How do you build an AI agent without code?
You create a bot in Pachca in five minutes, and an n8n workflow does its thinking: the bot’s events go to a Webhook node, an AI Agent looks for the answer in your knowledge base, and the Pachca node replies in the question’s thread. Employees just @mention Kai in a chat. Walk through the question’s path yourself below.
Chatbots and webhooks
Creator
Daniel Reed (you)Bot type
For multiple chatsWho can edit bot settings
Who can add the bot to chats
You can let people find this bot in global search and message it directly
You create the bot in Automations → Integrations → Webhook → +. Check out the Outgoing webhook and API tabs, then click Save.
Why put an AI agent in your team chat
A new hire asks in the team chat where the policy is and waits for someone to free up. The answer comes an hour later, the link gets buried in the conversation, and a week later the next new hire asks the same thing.
- The answer lands where the question was asked. No copying the conversation into a separate AI chat and relaying the result to your teammates. An employee @mentions the agent in a chat, and the answer shows up in that message’s thread.
- Answers in seconds, not whenever a teammate is free. A new hire doesn’t have to wait for HR to get back to their laptop. The agent finds the document in the knowledge base and links to the source, so the answer is easy to check.
- Teammates stop getting the same questions. Policies, access, business trips and time off come up week after week. The agent handles them, and people answer what the docs don’t cover.
- The answer stays in the chat and shows up in search. Unlike a DM conversation, the thread with the question, answer and link doesn’t disappear. The next person with the same question will find it on their own.
- You see what’s missing from the knowledge base. Questions the agent couldn’t answer stay in the chat. That’s a ready-made list of what to add to your policies.
- Grows with you. Today it answers policy questions. Tomorrow it’s a “Create request” button, tracker tasks from a discussion, a customer brief before a call and a summary of a long thread. The bot in Pachca stays the same; only the handler changes.
How to set it up
About an hour, no developer needed. You'll need administrator rights in Pachca and n8n, either self-hosted or the cloud version.
In Pachca
- 1
Create a bot
Automations Integrations Webhook +. Give the bot a name, like "Kai," add a logo and turn on Public bot so employees can find it in search. - 2
Copy the token and add the bot to a chat
On the API tab, copy the token. You'll need it in n8n. Then open the chat settings Add members the Integrations tab. Only an admin or the chat's creator can do this.
In n8n
Four nodes: bot events from Pachca, an @mention filter, an AI Agent and a reply in the thread.
- 3
Install the Pachca node
Settings Community nodes the n8n-nodes-pachca package. Create a Pachca credential: put the bot token in Access token, choose Bot Token under Token type, enter the bot's User ID and set Base URL to:https://api.pachca.com/api/shared/v1 - 4
Build the workflow
Webhook Filter "content contains @Kai" AI Agent with a model, a "Knowledge base" tool and the prompt "answer from company documents and cite the source" the Pachca node: a message in the question's thread using the bot's credential. Turn on the workflow and copy the Webhook node's Production URL. - 5
Connect the bot to the workflow
In Pachca, open the bot settings and go to the Outgoing webhook tab. Paste the Production URL into Webhook URL and check New messages and Any messages. Ask the bot something in the chat with an @mention, and the answer will arrive in a thread.
Other options
If n8n isn't a fit, you can build the handler another way. The bot in Pachca stays the same.
- A
Claude Code or Cursor with Pachca skills
Get a token in Automations API and run three commands:
Then describe the handler in plain words: "Kai answers from the knowledge base in the General chat, in a thread, with a link to the source. Deploy it and give me the URL for the outgoing webhook." Paste that URL in step 5.npm install -g @pachca/cli echo "YOUR_TOKEN" | pachca auth login --token - npx skills add pachca/openapi - B
Your own service, built from the docs
Give this skill to a developer or their AI assistant. It covers receiving Pachca events with signature verification, replying in a thread, buttons, the knowledge base and choosing hosting in Russia.
In your team
- 6
Agree on how to call the agent
Ask the agent with an @mention in the chat and follow up in its thread, so the chat stays clean. If the agent can't find an answer, the topic owner picks up the question and adds the missing doc to the knowledge base.
The "Create request" button from the demo and actions in other systems are built with the API. The skill above covers them. To set them up by hand, see bot buttons and outgoing webhooks. For a look at how the agent works under the hood, read our architecture breakdown.
Troubleshooting
- The agent doesn't reply: the bot isn't in the chat, "New messages" isn't checked in the outgoing webhook, or the question was asked without an @mention.
- You can't create a token: API access is turned off in the workspace settings. Ask an administrator to turn it on or to issue a token from a service account.
- It worked, then stopped: the n8n workflow is turned off, or the outgoing webhook has the Test URL instead of the Production URL.