Skip to main content

How to plan live chat staffing for a small team

Live chat staffing for a small team comes down to your busiest hour, how long each chat takes and how many chats one person can handle at once.

Emily Lewis, Senior Content Editor at LenoChat
Emily Lewis
Content Editor

Key takeaways

  • Staff for your busiest hour, not your daily total. That hour decides how many people you need.
  • Most practitioners recommend 2 to 3 chats per agent at once, and no more than 2 for complex questions.
  • Only 39% of US consumers prefer contacting businesses between 9am and 5pm on weekdays.
  • Divide your peak workload by chats per person, then add about a third for breaks and admin.
  • A chatbot only works as cover if customers can reach a person. 84% of US consumers say that matters.

You can work out live chat staffing from three numbers: chats in your busiest hour, minutes per chat and how many chats one person can juggle. Speed counts: in a 2026 Moneypenny survey, 72% of US consumers said they’d probably go with whichever business gets back to them first. Here’s how.

How many chats will your business get?

A small online seller writing in a notebook beside an open laptop, with shipping boxes and a clothes rail around the desk
Photo: Kampus Production (Pexels License)

Your best estimate is the number of questions you already answer by email, phone and social messages, since chat mostly gives the same customers a faster way to ask them. Start there rather than guessing.

Count those questions for two ordinary weeks. Write down the hour each one arrived, not just the day. Staffing depends on your busiest hour, so a shop with forty questions spread evenly across the day needs fewer people than one that gets thirty of them over lunch.

Timing also tells you which hours you can’t ignore. A 2026 Moneypenny survey of 2,000 US consumers found 39% picked weekday office hours, 9am to 5pm, as the time they’d most like to get in touch, while 24% chose evenings from 5pm to 9pm. If your chat closes at five, the customers who’d rather write after dinner will find it shut.

Then mark the days that break the pattern. The morning after a weekend, the day a promotion email goes out and the week before a holiday are the ones to check. Your own count will show which of them spike, and those are the days you plan for separately.

If you already run chat, skip the estimate and read your reports. Chats per hour and first response time, the gap between a customer’s first message and your first reply, are the live chat metrics that show whether your team is keeping up.

How many chats can one person handle at once?

A support agent wearing a headset works at a laptop in a row of desks, with a colleague out of focus in front
Photo: Tima Miroshnichenko (Pexels License)

Two or three at once is the ceiling most practitioners named when Call Centre Helper asked them in 2022, and several capped it at two for complicated questions.

That number is concurrency: the average count of chats one agent works on side by side. You measure it by dividing the chats open at a given moment by the agents on shift. In a 2022 Call Centre Helper example, 15 open chats shared by 5 agents gives a ratio of 3 per agent.

The limit and the average aren’t the same thing. Call Centre Helper suggests planning with a concurrency of between one and two, because agents rarely sit at their limit all hour. Its 2022 worked example counted 39 chats, handled at an average of 1.3x side by side, as the same load as 30 single contacts.

Pushing past the limit has a cost. Contributors to the same Call Centre Helper article warned that each chat takes longer as an agent adds more, which can lower satisfaction, and that juggling several makes it easier to confuse one customer’s question with another’s. Sending one shopper’s order details to another is the kind of mistake people remember.

Two things raise a person’s safe limit. Saved answers cut typing time, so a short set of canned responses for live chat support pays for itself quickly. Experience matters too: start new hires on one chat at a time and add a second once they know your products, a routine covered in training new support agents for chat in their first week.

How do you turn chat volume into a headcount?

Multiply the chats in your busiest hour by the minutes each one takes, divide by your concurrency, then add room for breaks and round up. That gives you a sound starting number, not an exact one.

The standard call-center staffing tools won’t do this for you. Erlang models, the formulas contact centers use to turn call volume and a wait-time target into a headcount, are built on the idea that an agent finishes one contact before starting the next, so they need adjusting before they work for chat. Jonty Pearce, founder of Call Centre Helper, says it plainly: “There is no easy formula for web chat that I know of.”

Two terms make the math readable. Handle time is the minutes an agent spends on one chat from greeting to close. Shrinkage is paid time when agents can’t take chats, such as breaks, training and meetings.

His site’s 2022 worked example shows the adjustment. Picture a team receiving 100 chats every hour, each lasting 16.5 minutes, aiming to answer 80% inside 60 seconds and losing 35% of paid time to shrinkage. With every agent on a single chat, it needed 51 agents. At an average concurrency of 1.6, that fell to 32 agents, a figure the author offered as a sensible place to begin and not an exact count.

Your numbers will be far smaller, and you can do the same math on paper in five steps.

Step 1: Find your busiest hour

Take the hour with the most chats or questions from your two-week count. That’s the hour you staff for, even if the rest of the day is quiet.

Step 2: Time a typical chat

Note how long a few dozen chats take from first message to close and take the average. If you don’t run chat yet, time how long a typical email reply takes and treat it as a rough guide.

Step 3: Work out the hours of work in that hour

Multiply chats by minutes per chat and divide by sixty. Ten chats at twelve minutes each is two hours of work landing inside one hour.

Step 4: Divide by your concurrency

Use a planning figure between one and two, not your agents’ maximum. Two hours of work at a concurrency of one and a half comes to about one and a third people.

Step 5: Add shrinkage and round up

Add roughly a third for breaks, lunch and admin, close to the allowance the worked example used, then round up to whole people. In this example you need two people on at peak, and one may be enough for the quieter hours.

Then check the result against a speed target. A common benchmark has 80% of chats picked up inside 40 seconds, according to Lorikeet’s roundup of response time benchmarks. Lorikeet also reports that a 2025 study of more than 220 million chats from 2024 measured a typical 23.6 seconds before an agent sent a first reply. If two people can’t get near that at peak, add a third for that hour or let a chatbot take the first reply.

See how LenoChat handles this for small teams.

How do you cover lunch, nights and busy days?

You cover them with one of three setups: one person with fixed chat hours, two or three people on staggered shifts, or people by day with a chatbot for the gaps. Which one fits depends on how many chats arrive outside your core hours and how many of them are repeat questions.

The table compares the three on hours covered and who answers when nobody is at a desk. The hour windows follow the contact-time preferences in the Moneypenny survey. The chatbot figure comes from one contributor to a 2022 Call Centre Helper article, who estimated that scripted conversation flows backed by a knowledge base dealt with about 80% of routine chats.

SetupPeople on the rotaLive chat hours on a weekdayLunch covered by a personRepeat questions answered after hours
One person with fixed chat hours19am to 5pm, minus lunchNoNo, offline form only
Two or three people on staggered shifts2 to 39am to 9pmYesNo, offline form only
People by day, a chatbot for the gaps1 to 29am to 5pm with people, bot around the clockBot answers repeat questionsYes, about 80% of routine chats

One person with fixed chat hours

A boutique owner checking a tablet beside a rail of clothes in a small shop
Photo: Vitaly Gariev (Pexels License)

This setup works when your peak-hour math comes out at one person. You set chat hours to the times you can actually reply, and the widget switches to an offline message or form the rest of the time.

Lunch is the weak spot. Take it in your quietest hour and switch chat to offline for that slot, rather than leaving a green online dot with nobody behind it. A clear away message for nights, weekends and holidays tells people when they’ll hear back, which matters more than the hours themselves.

If most of your questions can wait a few hours anyway, weigh chat against a simple form. The trade-offs for a team this size are laid out in live chat versus a contact form for a small team.

Two or three people on staggered shifts

A cafe worker checks a tablet at the counter while two colleagues work behind him
Photo: Nathan Dumlao (Unsplash License)

Staggered shifts let a small team cover a long day without anyone working one. One person starts early, another starts late, and both are on during your busiest hour, so the overlap covers the peak and each other’s lunch.

Two people on staggered shifts cover more of the day than one person working longer hours, because the overlap absorbs lunch and the rush. A late shift that runs into the evening also reaches the 24% of US consumers in the Moneypenny survey who’d rather get in touch between 5pm and 9pm.

On busy days, put everyone on for the peak and move other work, like packing orders or admin, to the quiet hours. A shared inbox, where every chat and message lands in one list any teammate can pick up, keeps the handover between shifts clean.

People by day, a chatbot for the gaps

Two chat widget conversations in which the AI assistant offers a Chat with Support button to bring in a person

A chatbot can answer repeat questions at any hour and pass the rest to people when they’re back. One contributor to a 2022 Call Centre Helper article estimated that scripted conversation flows paired with a knowledge base handled about 80% of routine questions, and every chat the bot answers is one your team doesn’t have to staff for.

The handoff is what makes this setup work. In LenoChat, for example, the AI agent replies from your knowledge base and brings in a person, with the whole conversation attached, when it has no confident answer or the visitor asks for someone. Outside your chat hours it collects an email and message instead, so nobody sits in a queue that won’t move until morning.

When should a chatbot cover for a person?

Use a chatbot for repeat questions with one right answer, such as order status, shipping times and opening hours, and only when customers can reach a person easily.

Customers like chatbots less than business owners think. In the 2026 Moneypenny survey, 56% of US consumers viewed live chat favorably, next to 69% of business decision-makers. For chatbots the split was 26% of consumers against 52% of decision-makers, a gap twice as wide.

A clear route to a person changes the picture. The same Moneypenny survey found 71% of consumers are fine with AI handling their questions as long as they can move to a human, and 84% said keeping that option open matters to them.

Skip that route and you risk losing people. A 2024 Gartner survey, reported by Insurance-Canada.ca, found that 64% of customers would rather companies kept AI out of service altogether, and 53% would think about moving to a rival on learning that a company planned to bring it in. Their biggest worry was that getting hold of a person would become harder.

Many customers fear that GenAI will simply become another obstacle between them and an agent.

Keith McIntosh, senior principal in the customer service and support research practice, Gartner

So keep the bot on the questions it answers well, and send anything with feelings or money attached to a person: damaged orders, refunds, complaints and anything a customer has already asked twice. Then rerun the headcount math with only the chats the bot hands over. That smaller number is the one that decides whether you need a second hire.

Check the split once a week. Look at which chats the bot passed to a person and why, then add the missing answers to its knowledge base so fewer of the same questions reach your team next week.

Which setup should you choose?

Let your peak-hour math decide. If it comes out at one person and your evenings are quiet, start with fixed chat hours. If it comes out at two or more, stagger the shifts. If most of your chats are repeat questions, put a chatbot in front of whichever team you have before you hire.

One person with fixed chat hours. Best for owner-run businesses whose busiest hour brings a handful of chats. Not for stores with evening traffic or regular promotions.

Two or three people on staggered shifts. Best for teams whose peak needs two people and whose customers write in the evening. Not for businesses too small to keep a second person busy outside the peak.

People by day, a chatbot for the gaps. Best for stores where most chats are repeat questions about orders, shipping and returns. Not for businesses where most questions need judgment, like custom quotes or complaints.

Whichever you pick, redo the count after a month of real chats. Your first estimate is a starting point, and the real numbers will tell you whether to add a shift, move one or hand more questions to the bot.

Ready to answer every customer from one place? Start free with LenoChat.

Frequently asked questions

What is a good first response time for live chat?

A widely used benchmark is 80% of chats answered inside 40 seconds. A 2025 benchmark of 2024 data put the typical wait for an agent's first reply at 23.6 seconds.

Should a chat widget show as online when nobody can answer?

No. Set chat hours to the times someone can reply and show an offline message or form outside them. A widget that looks live but goes unanswered costs you the sale and the customer's trust.

Do WhatsApp and Instagram messages need their own staff?

Not usually. A shared inbox lets the same people answer every channel. Add those messages to your peak-hour count so your headcount reflects the real load across website chat and social channels.

What is shrinkage in support staffing?

Shrinkage is paid time when an agent can't take chats, such as breaks, lunch, training and meetings. A 2022 Call Centre Helper worked example allowed 35%, so plan for roughly a third more hours than the chats alone need.

Sources

  1. Moneypenny
  2. Insurance-Canada.ca (reproducing Gartner press release)
  3. Call Centre Helper
  4. Call Centre Helper
  5. Call Centre Helper
  6. Call Centre Helper
  7. Lorikeet

Image credits

Related Articles

Store owner filling in order paperwork at a desk stacked with packed shipping boxesMarketing

How to handle an order cancellation request chat without losing the customer

Most order changes and cancellations take one reply, once you know whether the parcel has shipped. Wrong addresses are costly: in Sendcloud’s 2025 survey of UK shoppers, 64.3% said a failed delivery would stop them reordering. Here’s how to run each order cancellation request chat. These chats come in three shapes: fix the address, swap […]

Two customer support agents talking while working at the same deskMarketing

How to run agency live chat for several clients

Table of contents How do you keep each client separate in one inbox? Who should see each client’s conversations? How do agents write in each client’s brand voice? How do you keep service consistent across clients? Which inbox setup should your agency choose? Agency live chat can cover several client brands without mixing conversations or […]

A support agent wearing a headset types a reply on a laptop at an office deskMarketing

Apology message templates for wrong items, late orders, double charges and missed replies

A good apology message template lets you say sorry fast without sounding like a form letter. Only 40% of US complainants came away fully satisfied with how their complaint was handled, according to a 2025 survey by Customer Care Measurement & Consulting. Below are ready-to-paste apologies for four common mistakes. What should every apology message […]