Most HR teams spend 23 hours screening candidates for a single hire, according to SHRM. A chatbot does not eliminate that work, but it shifts most of it off human calendars. The question is whether you buy one off the shelf, build one, or layer AI tools onto what you already have.
The cost spread is wide: $300 a month for a basic SaaS product, $30,000 or more to build something tailored to your process. The right answer depends on two things: how complex your hiring workflows are, and whether your compliance requirements let you use a generic tool at all.
What HR tasks can a chatbot realistically handle?
The gap between what vendors promise and what chatbots actually do well is worth naming before you look at pricing.
Chatbots handle structured, repetitive tasks reliably. Answering candidate questions about benefits, salary bands, and application status, screening resumes against a fixed set of criteria, sending scheduling links, and sending rejection notices: these work well because the inputs and outputs are predictable. LinkedIn's 2023 Talent Trends report found 67% of recruiters said automation helped most with high-volume screening, specifically the part where the same ten questions get asked to a hundred candidates.
Chatbots do not handle judgment calls. Deciding whether a candidate's non-linear career path is an asset or a liability, reading the subtext in a resignation letter, or coaching a manager through a difficult performance conversation: those require a human. Vendors will tell you their tool "uses AI to assess fit." What that usually means is keyword matching dressed up with a confidence score.
The tasks worth automating, with clearest ROI:
- Initial candidate screening (qualifying questions, knockout filters)
- Interview scheduling and rescheduling
- Onboarding FAQ (benefits enrollment, IT setup, first-day logistics)
- Status update messages throughout the hiring pipeline
- Off-hours responses to inbound candidate inquiries
If your process only needs two or three of these, a SaaS product covers you. If you need all five tightly integrated with your ATS, payroll, and calendar systems, a custom build starts making economic sense.
How does a recruiting chatbot screen and schedule candidates?
The screening workflow is where chatbots save the most recruiter time, so it is worth understanding how it actually works before you pay for it.
When a candidate applies, the chatbot sends them a conversational questionnaire. It asks about availability, compensation expectations, required certifications, location, and any role-specific knockout questions. Candidates answer on their phone in a few minutes. The chatbot scores each response against your criteria and routes qualified candidates forward, sending unqualified ones a polite rejection notice.
HireVue's 2023 benchmarking data found companies using AI screening reduced time-to-screen from 5 days to under 24 hours. The mechanism is straightforward: instead of a recruiter reading 200 resumes on Monday morning, the chatbot has already sorted them into three buckets by Sunday night.
Scheduling works through calendar integration. The chatbot checks interviewer availability in real time, offers the candidate a set of open slots, and books the meeting without a human touching it. Every reschedule request goes back through the same loop. Greenhouse and Lever both have native integrations that make this work without custom code.
The limitation is configuration. A chatbot only screens for what you tell it to screen for. If your knockout criteria are not written down somewhere, the chatbot cannot apply them. This sounds obvious, but many companies discover mid-implementation that their screening logic exists in a senior recruiter's head rather than in a documented rubric. The chatbot forces you to write it down, which is a useful side effect.
Is it cheaper to buy an HR chatbot product or build one?
For most companies, buying wins on cost until you hit a specific inflection point.
The leading SaaS products, Paradox (Olivia), Eightfold, and Phenom, charge based on employee count or hiring volume. A company with 200–500 employees typically pays $800–$2,000 per month. The product is live in a few weeks, the vendor handles compliance updates, and you get a support team when something breaks.
Custom builds cost more upfront but have no ongoing licensing fee and can be shaped to your exact workflow. With an AI-native development team, a custom HR chatbot integrated with your existing ATS, calendar, and HRIS runs $15,000–$30,000 to build. A Western agency quotes $60,000–$90,000 for the same scope, because their overhead and billing model have not changed since 2023.
| Approach | Upfront Cost | Monthly Cost | Best For |
|---|---|---|---|
| SaaS product (e.g. Paradox, Phenom) | $0–$5,000 setup | $800–$2,000/mo | Companies under 500 employees, standard workflows |
| Custom build (AI-native team) | $15,000–$30,000 | $200–$400/mo hosting | Companies with complex or regulated workflows |
| Custom build (Western agency) | $60,000–$90,000 | $200–$400/mo hosting | Same scope, 3–4x higher build cost |
| Bolt-on AI to existing ATS | $0 upfront | $200–$600/mo | Companies already using Greenhouse, Lever, or Workday |
The break-even point between SaaS and custom is roughly 18–24 months. If you are paying $1,500 a month for a SaaS product and a custom build costs $25,000, you break even in about 17 months and save money every month after that. That math holds only if your workflows are stable enough that you do not need to rebuild the custom tool every year.
There is a third option many companies overlook: their existing ATS already has a chatbot feature they are not using. Greenhouse, Lever, and Workday all added AI-assisted screening and scheduling in 2023. Before buying a separate product or commissioning a build, check what your current vendor already includes.
What compliance concerns apply to AI in hiring?
This is where cost estimates can jump, and where ignoring the fine print gets expensive.
The EEOC issued guidance in May 2023 clarifying that employers are liable for discriminatory outcomes produced by AI hiring tools, even if they did not build the tool themselves. Using a vendor's chatbot does not transfer your legal responsibility. If the tool systematically screens out candidates based on a pattern that correlates with a protected class, that is your problem.
New York City went further. Local Law 144, which took effect in July 2023, requires employers using automated employment decision tools to conduct annual bias audits and publish the results. Chicago and California have similar rules moving through their legislative process. Any company hiring across multiple US jurisdictions needs to know which rules apply where.
Compliance affects build vs buy in two ways. For SaaS products, ask vendors directly whether their tool has been audited under EEOC guidance and whether they can provide the documentation. Reputable vendors have this. Smaller vendors often do not.
For custom builds, budget $3,000–$8,000 for an initial bias review from a firm that specializes in algorithmic auditing. The build cost table above does not include this, and it should. A chatbot that makes a binary hire/reject decision without human review carries more legal exposure than one that ranks and routes candidates for a recruiter to approve. The latter is easier to audit and easier to defend.
How do I measure whether the chatbot improves time-to-hire?
Buying or building without defining success metrics first is how chatbot projects turn into shelfware.
Time-to-hire is the headline metric, but it is a lagging indicator. By the time you see it improve, you have already run three or four hiring cycles. The leading indicators to track from day one:
- Screening completion rate: what percentage of candidates sent the initial questionnaire actually finish it? Under 60% suggests the questionnaire is too long or poorly worded.
- Time from application to first human touchpoint: the chatbot's job is to shrink this number. Benchmark it before you go live.
- Recruiter hours per hire: SHRM's 2023 data puts the average at 23 hours. If your chatbot brings it to 14 hours, that is nine hours back per hire per open role.
- Offer acceptance rate: automating the wrong parts of the process can make candidates feel like they are talking to a wall. If acceptance rate drops after deployment, the screening or scheduling experience is likely the reason.
Set a 90-day checkpoint. If screening completion rate is above 65%, time to first human touchpoint is under 48 hours, and recruiter hours per hire has dropped by at least 25%, the tool is working. If two of those three are off, diagnose before renewing the contract or expanding the build.
The companies that get the most out of these tools treat the first three months as a tuning period. Revisit knockout criteria every quarter. Update the onboarding FAQ when your benefits change. A chatbot accurate in February will drift out of date by July if nobody maintains it.
