AI customer service
An AI agent that answers questions on WhatsApp, on your website or by email from your own approved information, and hands over to a person when it should.
AI for business
We build AI agents and assistants that answer customers, read documents and support your team, using your own information and working inside the tools you already use.
I'm an AI agent too. Want to see what I know?
Many business owners have tried a chatbot or asked an AI tool to write an email. It is impressive at first, and then it is unclear how it helps the business: it doesn't know your prices, can't see your calendar and sometimes simply makes things up. The real gains come when AI is connected to your information and your workflows, with clear limits.
Generic AI tools that know nothing about your business
Customer questions piling up after hours and at weekends
Hours lost reading, sorting and retyping documents
We start with one clear job for AI to do, such as answering customer questions, sorting incoming requests or pulling data out of documents. Then we give it the right knowledge (your services, policies, prices and past answers) and connect it to your WhatsApp, CRM or internal systems. We set rules for what it may answer, what it must pass to a person and how its work is checked. The result is an AI agent that saves real time and that you can put in front of customers with confidence.
AI for business
An AI agent that answers questions on WhatsApp, on your website or by email from your own approved information, and hands over to a person when it should.
Your WhatsApp bot understands messages written in the customer's own words and language, not just button taps, so conversations feel natural.
Automatic summaries of calls and chats, suggested next steps, lead prioritization and draft replies, right inside each customer record.
AI reads invoices, forms, contracts, CVs and emails, extracts the data that matters and sends it to the right system, with a person checking anything unclear.
A private assistant that answers your team's questions from your procedures, price lists, product specs and past projects, so knowledge doesn't live in one person's head.
First drafts of product descriptions, social posts, newsletters and replies, written in your brand voice and ready for a human edit.
How it works
Four panels, like a yonkoma comic: setup, build, twist and payoff.
We look at where your team spends time on repetitive reading, writing and answering, and choose a first use case with clear value.
We gather the information the AI needs, tidy it up and define its limits: what it answers, what it never touches and when it calls a human.
We build the agent, connect it to your tools and test it against real customer questions and documents before anyone outside sees it.
We go live with human oversight, review conversations and results with you, and keep improving the knowledge and rules over time.
Examples
An online skincare store in Paris uses an AI agent on WhatsApp to answer questions about ingredients, skin types and shipping, using only its own product data. When a customer reports a reaction or a problem with an order, the agent passes the chat to the team straight away.
A restaurant in Tokyo gets questions from visitors in many languages about the menu, opening hours and reservations. The AI agent replies in each guest's language and books tables, while any question about allergies always goes to a member of staff.
An accounting firm in London asks clients to send receipts and invoices by email or WhatsApp. AI reads each document, extracts the date, supplier and amounts, and adds them to the right client file for an accountant to approve.
A property management company in Dubai receives maintenance requests at all hours, often with photos. AI reads each message, sorts it by urgency and type, opens a ticket and tells the tenant what happens next.
A creative agency in Lisbon uses an internal AI assistant trained on its past proposals and brand guidelines. Account managers start every proposal and monthly report from a solid first draft instead of a blank page.
Kaede
Kaede, the character on this site, is our own AI agent. We build for clients the same kind of agent we trust to welcome our own visitors.
Our AI works inside the WhatsApp bots, CRMs and business systems we build, so it can look things up and take action, not just chat.
We tell you honestly what AI can and can't do for your business today, start small and scale only what proves useful.
AI has moved well past writing clever emails. An AI agent is a system that understands requests in plain language, looks up the information it needs and takes a next step: answering, sorting, filling in a record or handing over to a person. For a small business, that means AI can take on part of the reading, writing and answering that fills everyone's day.
The key word is 'part'. Good AI projects pick specific jobs, give the AI the right information and connect it to the tools where the work happens. Vague plans to 'add AI everywhere' tend to disappoint.
Customers ask the same things in endless different ways. AI understands those variations and answers from your approved information, on WhatsApp, on your website or by email. Complex or sensitive cases go to your team with a summary, so the customer never has to tell the story twice. This is where AI customer service earns its place fastest.
Inside a CRM, AI can summarize a long WhatsApp conversation, suggest the next step for a deal, draft a follow-up message and point out leads that look ready to buy. Salespeople stay in charge, but they start every call better prepared.
Invoices, delivery notes, application forms, contracts and CVs all contain data that someone has to type into a system. AI can read them, pull out the right fields and pass them on for a quick human check, turning a slow manual job into a short review.
Every business has knowledge spread across documents, chats and people's heads. An internal assistant trained on your procedures and past work lets new and experienced staff find answers without interrupting a colleague.
The terms are often mixed up. A classic chatbot follows a script: it recognizes keywords and offers buttons. An AI agent understands intent, works with your real data and can act through connected tools. In practice the best solution often uses both, with scripted paths where you want predictability and AI where customers need flexibility.
We define what the AI may answer, what it must never answer and when it hands over. Medical, legal, financial and safety questions usually stay with people.
The agent answers from sources you approve, such as your website, price list, policies and FAQs. When those change, the answers change too.
We choose providers and settings that match your privacy needs, limit what data the AI can see and keep sensitive information out wherever it isn't needed.
In the first weeks after launch we review conversations and outputs closely with you. Mistakes are fixed at the source, by improving the knowledge or the rules, rather than patched one case at a time.
A separate AI tool that nobody opens doesn't help anyone. We put AI where your work already happens: inside your WhatsApp bot, inside your custom CRM and inside the business systems that run your operations. That way the AI can check a booking, update a record or create a task, and every action it takes is logged.
The best first project is frequent, well defined and easy to measure, such as answering common WhatsApp questions or extracting data from incoming documents. Once it works, the same knowledge and connections can power the next use case. Step by step, AI automation becomes part of the journey from the first click to the closed deal, without losing the human touch your customers value.
Kaede, the ink drawn character on our site, is our own AI agent. You can ask her about our services right now, and she is a good preview of what an agent trained on your business could do for your customers.
A regular chatbot follows a fixed script of buttons and keywords. An AI agent understands free text, uses your business information to answer and can take actions through connected tools, such as checking availability, creating a lead or opening a ticket. It still works within the rules you set.
Yes, always. You can take just an AI project and nothing else. If you want to connect more pieces later, everything is built to plug in easily. The simplest way to start is a short consultation call on WhatsApp: tell us what your business needs and we will tell you honestly what is worth doing and what is not.
Book a consultation on WhatsApp
Any AI model can get things wrong, so we design around that. The agent answers from your approved information, is instructed to say when it doesn't know, and passes sensitive or unclear questions to a person. We test it with real questions before launch and keep reviewing conversations afterwards.
We set up AI services on business plans and settings where your data is not used to train public models, and we tell you which provider is used and where the data is processed. You decide what information the AI may access, and sensitive data can be kept out entirely.
We are not tied to one provider. Depending on the task, the languages and your privacy needs, we can work with models from providers such as OpenAI, Anthropic or Google. We choose what fits the job and can switch later if something better comes along.
It can take over the repetitive questions and the after hours messages, which frees your team for the conversations that need empathy and judgement. We design AI to work alongside people with a smooth handoff, not to leave customers alone with a machine.
No. Your website, price list, FAQs, policies and some example conversations are often enough for a first agent. We handle the technical side and show your team how to update the knowledge when something changes.
Usually something frequent and well defined: answering common customer questions on WhatsApp, sorting incoming requests or extracting data from documents you receive every day. A focused first project shows its value quickly and teaches you where AI should go next.
From the journal
AI for businessCustomers don't mind AI. They mind being stuck with it. Here is how to use AI in customer service to answer faster while keeping the warmth that makes people come back.
AI for businessAI agents can now answer customers, qualify leads and handle paperwork, but they are not magic employees. Here is what works today, where the limits are and how to start.
Related services
AI for business
Tell us where your team spends its time answering, reading or retyping. We'll suggest where AI can help first, and where it shouldn't be used at all.
Kaede is an AI. Please do not share sensitive information. Privacy