See the little K button in the bottom-right corner? That’s Kyn, and Kyn is exactly this service. It’s not a faked demo or a script: it’s an AI chatbot trained on everything KyndredAI does, answering live. Before you read on, try it: open Kyn and ask it anything. What it answers is, more or less, what a chatbot like this can do for your business.

What is an AI chatbot?

An AI chatbot is an assistant that converses in natural language and answers using your business’s own information: your prices, services, and policies. It understands what a person types in their own words, replies instantly in your brand’s voice, and when it doesn’t know something it hands off to a human instead of making things up. It lives where your customers already are: the web, WhatsApp, Instagram.

The old-school chatbot followed a tree of buttons: “Press 1 for sales, 2 for support.” Rigid, and it broke the moment a customer typed something that wasn’t in the menu. An AI chatbot is a different animal: it understands what the person writes in their own words, connects it to what it knows about your business, and answers the way someone on your team who knows everything by heart would.

The difference that matters is that it doesn’t make things up. A well-built chatbot isn’t generic ChatGPT answering from general knowledge: it’s tied to your information — your pricing, your services, your policies, your FAQs — and answers from there. If it doesn’t know something, it says so and hands you off, instead of improvising. That’s exactly the line between a tool that helps and one that gets you into trouble.

How it works

Under the hood there are three pieces. The model that understands and writes (I use Claude, and I’ll get to why). Your knowledge: your business material loaded so the bot consults it before answering. And the rules: the tone, what it can and can’t say, when to escalate to a human. When someone writes, the bot interprets the question, looks up the right answer in your information, and writes it in your brand’s voice.

And here’s what really sets it apart from a robot that fields loose questions: we don’t load it four answers, we teach it your whole business. Where you are, your rules, your tone, your style, your hours: all of it lives in its brain and shapes every reply. That’s what’s valuable, and it’s exactly what KyndredAI does: build everything you know about your company into it.

It lives where your customers already are: the website bubble (like Kyn here), WhatsApp, Instagram. You don’t have to send anyone to a new app; the bot goes to where people already message you. The best way to understand the “how”? I won’t tell you, better to try it: ask Kyn how we work and see how it answers.

What it does for your day to day

The pattern is clear: if you get the same questions over and over, a chatbot takes them off your plate.

Support that never sleeps. Most enquiries come in off-hours, at night, on the weekend, when you’re not around: 39% of customer chats happen outside 9-to-5 (Drift platform data, 2024). A customer who asks a price at 11pm and doesn’t hear back until Monday has probably already bought from someone else by Monday. The bot answers instantly, always.

The same questions as always. Hours, prices, payment methods, “do you ship?”, “do you cover my area?”. It’s the bulk of the messages and exactly what eats your time. Automated, your energy goes to the enquiries that genuinely need your judgement.

Qualify and book. A good bot doesn’t just answer: it works out whether the person writing is a serious lead, asks for the details that matter, and books the call straight into your calendar. You start your day with meetings, not with an unopened inbox.

Many at once. A person handles one chat at a time; the bot handles a hundred without breaking a sweat. In a demand spike, nobody is left waiting.

What I build them with, and why

The brain of my chatbots is Claude, by Anthropic. For two concrete reasons: it’s the model that writes most naturally — it sounds human, not like a translated robot — and it’s the one that “makes things up” least when it doesn’t know something, which in customer support is what matters most. A bot that invents a price or promises something that doesn’t exist does more harm than no bot at all.

But the model is only one part. What makes a chatbot yours and not generic is how your knowledge is loaded and how the limits are set. That’s where the careful work is: making it answer in your voice, know exactly how far it goes, and hand off well when it doesn’t know. It’s the same approach I use in Kyn and in Bayard Writer, the assistant I built for a school: build the business into it, rather than putting a layer of AI on top.

How long it takes to ship

A chatbot that answers FAQs about your business is usually ready in one to two weeks. If it also qualifies leads, books, and connects to your CRM through automations, two to four. What takes the most time isn’t the AI: it’s gathering and organising your information — prices, policies, answers — so the bot has something solid to answer from. The clearer you have that, the faster it ships.

A real example

The most honest example I can give you is the one you’re already looking at. Kyn is trained on everything KyndredAI: what I build, how I price, how I work. It answers the questions of whoever lands on the site, explains the services, and when someone wants to move forward seriously, it steers them toward a call with me. What it doesn’t know or what gets complex, it hands straight to my WhatsApp instead of inventing.

That’s what a well-built chatbot does for any business: it filters the noise, answers the repetitive stuff, and leaves only what needs a person for you. Try asking it a hard question and watch how it handles what it doesn’t know.

Frequently asked questions

Won’t it make up answers and embarrass me?

It’s the right fear, and the answer is in how it’s built. The bot is tied to your information and given a clear limit: if the answer isn’t in what it knows, it doesn’t invent it: it hands off to a human. Claude is also one of the models best at recognising when it doesn’t know something. It’s not infallible magic, but a well-built bot gets it wrong far less than a tired employee on a Friday afternoon.

Does it sound like a robot?

No, if it’s done right. The whole job is loading your brand’s voice: if your business is formal, it speaks formally; if it’s warm, it speaks warmly. Try Kyn and judge for yourself: it’s built with that same care.

Does it replace my support team?

It replaces the repetitive part, not the judgement. The bot handles the 80% of identical enquiries so your people can properly handle the 20% that matters: the delicate complaint, the big client, the odd case. In practice, your team stops firefighting and starts serving better.

Is it safe with my customers’ data?

Yes. Conversations are handled with the same care as everything else I do: data minimisation and no exposing sensitive information. You define what it can and can’t see.

Do I own the chatbot?

Yes. The knowledge is yours, the configuration stays documented, and you’re not locked into a platform that charges you forever. If you ever continue on your own, you take it all with you.

The fastest way to know if this is for you is to try it. Message Kyn right now, or if you’d rather talk to me, book a free twenty-minute call. What you see in Kyn is what I can build for your business.