A custom AI chatbot serves your customers 24/7 on web and WhatsApp, answers with your company's information and performs real actions (check an order, book an appointment, qualify a lead) connected to your systems. We design it to reduce support load and capture more opportunities, with handover to a person when needed.
By Oliver Guirado · Updated on October 28, 2026 · 7 min read
What an AI chatbot is and why it is not "the usual chat"
A classic chatbot follows a tree of options: if the customer strays from the script, it fails. A generative AI chatbot understands natural language, keeps the thread of the conversation and answers with your company's real information. Well built, it also carries out actions: it checks the status of an order, books an appointment, opens a ticket or qualifies a contact and saves it in the CRM.
The key to making it reliable is how it connects to your knowledge. Our chatbots answer from your documents, catalogue and policies through RAG, cite the source when appropriate and refuse to make things up when they lack data. And when the conversation requires it, they hand over to a person with the summary already prepared.
Answers anchored to your information: the chatbot only replies with what it retrieves from your sources and cites it.
Scope limits: which topics it can handle and which it must pass on (for example complaints, legal or medical matters).
Evaluation before and after launch: a set of real questions with expected answers that runs before every change.
Handover to a person with the history and a summary, so the customer repeats nothing.
Transparency: the user knows they are talking to an AI system, something European regulation also requires.
Conversation review to detect gaps and improve the knowledge base every week.
Sample AI chatbot projects
Sample project · E-commerce
Orders and returns assistant for an online fashion store
Problem
The support team answered the same questions all day (where is my order, how do I return, which size should I pick) and during campaign peaks response times soared.
Solution
A chatbot on web and WhatsApp connected to the store and the carrier: it reports order status, starts returns with their label, recommends a size using the brand guide and passes complaints to a person.
Steps
Analysis of the most frequent queries and of the shipping and returns policies.
Connection to the store platform and to carrier tracking.
Knowledge base with size guides and conditions, with cited answers.
Tests with historical conversations, tuning of tone and handover rules.
Staggered launch and weekly review of conversations.
Target outcome
Fewer repetitive queries for the team, an immediate response at any hour and human attention concentrated on the cases that need it. More on custom online stores.
Technology
WhatsApp Business API, e-commerce platform, knowledge base with RAG, language model with EU data.
Sample project · Healthcare
Appointment and administrative chatbot for a clinic
Problem
Reception handled calls and messages for appointments, changes, test preparation and invoices, with busy lines and patients who gave up.
Solution
A chatbot that manages appointments on the real diary, sends reminders and prior instructions and answers administrative questions. It gives no medical advice: faced with symptoms or clinical doubts it hands over to a person and says so clearly.
Steps
Map of contact reasons and of diary rules by specialty.
Integration with the clinic management software and with WhatsApp.
Definition of forbidden topics and of the handover protocol.
Privacy and consent review before launch.
Pilot in one specialty and gradual roll-out.
Target outcome
Fewer admin calls, fewer appointments lost for lack of a reply and a reception dedicated to in-person care. More in apps and software for healthcare.
Technology
Clinic management software via API, WhatsApp Business, RAG over administrative protocols, consent log.
The projects on this page are illustrative, anonymised examples of how we approach this kind of work. Every real project is defined with your information, your systems and your goals.
How we develop your chatbot
Discovery: channels, query volume, systems to connect and measurable goals (resolution without a person, leads, appointments).
Knowledge base: we organise and prepare your documents, FAQs and catalogue so the chatbot answers accurately.
Conversation design: brand tone, limits, flows with actions and handover rules.
Development and integrations: CRM, store, diary, ERP or helpdesk, with minimum permissions.
Evaluation: tests with real conversations and accuracy measurement before opening it to the public.
Launch and continuous improvement: metrics dashboard, conversation review and expansion of cases.
Channels, integrations and metrics
We deploy the chatbot on your website (own widget), on WhatsApp Business, in mobile apps and, if needed, on Instagram or Telegram. The usual integrations are CRMs, e-commerce platforms, diaries, ERPs and support tools. If your product is an app, we plan it as part of the AI integration in your software.
We measure what matters to the business: resolution rate without human intervention, handover rate, first response time, satisfaction, qualified leads and appointments generated. With those metrics the chatbot improves every week instead of staying a static project.
Reference prices for an AI chatbot
Scope
Indicative investment
Typical timeline
FAQ chatbot with AI on one channel
€4,000 – €9,000
3 – 5 weeks
Assistant over your documents (RAG), one channel and metrics dashboard
€9,000 – €25,000
6 – 10 weeks
Chatbot with actions in your systems (CRM, orders, appointments) and several channels
€20,000 – €60,000
2 – 4 months
Maintenance, improvement and conversation review
€300 – €1,500 / month
Ongoing
Indicative reference prices in euros, excluding VAT, for projects in Spain and Europe in 2026. The final amount depends on scope, integrations and data quality; we confirm it with a fixed quote after a first meeting.
On top of these figures come the cost of using the language model and, on WhatsApp, the messaging provider's fees, which we calculate with you according to your volume. Details of the factors in how much an AI chatbot or agent costs.
Privacy and regulatory compliance
We apply GDPR by design: data minimisation, clear information to the user, a data processing agreement with the providers and, when appropriate, hosting in the EU. The chatbot states that it is an AI system and always offers a way to speak to a person. Conversation data is kept for as long as necessary and can be deleted on request.
Frequently asked questions
How much does an AI chatbot cost for a company?
An AI FAQ chatbot starts between €4,000 and €9,000; an assistant over your documents with a metrics dashboard, between €9,000 and €25,000; and a chatbot with actions in your systems and several channels, between €20,000 and €60,000. These are reference prices excluding VAT.
Can a chatbot serve customers on WhatsApp?
Yes. We connect it to the official WhatsApp Business API. You need to take into account the provider's messaging fees and the approved templates for messages initiated by the company.
How do you make sure it does not give wrong answers?
By answering only with your retrieved and cited information, limiting the topics it can handle, evaluating with real questions before launch and handing over to a person when confidence is low.
Does it integrate with my CRM, store or diary?
Yes, as long as the system has an API or connector. That is what lets the chatbot take real actions instead of only answering.
How long does it take to have a chatbot running?
A simple chatbot can be ready in 3-5 weeks and one with integrations and several channels in 2-4 months, including evaluation and tuning.
What is the difference between a chatbot and an AI agent?
The chatbot converses and answers; the agent also plans and carries out complete tasks in your systems. Many projects start with a chatbot and evolve into an agent.
Shall we talk about your project?
Tell us about your case and we will reply within 24-48 hours with a proposed scope and an indicative quote.
