AI sales agent

An AI sales agent that answers every lead in seconds

A lead who writes at 11 pm and hears back at 10 the next morning has usually written to someone else in between. The agent replies at once — on the website, by email, in WhatsApp — asks what you would ask, qualifies against your rules, books the meeting in your calendar and writes it all to the CRM. Your team walks into calls that are already warm.

from €2,250Six-month warrantyReplies in secondsYour qualification rules
Works in
  • WhatsApp
  • Telegram
  • Instagram
Try it

Move a lead. Watch the follow-up happen.

Drag a deal to the next stage: the welcome email goes out, the task lands with the right person, the proposal is generated, the invoice is raised. The agent does the same the moment a lead writes.

It’s live — try it right here, nothing saved.

Sales pipeline
Drag a deal to the next stage and watch what the automation does.
DEMO
New lead2
Nordlicht Bakery
Lena
€4,200
Atlas Dental
Marco
€12,800
Qualified2
Kite Studio
Sofia
€3,600
Green Mile Logistics
Jonas
€24,500
Proposal1
Bloom Florists
Amelie
€2,900
Won0
What the automation did
  1. Nothing yet. Move a card forward — every step fires real-world actions.
Drag a card, or use the arrow on it.
What you get

What the agent does with a lead

  • Answers first, in seconds, in your tone

    Website chat, contact form, email, WhatsApp — whichever way a lead arrives, the agent replies at once with a real answer to what they asked, not an autoresponder. It knows your services, prices, timelines and cases, and says so in your voice.

  • Qualifies against your rules

    Budget, timeline, decision maker, fit: the agent asks what you would ask, one question at a time, and scores the lead against the criteria you set. Good fits are booked; poor fits get a polite, useful answer; edge cases go to a person.

  • Books the call and keeps the CRM current

    The agent offers slots from your calendar, books the meeting, sends the confirmation and writes the conversation, the qualification and the next step to the CRM. Nobody re-types anything, and nothing waits until morning.

  • Follows up when a lead goes quiet

    Someone asked for a price and went silent. The agent follows up on a schedule you set, in the channel the lead used, and stops when they answer or after your last step. Every message is logged, so your team sees the whole thread before the call.

Who it is for

An AI sales agent earns its keep where leads come in faster than a small team can answer, and where the first conversation is mostly the same questions. Typical cases:

  • Service businesses with a quote step. Agencies, studios, contractors, consultancies. The lead wants a rough price and a date. The agent gives the ranges you approve, asks what affects the quote and books the call where the real estimate is made.
  • B2B companies with inbound demand. Demo requests and contact forms that sit until a rep gets to them. The agent answers, checks fit against your criteria and books the demo with the right person.
  • Businesses that sell in messengers. Leads who write in WhatsApp or Telegram expect an answer in minutes, not the next working day. The agent answers there, in the same thread.
  • Teams that lose leads after hours. Evenings, weekends, time zones. The agent covers the hours when nobody is at the desk, and the morning starts with booked calls instead of a pile of messages.

Your process, made instant

We start from how your best salesperson handles an inbound lead today — the questions, the order, the answers to objections, when a call is offered. That conversation becomes the agent’s script; your service pages, prices and cases become its knowledge base.

Rules set what the agent decides alone and what it hands over: a budget under your minimum, a request outside your services, a lead who asks for a person. Every conversation is logged to the CRM with a summary, and the first weeks run with your team reading the transcripts.

The agent runs on the channels you use and in your accounts — website, email, WhatsApp, Telegram — connected to your calendar and CRM through their APIs. It is the same kind of agent that greets visitors on this site.

AI sales agent, chatbot or a contact form?

All three sit where the lead arrives. They do very different amounts of work.

  • First reply

    Contact form + autoresponder
    “Thanks, we’ll get back to you”
    Rule-based chatbot
    Instant, from a fixed menu
    AI sales agent
    Instant, a real answer to what was asked
  • Questions it can handle

    Contact form + autoresponder
    None
    Rule-based chatbot
    The ones in the script
    AI sales agent
    Anything covered by your services, prices and cases; the rest goes to a person
  • Checking fit

    Contact form + autoresponder
    Your team, later
    Rule-based chatbot
    Buttons and fixed branches
    AI sales agent
    A conversation against your criteria, with a score and the reasons
  • Booking

    Contact form + autoresponder
    Your team, by email
    Rule-based chatbot
    A calendar link
    AI sales agent
    Offers slots, books, confirms
  • CRM

    Contact form + autoresponder
    Manual entry
    Rule-based chatbot
    Sometimes a new contact
    AI sales agent
    Conversation, answers, score and next step written to the CRM
  • When it’s enough

    Contact form + autoresponder
    Very few leads, long sales cycle
    Rule-based chatbot
    Simple, repeatable requests
    AI sales agent
    Enough leads that speed and follow-up decide who wins

A rule-based bot is cheaper and sometimes the right answer. We’ll say so if it is. See chatbot development for that route.

What the agent doesn’t do

  • It doesn’t close the deal. Pricing exceptions, negotiation and the final yes stay with your people. The agent’s job is to get a well-prepared lead into that conversation fast.
  • It doesn’t cold-message lists. It works with people who came to you. Outbound to cold lists is a different scope and not what this page sells.
  • It doesn’t invent answers. It speaks from your approved information. A question outside it is handed over, and the gap is listed so it can be covered.
  • It doesn’t hide that it’s an AI. It introduces itself and hands over the moment a lead asks for a person.

For a wider look at which sales tasks to give to AI first, and how to measure the result, read our guide AI for sales teams: what to automate first. It covers lead qualification, CRM entry, follow-ups and a rollout plan for a small team.

AI sales agent cost and timing

from €2,250for an AI sales agent on one channel, with your qualification rules, calendar booking and CRM logging.

The estimate depends on the channels, the CRM and calendar integrations and how much of your sales conversation the agent has to cover. Model usage is billed at cost and listed separately.

A first agent can be live from around three weeks once your service descriptions, prices, qualification rules and a sample of past lead conversations are ready.

Six-month warranty. Coverage and ongoing support are set out in your project agreement.

How it works

From first message to launch.

  1. Tell us what needs to work

    Share the goal, the users and any systems you already use. A short message is enough to start.

  2. Agree the scope and quote

    We define the first release, responsibilities, price and delivery plan before development begins.

  3. Review the build

    You see the agreed screens and working flows as the project develops. Changes are assessed before extra work starts.

  4. Launch and hand over

    We prepare the release, test the agreed flows and hand over the project as set out in your agreement.

Proof

Projects behind it

Vera takes over a wedding agency’s daily conversations and brings the manager in only when it matters; Matchbound turns an inbox of invoices into decisions with a person only where the numbers disagree. Both are the pattern a sales agent follows: instant, grounded, handed over at the right moment.

Technologies

The stack we build with.

TypeScriptReactNext.jsNode.jsPythonFastAPIPostgreSQLMongoDBPrismaTailwindJavaScriptSwiftKotlinFlutterUnityStripeTelegramAWSVercelTypeScriptReactNext.jsNode.jsPythonFastAPIPostgreSQLMongoDBPrismaTailwindJavaScriptSwiftKotlinFlutterUnityStripeTelegramAWSVercel
FAQ

Fair questions, straight answers.

Yes, and it should be said plainly — leads mind slow, vague answers far more than an honest assistant that replies instantly and books a call with a person. The agent introduces itself and hands over the moment a lead asks for someone.
Against the criteria you give it — budget, timeline, decision maker, fit with your services — asked as a conversation, not a form. You see the score and the reasons in the CRM, and the rules are tuned on real leads in the first weeks.
We connect the CRM you use (HubSpot, Pipedrive, Salesforce, Notion and others through their APIs) and your calendar (Google, Microsoft, Calendly). Leads, conversations, qualification and bookings are written where your team already works.
Yes. Follow-ups on a schedule you set, in the channel the lead used, stopping when they answer or after the last step you define — with every message logged. Outbound to cold lists is a different scope and is not what this page sells.
The languages your leads write in; the knowledge base can be in one language and the answers in several. We agree the set in the scope and test each one.
A chatbot follows a script of buttons and branches. An AI sales agent understands the question in the lead’s own words, answers from your services, prices and cases, and does the next step itself: checks fit, books the call, updates the CRM. When a lead is outside its rules, it hands over instead of looping.
No. It takes the first reply, the routine questions, the booking and the data entry — the work that happens before a real sales conversation. Your people get more calls that are worth taking, and closing stays with them.
Your service descriptions and prices, the questions you ask before offering a call, the rules for a good and a poor fit, access to the CRM and calendar, and a sample of past lead conversations. With those ready, a first agent can be live from around three weeks.

Show us how you answer a lead today

Send a few real lead conversations and the questions your team asks before offering a call. We’ll turn them into the agent’s script, name the integrations and estimate the work.