How to Build a Recruitment Chatbot

Created by

Do not index

Created time

May 28, 2026 10:54 AM

Most recruitment chatbots only know how to ask. They march every applicant through the same scripted questions, years of experience, location, notice period, while the candidate is sitting there at 11pm on a Sunday with their own questions: What does the role actually pay? Is the team remote? What is the growth path? A flow-builder bot cannot answer any of that. It just keeps asking. So the candidate closes the tab, and a qualified applicant is gone.

A recruitment chatbot earns its place when it does both jobs at once: it answers the candidate's questions about the role, the team, and the benefits, and it pre-screens that candidate on the criteria a recruiter actually needs. This tutorial covers how to build one with SiteGPT, trained on your careers page and job descriptions, so it can hold a real conversation, qualify the applicant on experience, location, and work authorization, then route the strong ones straight to a recruiter. The honest framing: this chatbot does not replace the recruiter. It gives the recruiter their evenings back by making sure every candidate they speak to is already informed and already qualified. The closest existing parallel on this blog is the career guidance chatbot case study with The SaaS Jobs, where a SiteGPT bot mapped roles to candidate backgrounds in exactly this way.

TL;DR: Train SiteGPT on your careers page, job descriptions, and benefits docs, set a recruiter-style persona, write custom instructions that tell the bot to answer the candidate first and screen second, enable native lead collection for experience, location, work authorization, and notice period, then escalate qualified candidates to a recruiter and push their details into your ATS by webhook. Total setup time: 25-35 minutes, no flow builder or coding required.

What You'll Build

By the end of this tutorial, you will have a recruitment chatbot that:

Answers a candidate's real questions ("Is this role remote?", "What is the salary band?", "What does the interview process look like?") using your actual careers content, before it asks anything in return

Pre-screens each applicant on years of experience, location, work authorization, and notice period using built-in capture fields, no separate form tool required

Escalates qualified candidates to a recruiter and pushes their captured details into your ATS automatically, so recruiters only spend time on people who already fit

Time to complete: 25-35 minutes

Difficulty: Beginner

What you'll need: A SiteGPT account (free trial available), your careers page URL, and an ATS or email inbox where qualified candidates should land

This is the screening side of the same conversational model that powers a lead qualification chatbot, only here the goal is a shortlist of informed applicants instead of a pipeline of sales leads.

Why Recruitment Chatbots Work

A recruitment chatbot pays off on two fronts at once: it improves the candidate experience by answering questions instantly, and it cuts the recruiter's screening load by qualifying applicants before a human ever reads a resume. The order is what makes it work. Answer the candidate first, screen second, and applicants finish the conversation. Lead with the screening questions and they bail, the same way they abandon a long application form.

It Screens Applicants Without Adding Headcount

Recruitment chatbots are no longer a fringe experiment. 20% of the Fortune 500 now use a recruitment chatbot for hiring, according to SelectSoftwareReviews, and among them, 10% already use the bot to assess candidates with job-qualifying questions and 15% to answer the FAQs that otherwise pile up in a recruiter's inbox. A chatbot that does both runs a first-pass screen on every applicant, around the clock, without a recruiter staying late to do it.

It Answers Candidates at the Moment They Apply

People apply when they have a spare ten minutes, often late at night or over a weekend, exactly when no recruiter is online. A bot trained on the careers page answers "is this remote?" or "what is the comp band?" in the moment, while the candidate is still interested. That instant response keeps strong applicants from drifting to a competing posting that replied faster.

It Gives Recruiters a Pre-Screened Shortlist

Not every applicant fits the role. A chatbot that confirms work authorization, minimum experience, and location before booking anything means recruiters stop opening resumes that were never going to work. Fewer candidates reach the recruiter, but the ones who do are already informed and already screened, which raises the quality of every conversation that follows.

Why SiteGPT Works Well for Recruitment Screening

SiteGPT is best suited for teams that want a recruitment chatbot trained on their own careers content, with both the candidate's questions and the screening handled natively, without building a flow chart or wiring up separate form tools.

The defining feature for recruitment is that SiteGPT is trained on your careers page, job descriptions, and benefits documentation, so it can hold a genuine conversation before it screens. A linear flow builder can only ask what you scripted it to ask. A SiteGPT chatbot can field an unscripted question like "do you sponsor visas?" with an accurate answer pulled from your own content, then pivot to "are you authorized to work in the country this role is based in?", all in the same thread. This is the answer-the-candidate-and-ask-the-candidate model that separates a real conversational AI chatbot from a scripted question tree.

The second feature that matters is native candidate capture. SiteGPT collects years of experience, location, work authorization, notice period, and any custom field directly inside the chat, then escalates to a recruiter or pushes the details onward. Where a flow-builder tool like Landbot routes captured data through a separate integration, SiteGPT stores screened candidates natively and hands them off. If you are weighing scripted recruitment bots against trained chatbots, the Landbot alternatives breakdown covers that contrast in depth.

The setup takes well under an hour and requires no technical background. As one verified G2 user put it:

"SiteGPT makes it easy & intuitive to get your chatbot setup & working in no time at all - anyone can do it." Verified User, G2

Step-by-Step: How to Build a Recruitment Chatbot with SiteGPT

The full setup takes 25 to 35 minutes and requires no coding or flow building. Each step below includes a time estimate and a screenshot of exactly what to click.

Step 1: Create Your SiteGPT Account and Chatbot (~2 min)

Start by creating a free SiteGPT account and setting up a new chatbot. No credit card is required for the trial.

Go to SiteGPT and sign up for a free account

From the dashboard, click "Create Your Chatbot Now"

Name the chatbot something a candidate will recognize, for example "Acme Careers Assistant" or "Talk to Recruiting"

Set a welcome message that signals the bot can actually help, such as: "Hi! I can answer questions about our open roles, the team, pay, and benefits, and help you see if you're a fit. What would you like to know?"

Pro tip: A welcome message that names what the bot can answer ("roles, the team, pay, benefits") pre-frames the conversation as helpful, not as another application form. That framing is what earns the bot the right to ask screening questions later.

Step 2: Train It on Your Careers Page, Job Descriptions, and Benefits Docs (~5-10 min)

Training is what lets the chatbot answer a candidate's question before it asks its own. For recruitment, the priority content is everything an applicant reads while deciding whether to apply: the careers page, individual job descriptions, the benefits and perks documentation, and any team or culture blog posts. The same principle is laid out in this guide to chatbot training and what to feed your AI chatbot.

From the dashboard, click "+ Add Links"

Choose "Scrape Website" to pull the entire careers site automatically, or "Multiple Links" to add specific job postings

Prioritize the pages that answer candidate questions: the careers page, each open job description, the benefits page, and any "life at the company" or team blog posts

If your benefits summary, compensation philosophy, or interview-process overview lives in a PDF or an internal doc rather than on a public page, upload those files directly so the bot can answer accurately.

What to train for recruitment:

Careers page and every open job description (so the bot can answer "what does this role involve?")

Benefits, perks, and compensation philosophy docs (so it can answer "what's the pay band?" and "is there equity?")

Remote and work-location policy (so it can answer "is this role remote or hybrid?")

Team, culture, and "day in the life" blog posts (so it can build a real picture of the company before asking the candidate anything)

Step 3: Set a Recruiter-Style Persona (~2 min)

The persona sets the chatbot's tone. For recruitment, the bot should be warm and helpful, the way a good recruiter is on a first call, while still steering the conversation toward the criteria that determine fit.

In the left sidebar, go to Customizations > Chatbot Persona

Select a persona that fits a recruiting conversation. The "Problem Solver" persona works well here because it is oriented toward understanding a person's situation and guiding them to the right outcome

Confirm the description matches the tone you want a candidate to experience

A recruiting conversation is closer to helping someone find the right fit than to closing a sale, so a problem-solving, advisory tone lands better with candidates than a hard sales persona. The persona keeps the bot answering questions and gently advancing toward screening, rather than answering once and ending the thread.

Step 4: Write the Screening Instructions (~5 min)

Custom instructions are where a general Q&A bot becomes a recruitment chatbot. This is the step that encodes the answer-first, screen-second sequence, and it is the single most important configuration for this use case.

Go to Customizations > Chatbot Instructions

Click "Add Instruction"

Paste the instructions below and adapt the bracketed details to your roles

Screening instructions (copy-paste ready):

You are a recruiting assistant for [Company Name], helping candidates explore
[role or department] openings. Always answer the candidate's question first using
accurate information from the training content, the role, the team, pay range,
benefits, remote policy, and the interview process. Only after you have helped
them, ask one screening question at a time to understand fit: years of relevant
experience, current location and time zone, work authorization for [country or
region], and notice period or earliest start date. Keep it conversational, never
interrogate. If the candidate clearly meets the minimum bar (relevant experience,
authorized to work, reasonable start date), tell them they look like a strong fit
and offer to connect them with a recruiter or book a screening call. If they do
not meet the bar, be kind and direct, point them to other openings that may fit,
and offer to keep their details on file. Never ask for contact details before
providing value, and never make a hiring decision, only collect information and
flag fit.

The structure matters more than the exact wording. Answer first. Ask one question at a time. Screen on experience, location, work authorization, and notice period. Offer a recruiter handoff to strong fits and a polite alternative to everyone else.

Step 5: Configure Screening Questions and Candidate Capture (~6 min)

This is the core of a recruitment chatbot. SiteGPT collects screening data, years of experience, location, work authorization, notice period, and any custom field, directly inside the chat, with no separate form tool needed.

Go to Lead Collection in the left sidebar and toggle it "On". For recruitment, treat this as candidate capture

Set the timing to collect details after the first message, not before the conversation starts, so the bot answers a question before it asks for anything

Add the screening fields that define a qualified applicant for your roles: name, email, years of experience, current location, work authorization status, and notice period

For experience, authorization, and notice period, use dropdown or multiple-choice fields so candidates self-select cleanly (for example, "0-2 / 3-5 / 6-10 / 10+ years", "Authorized / Need sponsorship", and "Immediately / 2 weeks / 1 month / 2+ months")

You can start from a pre-built template and rename the fields to match a screening flow, which saves time over building each field from scratch

Once the fields are set, configure where screened candidates go.

Go to Notifications and add the recruiter email addresses that should receive new qualified-candidate alerts. Each alert includes the captured fields and the full conversation transcript, so the recruiter sees the experience level, location, authorization, notice period, and the actual questions the candidate asked

A note on sensitive candidate data: Recruitment screening collects personal information, so handle it with the same care as any regulated data. Ask only for what the role genuinely requires, avoid screening questions that touch protected characteristics, and review where the captured data is stored and who can see it. SiteGPT's guidance on secure enterprise chatbot deployment covers the access-control and data-handling practices worth applying here.

Pro tip: Set work authorization and minimum experience as required fields, and notice period as optional. Requiring authorization filters out applications that cannot proceed, without adding so much friction that strong candidates abandon. The goal of a screening bot is fewer, better-matched candidates, not maximum application volume.

Step 6: Escalate Strong Candidates and Route to Your ATS (~5 min)

Screening a candidate only pays off if a strong fit reaches a recruiter quickly and their details land in your ATS without manual copy-paste. This step closes that loop.

Enable Human Support Escalation so a qualified, high-intent candidate can be passed to a live recruiter mid-conversation. This is the "escalate to recruiter" handoff at the heart of the flow: the bot does the first-pass screen, then a human takes over for anyone who clears the bar

To let a strong candidate book a screening call on the spot, connect a booking integration so the bot can offer a recruiter's calendar once a candidate qualifies

Connect SiteGPT to your stack through the integrations directory: push qualified candidates into a Slack channel so recruiters see them in real time, and use a webhook or Zapier to map the captured fields, experience, location, authorization, notice period, into the matching properties in your ATS

For any ATS without a native connector, the webhook route lets you create a candidate record automatically the moment the bot finishes screening, so no application is retyped by hand

Pro tip: Send qualified candidates to a dedicated Slack channel as well as your ATS. A real-time ping with the candidate's experience and notice period lets a recruiter reach out while the applicant is still excited about the role, which is exactly when a strong candidate is most responsive.

Step 7: Customize the Appearance and Deploy (~4 min)

Matching the chatbot to your brand makes it feel native to the careers page, and deployment is a single snippet, no plugins required.

Go to Customizations > Chat Interface Colors and set the primary color to match your careers branding

Set the widget position (bottom-right is standard for a careers page)

Go to Installation, copy the JavaScript embed code, and paste it before the closing `</body>` tag of your careers page

SiteGPT supports WordPress, Shopify, Squarespace, Wix, Webflow, and any custom HTML site

Once deployed, open the careers page in an incognito window and run a test conversation. Ask a question about a role, confirm the bot answers from your content, then check that it asks the screening questions and offers the recruiter handoff only after you signal interest.

A Real Screening Flow: Engineering Role on a Careers Page

To make the answer-first, screen-second model concrete, here is how the conversation plays out for a candidate browsing an engineering opening:

Candidate asks a real question: "Is this role remote, and what's the salary band?" The bot answers accurately from the trained job description and benefits docs, then adds: "It's fully remote within the EU, and the band is published on the listing. Mind if I ask a couple of quick questions to see if it's a fit?"

Bot screens one step at a time: Based on the reply, it asks years of relevant experience, then current location and time zone, then work authorization, then notice period, never all at once, always after delivering something useful.

Bot routes by fit: A candidate with the right experience, valid authorization, and a reasonable start date is told they look like a strong fit and offered a recruiter's calendar on the spot. A candidate who needs sponsorship the company cannot offer is gently told so and pointed to roles that do qualify.

The recruiter gets a screened candidate: The recruiter receives a Slack ping and an ATS record with the captured experience, location, authorization, notice period, and the full transcript, before they have read a single resume.

The same pattern adapts to high-volume hiring. For retail or hospitality roles, swap the engineering criteria for availability, shift preference, and proximity to the location, then route qualified applicants straight to a hiring manager. For roles that screen international candidates, train the bot on content in more than one language so it can qualify applicants accurately regardless of where they are applying from; the rundown of multilingual AI chatbots covers how that works in practice.

Tips to Screen Better Candidates, Not Just More of Them

Deploying the chatbot is the starting point. These adjustments consistently improve the quality of the candidates that reach your recruiters.

Tip 1: Screen on the one criterion that disqualifies fastest, first

Work authorization and minimum experience are usually the hardest disqualifiers, so ask them early in the screening sequence, right after the bot has answered a question or two. There is no point collecting a full profile from someone the role can never hire. Front-loading the hard filter keeps the conversation short for candidates who were never going to fit.

Tip 2: Review conversation logs weekly to refine the questions

SiteGPT logs every conversation. Reviewing them weekly shows which questions candidates answer freely and where they drop off. If applicants abandon at the notice-period question, move it later or make it optional. If everyone asks the same thing the job description should already answer, that is a signal to update the listing, not just the bot.

Tip 3: Let the bot sell the role, not just screen for it

Candidates have options, so the bot should give as much as it asks. Train it on the genuinely attractive parts of the offer, the growth path, the team, the benefits, so that while it screens, it is also making the case for why a strong candidate should want this role. A screening conversation that leaves the candidate more excited about the company is doing two jobs at once.

Tip 4: Place the bot on individual job postings, not just the careers homepage

A candidate reading a specific job description is the most engaged they will ever be. Deploying the recruitment chatbot directly on each posting, where the applicant is already evaluating that exact role, concentrates the screening where intent is highest and produces the most qualified handoffs.

Frequently Asked Questions

What is a recruitment chatbot?

A recruitment chatbot is an AI chatbot that talks to job applicants on a careers page or job posting. The best ones do two things: they answer the candidate's questions about the role, team, pay, and benefits, and they pre-screen the applicant on criteria like experience, location, and work authorization. A bot built with SiteGPT is trained on your own careers content, so it answers accurately first, then collects screening details, and hands qualified candidates to a recruiter.

How does a recruitment chatbot screen candidates?

It screens by asking targeted questions inside the conversation, then capturing the answers as structured fields. With SiteGPT, custom instructions tell the bot to ask about years of experience, current location, work authorization, and notice period one at a time, after it has already answered something for the candidate. The captured answers are scored against your minimum bar, and applicants who clear it are flagged and routed to a recruiter, while the rest get a helpful, honest response.

Will a recruitment chatbot replace recruiters?

No. A recruitment chatbot handles the repetitive first pass, answering FAQs and running an initial screen, so recruiters spend their time on candidates who already fit. The chatbot does not make hiring decisions or judge a candidate's quality beyond clear factual criteria like authorization and minimum experience. The honest framing is that it gives recruiters their evenings back by ensuring every applicant they speak to is already informed and already qualified.

How is a SiteGPT recruitment bot different from a Landbot flow?

Landbot builds linear question trees: the bot can only ask what you scripted, in the order you scripted it, and it cannot answer an unexpected candidate question. SiteGPT is trained on your careers page and job descriptions, so it can answer an unscripted question like "do you sponsor visas?" and then screen in the same conversation. There is no flow chart to build, and candidate capture is native, so there is no separate form tool to wire up.

Can a recruitment chatbot answer candidate questions about the role?

Yes, and that is the point of building it on trained content rather than a flow chart. When SiteGPT is trained on the job description, benefits docs, and remote policy, it can answer "what's the pay band?", "is this hybrid?", or "what's the interview process?" accurately, at any hour. Answering the candidate first is what earns the conversation the right to ask screening questions, which is why response and completion rates are higher than a bot that only asks.

What questions should a recruitment chatbot ask to pre-screen applicants?

The standard set is years of relevant experience, current location and time zone, work authorization for the role's country, and notice period or earliest start date. Ask them one at a time, in plain language, after the bot has answered something useful, so it feels like a helpful first call rather than an interrogation. Avoid any question that touches a protected characteristic, and keep the screen focused on clear, role-relevant facts.

How does the chatbot hand qualified candidates to a recruiter?

SiteGPT has built-in human support escalation, so a qualified candidate can be passed to a live recruiter mid-conversation. It also sends an email notification with the captured fields and full transcript the moment a candidate clears screening. From the integrations directory, you can push qualified candidates into a Slack channel and use a webhook or Zapier to create a record in your ATS automatically.

Can a recruitment chatbot connect to my ATS?

Yes. Captured candidate data, experience, location, authorization, notice period, can be pushed into your applicant tracking system through a webhook or Zapier, mapping each field to the matching property in the ATS. That means a screened candidate becomes a record in your hiring pipeline without anyone retyping the details. For tools with a native connector, the integrations directory lists the available options.

Is candidate data collected by the chatbot secure?

Candidate screening collects personal information, so it should be handled like any sensitive data. Ask only for what the role requires, avoid questions about protected characteristics, and control who can see captured records. SiteGPT's secure enterprise chatbot deployment guidance covers access controls and data handling; enterprise plans add options such as HIPAA and BAA eligibility for teams with stricter compliance needs.

Does a recruitment chatbot work for international candidates?

Yes. SiteGPT detects the candidate's language and responds accordingly, based on the content it was trained on. Teams hiring across regions can train the bot on careers content in multiple languages so it answers and screens accurately regardless of where the applicant is searching from. The breakdown of multilingual AI chatbots explains how multilingual handling works in detail.

Conclusion

A SiteGPT recruitment chatbot, trained on your careers page, job descriptions, and benefits docs, set to an advisory persona, with answer-first instructions and native candidate capture, answers every applicant's questions and pre-screens them on experience, location, and work authorization, then escalates the strong ones to a recruiter and pushes their details into your ATS automatically. It does not replace the recruiter. It makes sure every candidate the recruiter speaks to is already informed and already qualified.

SiteGPT starts at $39/month for a single chatbot with 4,000 messages per month. The highest-return starting point is training the bot on your busiest open job description and its benefits page first, then deploying it on that exact posting, where candidates are already evaluating the role and ready to be screened.

Get Started with SiteGPT Free →

Last updated: May 2026. All SiteGPT features and pricing verified as of May 2026.