Website AgentCopy link to this section
The website agent is Floe at the top of the funnel. Before a visitor is ready for a full demo, a compact Ask Floe launcher — or a Talk to me presenter when you enable and configure a video avatar — sits on your marketing site and answers their questions, grounded in your product. When they ask to see something live, the same agent hands off in place to a live demo. No new tab, no form, no wait.
It runs on the same session as the demo agent. A visitor can start by typing a question and finish by watching the agent drive your real product — one continuous conversation.
Answer questions before the demoCopy link to this section
Not every visitor is ready to sit through a demo. Most have a question first: "Does this integrate with Salesforce?", "How is this different from Storylane?", "What does setup involve?"
The website agent answers those questions right on the page. A visitor clicks the primary Ask Floe or Talk to me launcher and the agent begins connecting immediately with the visitor's microphone off. In floating mode the conversation grows from that same corner; sidebar mode opens its rail, and phones use the full screen. An avatar does not open a chat window at all: the presenter stays where it was and comes alive in place, with what it is saying in a bubble beside it and a compact call bar below that. See the avatar call surface. Opening the surface is itself a request for a response: without a supplied prompt, the agent immediately welcomes and orients the visitor; with a typed question or selected nudge, it answers that prompt directly instead of giving a generic greeting. Only preparation that happens before the visitor opens the surface stays silent, and an unused session still expires automatically. While live-demo readiness is prepared in the background, Q&A stays focused on the visitor's question and never exposes preparation or loading status. That warm start keeps the first exchange fast without asking for device access. Answers come from your documentation and product content — the same knowledge that powers the demo agent. It's the top-of-funnel conversation that used to require a rep or a form.
Grounded answers, honest about gapsCopy link to this section
Answers are grounded in your ingested content. When the agent can answer from that content, it does. When it can't — when a visitor asks about something you haven't documented — it says so and offers to connect them, rather than inventing an answer. A wrong answer on your marketing site is worse than an honest "I don't have that detail."
This is the same grounding discipline as the demo agent. The agent only speaks to what it learned from your content and product.
Contextual nudges as they scrollCopy link to this section
As a visitor moves through a page, the website agent can surface a short, contextual nudge tied to the section they're reading — a prompt like "Curious how scoring works? Ask me." Nudges are page-scoped: each one belongs to a specific page and section, so the prompt matches what's actually on screen, and follows the visitor as they navigate your site. Clicking a nudge sends that exact question as the visitor's next turn — or as the first turn of a newly opened conversation.
The standard pill launcher keeps its prompt visible while the visitor scrolls. It scales down as one surface — icon, copy, and all — then returns to full size after five seconds without scrolling. Each new scroll folds away an earlier pointer reveal; moving the pointer over the launcher again restores it immediately. A launcher reached with the keyboard stays expanded and keeps its controls available while that visitor scrolls, until focus leaves it. Section-aware nudges remain available and update with the visitor's position instead of disappearing during the scroll. An avatar launcher keeps a stable Talk to me action while its contextual nudges remain separate at the bottom center, where they stay visually connected to the presenter.
Nudge copy is generated from your content and prioritizes your buying-intent pages (pricing, product, comparisons). You control every nudge from the dashboard — edit the copy, or switch a page's nudge off entirely. See Website Agent Setup.
Suggested questionsCopy link to this section
To get a conversation started, the chat offers suggested question chips. There are two kinds:
- Openers — shown when the chat first opens, generated from the page context or your product content (with a per-site override in the dashboard). They give a visitor an easy first thing to ask. Floe does not invent generic questions when no relevant opener is available; if the site has a live demo, it offers Show live demo instead.
- Follow-ups — after the agent replies, it suggests a few relevant next questions based on the conversation so far, so the visitor can keep going with one tap.
Type or talkCopy link to this section
The visitor's microphone starts off, and opening the agent does not initialize the microphone or request browser permission. The call controls label that state Mic off · Tap to talk so a visitor does not mistake the warm conversation for an active microphone. Any background preparation before a deliberate open remains silent. Choosing the primary launcher or Talk to me immediately starts an agent-first voice conversation: the agent welcomes and orients the visitor with its speaker on while the visitor remains unheard. Floe requests microphone access only when the visitor explicitly turns the mic on. If access is blocked or no usable microphone is available, the same control explains the problem and remains available to retry; typed chat continues to work.
For a quiet text start, hovering or keyboard-focusing the launcher reveals a separate keyboard action (inside the pill or beside the avatar presenter). It opens with the agent's speaker muted, immediately renders the same concise welcome and orientation in writing, and puts focus in the composer. With a video avatar that start opens the call bar directly into its composer rather than a chat window, so the presenter stays on the page. Selecting a nudge instead sends that exact prompt as the first turn, so the agent answers it directly. Once the conversation is open, the visitor can type, mute or unmute the agent's replies, or turn their own microphone on. Without an avatar the composer is always on screen; with one it is one tap away on the call bar, so a voice call is not sharing its surface with a text box nobody is using. Either way it is the same agent and the same warm session.
Answers render as formatted text — lists, links, and emphasis — so a multi-part answer stays readable in the chat.
The avatar call surfaceCopy link to this section
With a video avatar configured, the agent is a presenter standing on your page rather than a chat window that opens over it. Choosing Talk to me does not open a panel: the presenter stays exactly where it was and comes alive in place.
Two things attach to it. What the agent is saying appears in a bubble beside the presenter, with the turn before it kept above in grey — so a visitor who looks away mid-answer still has the thread. Earlier expands that same bubble into a scrollable transcript of the conversation, labelled by speaker, and collapses back to the live line.
Below the bubble is a compact call bar: microphone, end call, and a control that reveals the rest — type instead, and mute the agent's replies. A voice call shows no text field until the visitor chooses one, and choosing it turns the bar itself into the composer rather than opening a second surface. The presenter never becomes a thumbnail inside a card.
Sites without a video avatar are unchanged: the launcher opens the standard chat window with its composer and controls.
Hand off to a live demo, in placeCopy link to this section
This is what separates the website agent from a support chatbot. When a visitor asks to see something — "Can you show me?" — the agent doesn't dead-end at text. It transitions the same session into a live demo: it opens your real product in a separate demo stage opposite the anchored chat. The chat and its outcome actions stay in place while the visitor watches the agent drive.
Expanding the demo grows the video behind that anchored chat instead of moving the conversation or its controls. A visitor can minimize the message window to reveal more of the demo underneath while keeping the compact interaction controls available. With a video avatar there is no window to minimize: the presenter and its call bar stay exactly where they were and the demo takes the space beside them.
There's no navigation away, no second form, no scheduling. The question becomes a demo in one continuous flow. From there it's the full demo agent: voice-led, driving your real UI, qualifying against your ICP.
The See it live action is available in the open chat surface until the visitor selects it. It disappears immediately so the visitor is not invited to start the same demo twice while the hand-off begins. The avatar call surface does not carry it as a permanent button — the agent offers the demo in conversation instead, and asking for one works the same way. You can turn the demo hand-off off, or set the label on its call-to-action, from the dashboard.
Know your visitorsCopy link to this section
Visitors do not need to share an email to start. When visitor data collection is enabled and any required consent is available, Floe uses a first-party browser identifier to keep activity from that browser together. If you choose to, the agent can ask for an email (and optionally name and company) inline, at a moment you configure: after the first answer, after a chosen number of answered questions, or when the visitor shows demo intent. The card never delays an answer.
When consent is required by the embed, identity capture stays hidden until the host grants it and disappears immediately if consent is denied. If a visitor shares an email, Floe records it as a contact claim attached to that exact browser and conversation so your team can review the source interaction. The claim does not create or merge a canonical person record. Another browser can share the same address and remains a separate visitor with separate history.
A browser identifier is a continuity aid, not authentication or proof that two people are the same. On a shared device, more than one person can use the same browser identity, and that continuity does not follow someone to a different browser or device. See Security & Privacy and the SDK consent controls.
Floe can continue an active Website Agent conversation across pages and tabs in the same browser. If the previous thread is no longer active, Floe starts a new conversation and, when collection and consent settings permit, can use retained context from that browser to greet naturally and avoid asking the same questions again. If that browser continues into a demo, the demo stays attached to the same source record. The agent does not use an email address alone to recall conversation or demo context from another browser. Clearing the site's browser storage starts a new anonymous visitor record.
Which nudges get shown and clicked is tracked so you can see what's drawing interest across your site.
What reaches your sales teamCopy link to this section
A Website Agent conversation can produce the same internal follow-up as a full demo even when it ends before a live hand-off. A completed or abandoned website session becomes eligible only when that exact interaction produced an active email claim or concrete qualification or intent evidence, such as buying signals, budget, qualification answers, pain points, product interest, or a lead score/tier. A name, page view, or generic Q&A by itself does not trigger sales traffic.
Internal email, Slack, and Teams identify the record as Confirmed contact, Unverified email claim, or Anonymous lead. CRM contact writes require a confirmed claim; an unverified or anonymous lead remains internal. Claim-backed capture and CRM delivery are staged deployment capabilities, so contact Floe to enable them. See Integrations for the delivery and late-confirmation behavior.
How it's different from a chatbotCopy link to this section
A support chatbot answers text from a static knowledge base and stops there. The website agent:
- Is grounded and honest — it answers from your product content and declines when it doesn't know, instead of guessing.
- Is proactive — it surfaces contextual nudges tied to the page a visitor is on, rather than waiting to be opened.
- Ends in a demo, not a transcript — when the visitor is ready, the same agent drives your real product live. The chat is the on-ramp to the demo, not a replacement for it.
Getting startedCopy link to this section
The website agent shares its knowledge with the demo agent, so setup builds on the same foundation:
- Create a site for your product
- Add content so Floe can answer accurately
- Enable and configure the website agent — see Website Agent Setup
- Embed Floe on your marketing site with the Browser SDK
FAQCopy link to this section
Can it answer questions before starting a demo? Yes — that's its main job. The Ask Floe launcher answers questions right on your marketing site, grounded in your product content, before a visitor commits to a full demo.
Does it show contextual nudges as I scroll? Yes. Nudges are page-scoped and section-aware: as a visitor reads, the agent can surface a short prompt tied to what's on screen. You control the copy and can turn any page's nudge off.
Can it hand off to a live demo in place? Yes. When a visitor asks to see the product, the same session transitions into a live demo — the agent opens your real product in a stage opposite the anchored chat. Expanding the video keeps the chat and controls in place, and minimizing the chat reveals the demo beneath it. No new tab, no form.
Is the website agent the same as the demo agent? They're two phases of one agent on one session. The website agent handles the top-of-funnel Q&A; when the visitor is ready it becomes the demo agent and drives your product. Same knowledge, same conversation.
Do visitors have to talk? No. The visitor's mic starts off and no permission is requested on open. Choosing the primary launcher or Talk to me makes the agent welcome them immediately with its speaker on; when that spoken welcome finishes, a one-time Ready to talk? invitation offers to turn on the mic. The visitor can dismiss it and type at any time. Choosing the keyboard action produces the same immediate welcome in writing with the speaker muted. Any background preparation before a deliberate open stays silent, and Floe requests microphone access only when the visitor explicitly chooses Turn on mic or taps the microphone control.
Does the chat remember a visitor across visits? A returning browser can continue an active conversation across pages and tabs. Once that conversation is no longer active, a later visit starts a new conversation and may use that browser's retained conversation context to pick up a relevant topic naturally. Floe never treats a browser identifier as proof of an authenticated person or uses an email address alone to recall conversation context from another browser or device.