Table of Contents
- Key takeaways
- What is a 24/7 customer support chatbot?
- How much does 24/7 human coverage cost compared with a chatbot?
- What should a 24/7 chatbot answer after hours?
- How to set up 24/7 customer support without hiring a night shift
- What happens when a customer asks for a person and nobody is online?
- How to choose your after-hours model
- Tools for a 24/7 customer support chatbot
- 1. SiteGPT: best for answering from your own content around the clock
- 2. Tidio: best for small stores that want live chat and an AI agent in one inbox
- 3. Fin: best for helpdesk teams that pay per resolved conversation
- How the three compare for after-hours coverage
- Common 24/7 chatbot mistakes
- Frequently asked questions
- Basics
- What is a 24/7 customer support chatbot?
- Can a chatbot replace a night shift?
- How do I set up 24/7 customer support without hiring an in-house team?
- Cost
- How much does 24/7 customer service cost?
- Is a per-resolution price cheaper than a flat plan at night?
- Handoff and hours
- What does a customer see when they ask for a person at night?
- How does the team know a conversation needs a person?
- Should the bot answer or stay silent outside working hours?
- Channels
- Does a 24/7 chatbot work on WhatsApp?
- Can a chatbot answer customer emails at night?
- Sources

- Covering one chat seat around the clock takes 4.2 full-time people (168 hours a week divided by 40). At the 2025 US median wage for customer service representatives, that is about $188,000 a year in wages alone.
- A chatbot trained on your help pages, policies, and product docs answers the repeat questions at night. It does not replace the person for account changes, judgment calls, or urgent problems.
- Set your team's working hours in the bot. Outside them, the bot keeps answering, sends an away message on handoff, and collects contact details so the morning shift can reply.
- Page an on-call person only for the few cases that cannot wait, such as outages, safety issues, or money at risk.
- SiteGPT Growth costs $129 a month billed monthly, or $79 a month billed yearly ($948 a year), with 10,000 GPT-4.1-mini messages a month.
Key takeaways
Question | Short answer |
What does a 24/7 chatbot cover? | Questions your content already answers: policies, how-to steps, order and shipping information |
What does it not cover? | Account changes, refunds outside policy, custom quotes, and anything urgent that needs a person |
What happens when nobody is online? | The bot sends an away message, collects contact details, and queues the transcript for your team |
What does human 24/7 cover cost? | About $188,000 a year per seat in median wages (4.2 people), before benefits and overtime |
What does the bot layer cost? | SiteGPT from $59 a month billed monthly, or $39 a month billed yearly |

What is a 24/7 customer support chatbot?
How much does 24/7 human coverage cost compared with a chatbot?
- A week has 168 hours. One full-time person works about 40.
- 168 divided by 40 is 4.2 people for one seat that is always staffed. Holidays, sick days, and training push the real number higher.
- The US Bureau of Labor Statistics puts the 2025 median pay for customer service representatives at $44,770 a year, or $21.53 an hour.
- 4.2 people at $44,770 is $188,034 a year. That is wages only, before benefits, night premiums, and management.
Coverage model | People needed per seat | Yearly cost | What you get |
Human seat, 24/7 | 4.2 | $188,034 in median wages | A person answers every hour |
Human seat, nights and weekends only | 3.2 | $143,264 in median wages | A person answers outside office hours |
SiteGPT Growth plus day team | 0 extra | $948 billed yearly, or $1,548 billed monthly | The bot answers from your content, people reply in the morning |
SiteGPT Scale plus day team | 0 extra | $3,108 billed yearly, or $5,148 billed monthly | Same, with 40,000 messages a month and webhooks for on-call alerts |

What should a 24/7 chatbot answer after hours?
Question type | Bot answers overnight? | Why |
Opening hours, shipping times, return policy | Yes | One answer, already on a page |
How-to steps and setup questions | Yes | Help center articles cover them |
Pricing and plan differences | Yes | The pricing page is the source, keep it synced |
Order status or account details | Only with a connected system | A bot trained on pages cannot see one customer's order |
Refunds outside policy, discounts, exceptions | No, hand off | A person has to decide |
Outages, safety, fraud, money at risk | No, page on-call | Waiting until morning causes harm |
How to set up 24/7 customer support without hiring a night shift
- Pick the after-hours questions. Export a month of chats or tickets that arrived outside office hours. Mark each one as answerable from your content, needs a person, or urgent.
- Train the bot on the pages that answer them. Add your help center, policy pages, and product docs. Fix any page that contradicts another before launch.
- Turn on human support. In SiteGPT this is under Leads, then Human Support. It works on any plan.
- Set your working hours. Under Settings, then Chat Modes, choose Follow Working Hours, set your time zone and hours, and choose "The AI answers" for outside those hours.
- Write the away message. Say when the team will reply, for example "The team is offline now and replies from 9 a.m. Eastern, Monday to Friday."
- Write escalation rules. By default the AI escalates when the visitor asks for a person, the situation is urgent, or the visitor is frustrated. Add your own urgent cases in plain language.
- Route the alerts. SiteGPT emails the chatbot owner for each escalation. Growth and above add an iOS app with push notifications when a visitor asks for a person. Scale adds webhooks for new escalations.
- Test at night. Start a conversation outside your hours, ask for a person, and check that the away message, the contact form, and the email alert all fire.

What happens when a customer asks for a person and nobody is online?
- The visitor selects the button that asks for a person, or the AI decides to escalate.
- The widget asks the visitor to confirm, then shows your confirmation message.
- The chat asks for contact details if lead collection is on, so your team can follow up by email.
- The AI stops answering, and the conversation waits in Chat History with an Escalated badge.
Setting outside working hours | What the visitor gets | Use it when |
The AI answers | Answers all night, away message and contact form on handoff | Most questions are in your content |
Your team answers later | An away message only, the AI does not reply | Nearly every question needs a person |
Let the visitor choose | A question at the start: chat with the AI or leave a message for a person | Visitors who want a person should not have to talk to the AI first |
How to choose your after-hours model
If you need... | Pick... | Why |
Answers to repeat questions at night, replies to the rest next morning | Chatbot plus working-hours handoff | Lowest cost, and no customer is left without a reply time |
A person within minutes for a few urgent cases | Chatbot plus one on-call person | Escalation rules wake someone only for those cases |
A person in chat at every hour | Staffed shifts or an outsourced team, with the bot as first reply | Required when most night questions need a person |
Daytime cover in several regions | People in a second time zone, plus the bot | Each region gets a person during its own day |

Tools for a 24/7 customer support chatbot
- Answers come from your own content.
- The tool has working hours or offline handling for handoff.
- The cost at night volume is predictable.
- The tool reaches the channels your customers use.
1. SiteGPT: best for answering from your own content around the clock

- Training sources include URLs, sitemaps, files, Google Drive, Dropbox, OneDrive, Box, Notion, GitHub, and YouTube
- Follow Working Hours mode with an away message and per-day schedules in your time zone
- Escalation on request, plus AI escalation rules you write in plain language
- Email alerts on every plan, iOS app push alerts on Growth and above, webhooks on Scale and above
- Chat integrations on Growth and above: Slack, Crisp, Messenger, Google Chat, Freshdesk Messaging, Zoho SalesIQ, with Zendesk on Scale and above
- One flat plan price per month, with no per-resolution fee
- Working hours, away message, and contact capture are built into the human support settings
- Auto refresh keeps answers current: monthly on Growth, weekly on Scale
- No free plan. The 7-day trial needs a credit card
- Starter has no chat integrations and only manual content refresh
- No Intercom or HubSpot integration today
- Message quotas cap volume: Starter includes 4,000 GPT-4.1-mini messages a month
Plan | Billed monthly | Billed yearly | Messages a month |
Starter | $59 a month | $39 a month ($468 a year) | 4,000 GPT-4.1-mini or 400 GPT-4.1 |
Growth | $129 a month | $79 a month ($948 a year) | 10,000 GPT-4.1-mini or 1,000 GPT-4.1 |
Scale | $429 a month | $259 a month ($3,108 a year) | 40,000 GPT-4.1-mini or 4,000 GPT-4.1 |
Enterprise | Custom | Custom | Custom |
2. Tidio: best for small stores that want live chat and an AI agent in one inbox

- Lyro AI agent that answers from FAQs and scraped website content
- Operating hours and an offline auto-reply
- Live chat, ticketing, and Flows automation in one inbox
- Messenger, Instagram, WhatsApp, and email channels
- A free plan exists: 50 billable conversations and 50 Lyro conversations, one time
- Lyro is also sold on its own for other help desks
- Two meters: billable human conversations and Lyro AI conversations are priced separately
- Lyro's entry quota is small: 50 AI conversations for $39 a month billed monthly, which is $0.78 a conversation
Plan | Billed monthly | Billed yearly | Includes |
Free | $0 | $0 | 50 billable conversations, 50 Lyro conversations one time |
Starter | $29 a month | $24.17 a month | 100 billable conversations, operating hours |
Growth | From $59 a month | From $49.17 a month | From 250 billable conversations |
Lyro AI Agent | From $39 a month | From $32.50 a month | From 50 Lyro AI conversations |
3. Fin: best for helpdesk teams that pay per resolved conversation

- AI agent that answers and takes actions inside a helpdesk
- Works with the Intercom helpdesk or an existing helpdesk
- Spending limits and alerts on outcome charges
- Fin Voice for phone, priced through sales
- No charge when Fin cannot answer or the customer asks for a person
- At most one charge per conversation
- 14-day trial with no credit card
- Cost rises with volume: 1,000 outcomes a month is $990
- Helpdesk seats are a separate cost
Item | Price on the Fin pricing page |
Fin AI Agent | $0.99 per outcome |
Intercom helpdesk | From $19 per seat a month |
Fin Voice | $1.99 per voice outcome, through sales |
How the three compare for after-hours coverage
Tool | Best for | After-hours handling | Entry price | G2 |
Answering from your own content around the clock | Working hours mode, away message, contact capture | $59 a month billed monthly, $39 a month billed yearly | 4.9 (12) | |
Tidio | Small stores with live chat and AI in one inbox | Operating hours, offline auto-reply | Free, then $29 a month billed monthly | 4.6 (1,992) |
Fin | Helpdesk teams paying per resolution | AI agent inside the helpdesk | $0.99 per outcome | 4.5 (3,938) |
Common 24/7 chatbot mistakes
- Promising a reply time nobody meets. If the away message says "within the hour", someone has to answer within the hour.
- Letting the bot guess at night. With no person to catch a wrong answer, the bot should decline when the content has no answer. The guide to stopping chatbot hallucinations covers the guardrails.
- Losing the transcript. A customer who wrote at midnight should not have to explain again at 9.
- Paging people for everything. If every escalation wakes someone, the on-call rota fails within weeks. Keep urgent rules narrow.
- Never reading the night transcripts. The questions the bot could not answer show which pages to write next.






























