader work of answering website visitors after hours: it responds while the visitor is present and leaves the business with a record of the questions that arose when its team was unavailable.

Connect each question with the visitor’s journey

Lumi’s working first-party tracking records an anonymous visitor and session, the page path, UTM data, referring origin and language. The website or widget also supplies page views and other events. Together, these details help your team understand where the conversation started, which part of the site framed the visitor’s expectations and what source brought the person to the site.

Language is especially useful when reviewing demand from audiences who do not all communicate in the same way. For a related decision framework, see evaluating multilingual website assistance for seasonal tourism visitors. Lumi records the session language, while the conversation itself shows the words and questions the visitor used.

Separate known intent from an unexplained departure

Lumi can show what a participating visitor asked and what tracked actions occurred. It cannot tell you the private reason every person left. A visitor who never starts a conversation leaves page and session evidence, but no stated question. Even when somebody does talk to Lumi, ending the session does not reveal whether the answer was satisfactory, whether the visitor became distracted or whether they chose another course of action.

Ambiguous wording is one reason conversation records matter. A visitor asking to “cancel” could mean ending a membership or withdrawing from a class; the next useful question is how to disambiguate that request. The example in whether cancel means ending membership or withdrawing from class illustrates the type of uncertainty that raw traffic reporting cannot resolve on its own.

The same principle applies when a request depends on eligibility or policy. Who can request a beauty gift voucher transfer? is not answered by knowing which page somebody visited. It requires an answer grounded in the relevant business information, followed by a review of the actual question asked.

Use follow-up actions to understand what happened next

A conversation becomes more informative when it is read alongside the next recorded action. Customer Lumi can capture leads and hand visitors a booking link. That link opens the business’s scheduling URL with attribution and records a booking_started event. The link can point to the business’s selected scheduling provider, but Lumi does not know whether the appointment was completed unless a calendar source is connected for that business.

For commerce, a business with a connected SHOPLINE store can give Customer Lumi access to synchronized catalog information. Lumi can help shoppers find products and manage the real store cart. SHOPLINE supplies commerce and cart events plus checkout references, but not orders. Checkout reference tagging remains limited and should not be used as proof of a completed order or attributed revenue.

This action-based review is equally important when visitors ask about a service that may not be documented. Undocumented furniture assembly questions after hours show why the responsible response is to work from available business knowledge rather than inventing a service commitment. The resulting conversation still tells the business what information the visitor wanted and what its published material failed to answer.

Turn repeated questions into useful business intelligence

The value of visitor-intent data is not merely that a transcript exists. The business can review conversations, leads and journey context together in Lumi’s private workspace. Repeated questions point to subjects visitors care about, while the associated page paths show where those questions surfaced. This gives the team a concrete basis for reviewing website content, approved knowledge and the concierge’s answers.

Look for questions whose answers depend on distinctions the website must communicate clearly. Examples include comparing advisers’ published service scope and eligibility and published window substitution criteria before contacting installers. In each case, the important evidence is the visitor’s wording, the approved material available to answer it and the action recorded afterward—not an unsupported assumption about what the visitor decided.

Mixed-language follow-ups deserve the same treatment. A session language provides context, but the sequence of messages reveals whether the visitor changed language or refined the question. Testing mixed-language follow-ups about garage service options is a useful adjacent case for teams reviewing whether the original need remained consistent across follow-up messages.

For businesses in Orange County, Los Angeles and San Diego, CA

In these busy Southern California markets, service businesses often compete for visitors who compare several options online before contacting anyone. Start by reviewing Lumi’s conversation records to identify the exact services, policies, availability or pricing details visitors asked about. Then connect each question with its page path, referring origin, UTM data, language and subsequent tracked actions. Group recurring questions by location or service category to find where website information may need clarification. Keep conclusions evidence-based: a conversation can reveal expressed intent, while a departure without a question provides only session activity, not the visitor’s reason for leaving.

Set up a review process that respects the evidence

A sound visitor-intent process separates observations from interpretations. Record the question as asked, note the page and session context, identify any lead or subsequent action, and label the point where the evidence ends. This prevents a booking-link open from becoming a claimed appointment, a checkout reference from becoming claimed revenue or a page view from becoming an invented motive for leaving.

Lumi is part of the OceSha ecosystem and is also the AI Concierge for OceSha.ai. OceSha Ventures builds and operates AI-first solutions, including course creation, branded academies, AI assistants such as Lumi and business intelligence for businesses and organizations. Lumi is sold by OceSha Inc. and is branded “Powered by OceSha”; it is a product rather than a separate legal entity.

Subscriptions and trials are supported, with monthly AI credits and optional credit top-ups. A top-up of 100 credits for $12 has been tested. If you want to discuss how conversation and journey evidence would apply to your website, contact the Lumi team about visitor intelligence.

The bottom line