Set up your AI Customer Support Agent
Publish your help center, configure the agent, test it, and start it answering, end to end
A customer-facing agent that answers from your published articles and passes the question to your team when it cannot.
The AI Customer Support Agent answers your customers' questions on your public Help Center and in the help widget on your website. It answers only from the knowledge base articles you have published to that Help Center, and it shows the customer which article the answer came from. When it cannot answer, or when the customer asks for a person, it passes the question to your team as a ticket. A person on your team then replies to the customer by email.
The agent has its own tab: Service, then AI Agent. Until the agent is live, that tab shows a checklist headed Set up your agent, with the line Three things have to be true before your agent answers a customer. Each row carries the one action that finishes it: Publish, Configure and Add agent. Publishing and adding the agent open a dialog on the tab. Configure opens Service, Configuration, AI agent, where you set how the agent answers. The checklist goes away once the agent is live.
The order matters for step 1 and it does not for step 2. The agent has nowhere to answer until your Help Center is live, so publish first. But you can configure the agent, save it and test it before you buy one, so the voice, the guardrails and the handoff message are already right on the day it starts answering.
- A few articles worth answering from. The agent is only as good as your knowledge base. Five or ten solid articles covering the questions you answer every week beat fifty thin ones.
- Permission to edit your knowledge base, the
kb.editkey. Publishing the Help Center, changing how the agent answers and testing the agent all need it. Without it the first checklist row states the fact instead of painting a control you cannot use: An administrator with knowledge base access can publish your Help Center. Anyone who administers roles in your workspace can grantkb.edit. - Someone who can buy, holding the
billing.viewkey. Adding the agent is a billing action, and the price is only readable by abilling.viewholder, so a reader without it sees An administrator with billing access can add it. rather than a path. A reader who holdsbilling.viewbut does not manage the subscription sees its own line, An administrator who manages your subscription can add it. Both keys are granted by a workspace administrator under Roles & Permissions, and the buyer can be a different person from the one doing steps 1 and 2.
The public surface the agent answers on, and the article set it answers from.
Your Help Center is the branded, customer-facing view of your knowledge base, and it is the page your agent answers on. It is off by default, so nothing is public until you publish it, and you publish it without leaving the AI Agent tab. If your Help Center is already published, the first row reads Done: skim the article gates below and move on to step 2.
Press Publish on the first checklist row
Go to Service, then AI Agent. The first row of the Set up your agent checklist is Publish your Help Center, and its Publish action opens a dialog headed Publish your help center on the same page. If your plan is below Business the row reads Publishing needs the Business plan or higher., and if you cannot edit the knowledge base it names who can.
Choose your public address
The dialog has one field, Public address, sitting under the /help-center/ prefix. It is how the public portal resolves your organization, so choose something your customers will recognize. Use 3 to 40 lowercase letters, numbers, and single hyphens; a line under the field shows the finished address as you type, and a short list of names is reserved. Only a workspace admin can set it. If you are not one, the dialog reads Only a workspace admin can set the public address. Ask an admin to set one, then publish from here. See Your Help Center Address.
Publish it
Press Publish. Your public address starts resolving straight away, with whatever articles already clear the three gates below, and the checklist row changes to Done. If the address is taken or malformed the dialog stays open and says which, so you can fix it in place.
Add your branding when you want it
That dialog publishes and does nothing else. Brand color, Logo URL, Default language and the rest of the public page live in Service, Configuration, Help Center, and so does the switch that takes the Help Center back off. See Your Public Help Center for that page.
Get the right articles onto it
In Service → Knowledge, make sure the articles you want the agent to use are published, set to visibility: public, and verified. See Your Public Help Center and Knowledge Base Article Template & Authoring for the full walkthrough of each gate.
Your first agent needs one article it can use. A live agent keeps running.
The Where customers reach your agent card on the AI Agent tab counts the articles the agent can use, for example 4 articles the agent can use. Each must be published, public and verified., with an Open Knowledge link. Your first agent needs at least one. Until then the Add agent action is greyed out and the page tells you in one sentence: Publish 1 article first. The agent answers only from published, public articles. An Open Knowledge link beside it takes you to your articles.
Handling, sources, tone, instructions, guardrails and handoff, on Configuration > AI agent.
Everything about how the agent answers is set in Service, Configuration, AI agent. The checklist's Configure action opens it, and so does the Configure link beside Learn more on the AI Agent tab. The sub-tab opens for people who can edit the knowledge base; anyone else reads You need permission to edit the knowledge base before you can configure the AI agent. An administrator in your workspace can grant it. Changing how the agent answers needs the Business plan or higher. Below that the form is readable but not editable, and the page says Changing how the agent answers needs the Business plan or higher. above it. Every saved change applies to the next conversation, so you can keep adjusting it while it is live. The defaults are deliberately cautious, so an agent you configure lightly hands off rather than guessing.
What you can set
- Agent behavior. Its Handling choice is how far the agent goes on its own: Send everything to my team (every question becomes a ticket and the agent does not answer), Answer first, then hand off, or Let the agent handle it. With the last one the agent never offers a person; when it cannot answer, it points the customer to your articles and to your request form (see What customers see).
- Sources. Under Articles the agent may use, choose which of your public, published, verified articles the agent may answer from. Every eligible article is in by default; turn individual ones off when an article is accurate but not something you want quoted back conversationally.
- Identity and tone. The Tone it writes in, the Answer length (Short, a sentence or two, Medium, a short paragraph or Detailed, the full steps), and the Language it replies in. Leave the language blank to reply in the language the customer used.
- Instructions. Under What the agent should know, free text describing your product, your support hours, and anything you want the agent to mention or avoid.
- Guardrails. Switches for Never quote prices and Never promise refunds, plus Your own phrases to withhold. An Always on set sits underneath them and cannot be turned off.
- Handoff. The Handoff message is what the customer reads when the conversation moves to your team. When you leave it empty, the agent uses the built-in lines in What customers see below.
Nothing is stored until you press Save changes at the foot of the form. If you try to leave the page with unsaved changes, a dialog headed Discard unsaved changes? asks first, with Discard and leave and Keep editing.
A private chat that answers the way the live agent does, drafts included.
The panel headed Test your agent is in two places: the Test sub-tab of the AI Agent tab, beside Transcripts, and on Configuration, AI agent, beside the form on a wide screen and under it on a narrow one. On Configuration it reads Save your changes, then ask a question to see how the agent answers. It runs the same answer steps, guardrails and hand-off rules as the live agent, on your public articles, and adds It also reads your public drafts, so you can test an article before you publish it. An article that is not published shows in the sources as Draft: and its title. Team-only articles are never used, because the live agent never uses them either.
- Who can test. People who can edit the knowledge base, on the Ultra plan. Below Ultra the panel reads Testing the agent needs the Ultra plan.
- Before you add the agent. The panel reads Before you add the agent, you can run 10 test conversations a day at no cost. After the tenth it reads You have used today's 10 test conversations. Add the agent to keep testing, or try again tomorrow.
- After you add the agent. The panel reads Each test conversation uses 1 of your agent's conversations. When no article matches the question, the conversation is given back.
- What only you see. Under a reply that a customer would not get answered, a line says why, for example No public article matched this question., and A customer would be offered a person here.
- Private. When the agent is already live, the panel reads Customers can already ask this agent. Test conversations stay private. A test conversation never shows in Transcripts or on the Overview. A conversation holds up to 12 answers; press New conversation to start again.
With the Help Center published and the agent configured, the last step is to add the agent. The third checklist row carries an Add agent action that opens a dialog headed Add your AI Customer Support Agent, with the price, what is included, and your trial if you still have one. There is no separate go-live switch: the agent starts answering on your Help Center and in your help widget as soon as it is added. Your first agent needs one article it can use, so until one is published, public and verified the action is greyed out (see What the page tells you about your articles). Without billing access the row names who in your workspace can do it instead, and on a plan below Ultra it reads The agent is part of the Ultra plan.
Try it for 14 days
While your organization still has a trial, the dialog offers it first. Start it for a one-time $1.99 deposit, charged the day you start. The agent runs for 14 days, and the deposit is credited to your first invoice, so the trial costs you nothing extra if you keep it. The trial is limited to one per organization. This deposit is separate from the deposit on a Laureo plan trial, and paying one does not pay the other.
Or add an agent outright
Add an agent for $59.99 per month, each. An annual price is available, and the dialog quotes whichever your organization is billed on. You can run more than one agent, each billed separately. See AI Customer Support Agent Add-on for the full price list.
With no card on file, you finish on the billing page
If your organization has no payment method yet, the dialog says what happens before it moves you: you will add a payment method next, then land back on billing. It takes a second, deliberate press on Continue to payment to leave, and that hand off is the one part of setup that does not finish on the page.
Know what is included
Every agent includes 500 conversations a month. The allowance resets at the start of each month and unused conversations do not carry over. A conversation is counted the first time the agent answers, so a customer browsing your Help Center, a question the agent hands straight off, or a customer who asks for a person never counts.
Add conversation credits if you need more
If a month runs long, buy an extra conversation credit pack. Purchased credits do not reset with the month and are drawn only after the month's included allowance is used. If you run out entirely, the agent stops answering and keeps your articles in view, and it still offers Talk to a person, so customers can still find answers and still reach your team.
The card headed Where customers reach your agent sits on the AI Agent tab above the Transcripts and Test sub-tabs, and it stays there after setup is done. Its rows, in order:
- Link to your AI agent. Your Help Center address with
?ask=1at the end, for examplehttps://app.laureo.io/help-center/your-business?ask=1, with Copy, Open and QR code. Under it: Share it in emails, chats or on your site. It opens your Help Center with the question box ready. The link puts the cursor in the question box; it never types or sends a question. QR code offers Download PNG and Download SVG, for a poster, a receipt or a slide. - Help Center. Your full public address, with Copy and Open.
- Help button on your website. The code to paste, with Copy, and the line Paste this before the closing </body> tag on each page. A Help button appears, and customers ask the agent there. Under it, the card shows your Allowed websites setting, for example Any website can embed your widgets., with an Open Channels link for people who can open Configuration, Channels.
- Live chat and email go to your team. The agent answers on your Help Center and in the help widget.
- The article count, for example 4 articles the agent can use. Each must be published, public and verified., with Open Knowledge.
Copy confirms with Link copied or Code copied. If your browser refuses, it reads Could not copy. Select the link and copy it.
Under the rows, How it works gives the customer's side in three lines: A customer opens your Help Center, or clicks Help on your website. They ask our AI agent a question in their own words. The agent answers from your articles, or passes the question to your team, who reply by email. With the Let the agent handle it handling choice, the last line reads The agent answers from your articles.
The question box, the answer, and the path to a person, as your customers read them.
Your Help Center page and the help widget on your website show the same box and the same lines. Where a line names your team, it uses your Help Center's name, for example the Acme Help Center team, or our team when there is no name. Where it says when the team replies, it reads Team hours: and your hours when you set Live chat hours under Configuration, Help Center. Otherwise it reads They reply as soon as they can. The line never promises a time you did not set.
Asking
- The box is labelled Ask our AI agent, with the placeholder Type your question in your own words and, under it, It answers from this help center's articles. You can ask for a person at any time. Before the first question, a Talk to a person link sits under that line.
- The article search sits below it, labelled Search the articles, with the placeholder Or search the articles. It matches every word of the search, in any order. When nothing matches it reads No articles match and the search, then offers Ask this instead (sends the typed words to the agent), Submit a request when your Help Center takes requests, and Clear.
- While a security check runs, a pressed Ask waits with Checking you are human… and sends the moment the check passes. While the agent looks for an answer: Searching the help center…
- If the check cannot load: The security check did not load. Submit a request to reach the team. Too many questions at once: Too many questions in a short time. Wait a minute, then ask again. When answers are off for a moment: Instant answers aren't available right now. Browse the articles below, or submit a request to reach the team.
The reply
- An answer shows its Sources, links to the articles it came from, then Still need help? Talk to a person, then the line Generated by Laureo AI from this help center’s verified articles. Always double-check important details. The help widget shows the first sentence of that line only.
- An answer the agent is not sure of opens with I'm not sure this fully answers your question. and offers Talk to a person under it.
- No answer: I couldn't find this in our help articles. A person on the Acme Help Center team can help. They reply by email. and the reply-time sentence, with Talk to a person. Your own Handoff message replaces this line when you set one.
- The agent cannot answer right now (for example, it has used its conversations): I can't answer this automatically right now. A person on the Acme Help Center team can help. They reply by email. and the reply-time sentence. When the conversations are used up it adds You can also browse the articles below. Your Handoff message replaces the first line and the reply-time sentence here too.
- The customer types a request for a person, such as can I talk to a person: A person on the Acme Help Center team can help. They reply by email. and the reply-time sentence, and the contact step opens. That turn never counts as a conversation.
Reaching a person
- The contact step is headed Talk to a person, with A person on the Acme Help Center team will reply to you by email., a required Your email field, Name (optional), and Send to the team. A blank email reads Enter your email so the team can reply to you. and a malformed one Enter a valid email address.
- While it sends: Sending your question to the team… Then it confirms in three lines: Sent to the Acme Help Center team. They will reply to the typed address by email., then Reference and the request number with We sent a copy to the same address, then the reply-time sentence. Asking again after that reads Your question is already with the team. They will reply by email.
- If it does not go through: Your question did not reach the team. Try again, or submit a request. Try again resends the same request with the same email.
- The customer also gets an email headed We received your request: Your question has reached the Acme Help Center team. A person will reply to this address by email. When replies to it reach the ticket, it ends with You can simply reply to this email to add more detail to your request. Otherwise it ends with Replies to this email are not received. The team will write to you at this address.
- The request lands in your Inbox as a ticket with the whole conversation and the customer's email. Your public reply goes to that email.
Read the line at the top of the AI Agent tab
The line there says whether the agent is live, what customers have done with it in the last 7 days, and what is missing if it is not live. It never says the agent is answering unless the numbers show it. When everything is done and there are no numbers to show, it reads Live on your Help Center. with a View your help center link beside it that opens your public page in a new tab. If no customer has asked anything yet, it reads Live on your Help Center. No customer has asked yet. Beside it, Open your agent link opens the agent in a new tab, so you can ask the first question yourself.
Once customers ask, it reads Live. Last customer question and how long ago that was. A second line under it gives the week, for example Last 7 days: 12 conversations. 9 answered, 3 passed to your team. If nobody asked in that week, the second line reads No customer has asked the agent anything in the last 7 days. with an Open your Help Center link. These are the same numbers the Overview shows, and you see them only if you can view tickets.
If the agent has articles it can use, there were at least 3 conversations in the last 7 days and it answered none of them, the line reads Not answering: every recent question went to your team. in red, with Test and Transcripts links. Test opens the chat where you ask the agent a question yourself, and Transcripts shows the conversations it passed to your team.
If you chose Send everything to my team, it reads Every question goes to your team. The agent is not answering. instead, which is that setting working, not a fault. If you can view the knowledge base and the agent has no article it can use, the line reads Not answering: 0 articles the agent can use. in red, with an Open Knowledge link: the agent hands every question to your team until one article is published, public and verified.
Open your agent link as a customer would
Open the Link to your AI agent in a private browser window, signed out. You should see your branding, your public articles, and the box labelled Ask our AI agent with the cursor in it. A question asked there is a real conversation: it counts against the agent's allowance and shows in Transcripts. For a private check, use the Test sub-tab.
Ask something you know an article answers
A good first test is a question whose answer sits in one specific published article. You are checking two things: that the agent answers, and that its Sources name the article you expected.
Ask something it should refuse
Then ask about pricing, or for a refund, with those guardrails on. The agent should not answer, and should offer Talk to a person with your handoff message. A handoff here is the system working, not failing.
A separate switch for how new customer tickets are sorted, not part of the agent.
The Ticket triage switch is not a setting of the customer-facing agent, and the agent does not need to be set up for it. It lives in Service, Configuration, Autonomy, and its switch reads Sort new tickets and draft a first reply. On the Ultra plan, the Service AI sorts each new customer ticket by priority and intent and leaves a first reply draft for your team; it never sends, resolves or assigns anything. See AI Autonomy Controls for the switch, and Tickets Sorted on Arrival for what it sets on a ticket.
- AI Customer Support Agent: Ultra plan, as a paid add-on bought per agent. Ultra is what turns the agent on.
- Testing the agent: Ultra, with 10 test conversations a day at no cost before you add an agent.
- Public Help Center and help widget: Business and above.
- Changing how the agent answers: Business and above, so steps 1 and 2 can be finished before the plan that runs step 3.
- Knowledge base authoring and verification: included with the ticketing suite.