An AI agent, deployed at your front door.
A Claude agent that qualifies, books, or routes elsewhere. Honestly. It is the same agent brinvik.com runs, and we agree how success is measured before I build it.
FreeNo commitmentA yes, a no, or a plan
- 7certifications from Anthropic
- 18+ yearsselling, hiring and leading sales teams
- Fixed priceon every step, before you say yes
- You ownthe system when I am done
Leads that reach your team
Only the leads worth a sales call reach your team, booked straight into your CRM, while the rest get an honest referral. Generic chat widgets and contact forms catch noise, not qualified intent. A website AI agent does the filtering at your front door.
What it does with what it learns
Pricing
€4,025
The typical tier for a website agent. Agents with multiple integrations move to Project Core. The audit (€670, one week) comes first and is credited against the project fee.
Does this fit you? Seven taps, an honest answer.
The check for the website agent.
What's included
- 01Outcome agreement
Before any build, we agree what success looks like and how we measure it. That outcome is what you pay for. The agent is live as a working proof of concept in two to four weeks, and because results take some time to show, it is tuned against the agreed measures from there.
- 02Custom system prompt aligned to your ICP
The agent qualifies on what actually matters to your sales team, not generic chat.
- 03CRM integration
HubSpot, Salesforce, or Pipedrive. Qualified leads land where your team already lives.
- 04Booking integration
Real calendar slots via Calendly, HubSpot Meetings, Google Calendar, or your own scheduling tool. No back-and-forth email.
- 05Cloudflare Turnstile bot protection
Invisible challenge keeps bots out of the funnel without friction for humans.
- 06Rate limits and disposable email blocking
Cuts the noise before it reaches your team.
- 0730-day post-launch optimization
I tune prompts and qualification logic against the first 30 days of real conversations.
Timeline
Audit, ICP definition and the outcome agreement. I learn your sales motion in detail, and we agree how success is measured.
Agent design and deploy. System prompt, qualification logic, behaviour rules.
Integrations and testing. CRM, booking, Turnstile, rate limits.
Optimization. Tuning against the first conversations and the measures we agreed. Simple deployments finish in week 2 or 3.
How I work
First we agree the outcome: what a good result looks like for your inbound, and how we measure whether the agent is good enough. Then I build a custom Claude agent on your site, the same one running on brinvik.com: it qualifies, books, or politely disqualifies with a referral. I tune the system prompt to your real ICP, connect it to your CRM, and optimize against the first 30 days of real conversations. It is a website AI agent that earns trust by saying no when the fit is wrong.
Not sure an agent is the right answer yet? The guide covers what an AI agent is, what it is not, the three kinds companies actually use, and where they go wrong.
Read the guide to AI agents →When this is right
A fit when
- A company of 5 to 80 people with a CRM in place and inbound that needs sorting.
- Current website traffic above roughly 1,000 monthly visitors.
- A sales team currently overwhelmed by unqualified leads.
Probably not when
- Pure B2C consumer brands.
- Sites with under 500 monthly visitors. The economics don't pay back.
- Businesses without a CRM. Set the CRM up first; the agent is the second move.
FAQ
How do I add an AI agent to my website?
I build a custom Claude agent for your site, the same one running on brinvik.com. We agree up front what success looks like and how it is measured. It is tuned to your ICP, connected to your CRM and booking tool, and live in two to four weeks as a working proof of concept. Results take some time to show, so we measure them against what we agreed. A website AI agent qualifies real intent instead of catching noise like a generic AI chatbot for a website.
How do I qualify leads with AI?
The agent scores every visitor against your real sales criteria, books the good ones into your CRM, and refers the rest honestly. Your team stops sifting unqualified inbound and spends its time on leads worth a call. It works as an AI lead qualification agent, not a contact form.
No. It removes the unqualified noise so SDRs spend their time on real opportunities. It's a filter, not a replacement.
Yes. The qualification rules live in a system prompt and a small config file. Changing them is a 30-minute conversation, not a re-engagement.
It flows into your CRM under your account. Brinvik doesn't retain the data after handoff. The agent itself doesn't train on visitor conversations.
Those are chatbots optimized for sales playbooks. The Brinvik agent is honest disqualification first. It books the right leads and refers the wrong ones elsewhere. The economics are also fundamentally different: no per-seat fees, no engagement charges.
Other services
- WorkshopTeach the team to use Claude on the work they already do.→
- Internal ClaudeClaude inside the team, taking the repetitive work off the desk.→
- RetainerMonthly tuning once something is live, so it keeps getting better.→
- AdvisoryJudgement on call. Sessions and written decision memos, nothing built.→
- Public speakingA talk about AI with a real system on screen, in Danish or English.→
A website AI agent
that earns its keep.
Tell me what a good lead looks like. We agree how success will be measured, and I'll scope a Claude agent that qualifies, books, or refers, honestly. Either way, you leave with a plan.
You pick a slot in Kim's calendar. No form, no callback.
Certifications from Anthropic · Fixed price on every stage, outcome agreed before the build · Invoiced in DKK