A visitor reads the pricing or solutions page, scrolls to the bottom, does not fill the form and leaves. The traffic was paid for, the interest was real, and nobody knows who it was.
AI solution 03 · Lead capture
The anonymous website visitor
A website chat that opens only when a visitor shows real buying intent, answers from your own knowledge base, and sends the lead to the CRM with every page and question attached.
AI agent · Website
A chat that opens only on real intent signals: time on page, scroll depth and no clicked call to action. It asks a question about that page, answers from the company knowledge base, and asks for a phone number or email only once the visitor is engaged. Pages viewed, time spent and questions asked go to the CRM with the lead.
Anonymous traffic becomes named pipeline, and the rep gets the whole conversation, not just an email address.
The problem
The usual fix, a chat pop-up that opens on every page after five seconds, annoys the people who were only browsing and gets ignored by the ones who were serious.
Why it happens
- B2B buyers want answers before they want a sales call, and a form offers only the call.
- Generic pop-ups fire on time alone, so they interrupt everyone and help nobody.
- When a form is filled, the rep gets a name and an email, but not what the visitor read or asked.
How the system works
- 01
Intent signals, not timers
The chat watches for a combination of time on page, scroll depth and the absence of a clicked call to action on the pages that matter, such as pricing, solutions and product pages.
- 02
A question about that page
It opens with something specific to what the visitor is reading, not “How can I help you?”.
- 03
Answers first, details later
Questions are answered from your own knowledge base. Only once the visitor is clearly engaged does the chat ask for a phone number or email.
- 04
Full context to the CRM
The lead arrives with pages visited, time spent and every question asked, so the rep knows exactly what the buyer was looking at.
Where AI does the work
Rules first, AI second. The model runs only where judgement on text is needed, so running costs stay low as volume grows.
Deciding when to open
Time, scroll and click signals are rules, so the trigger costs nothing to run and never opens on a casual visit.
Answering questions
Reading the question and answering from your knowledge base is language work, so the agent does it, and only from approved content.
CRM handover
Creating the lead with the session history attached is plain integration code.
What changed
- Visitors who would have left without a trace become named leads.
- Reps get the conversation and browsing history, not just a contact detail.
- Casual visitors are left alone, so the site does not feel like a sales trap.
Works with your stack
Runs on any website, whatever it is built on, and writes to Zoho, HubSpot, Salesforce, Pipedrive, Freshworks or any CRM with an API.
Nothing is replaced. It is a layer on top of the tools you already use, built on your accounts, so the data stays yours.
Questions
Will it make things up about our products?
It answers only from the knowledge base we build with you. If something is not covered, it says so and offers to put the visitor in touch with a person.
Does it slow the website down?
No. The chat loads after the page and only activates when the intent rules are met.
Can we control which pages it appears on?
Yes. Most teams start with pricing, product and solution pages and leave blog and careers pages alone.
What does it cost to run?
We build it in your own environment, and the AI runs on your own Google (Gemini) or OpenAI account. You pay them directly, pay as you go, only for what is used. We add no markup and charge nothing per message or per seat. Because rules handle most of the work before any model is called, the AI bill stays small.
Running cost and data
What it costs to run
We build the system in your own environment. You pay only for the AI it uses, directly to the AI provider, with no markup from us.
The system is set up in your own cloud account. Conversations, recordings and CRM data move between your systems and the AI provider you choose. They do not pass through our servers or a separate chatbot platform.
The AI runs on your own account with Google (Gemini) or OpenAI. You pay them directly, only for what is used, at their published rates.
We do not resell AI usage or charge per message, per conversation or per seat. Our fee is for building and improving the system, and it is agreed upfront.
Rules and plain code do most of the work, and the model is called only where judgement on text is needed, so the AI bill stays small as volume grows.
WhatsApp and calling charges, where they apply, are billed by your provider as they are today.
Related solutions
A WhatsApp AI agent that answers every enquiry in under 30 seconds, at any hour, asks the qualifying questions and passes the lead to a rep only when it is worth a call.
Read the case study →06 · Data quality and qualificationEvery rep qualifies differentlyYour qualification criteria written down once and applied by AI to every lead the same way, with a score, the reasons behind it and a recommended next step.
Read the case study →04 · Data quality and qualificationThe half-empty CRMMissing CRM fields filled automatically from your own conversations and email signatures, without ever overwriting what a person has typed.
Read the case study →Sound familiar?
Thirty minutes is enough to map where leads come from and where they drop off. If there is a fit, you get a short written diagnosis of the two or three problems worth fixing first, with a rough investment range.