TL;DR
200 calls a week is about 33 a day and about 6 in the busiest hour, or 58 hours of phone work a month: a quarter of one person's time. But one person alone leaves about 2 in 7 busiest-hour callers unable to get through; two people or two AI lines cut that to about 1 in 18. An AI answering every call uses about 2,600 minutes a month.
A store getting 200 calls a week needs two lines free in its busiest hour: two people, or one person plus an AI that picks up when they are busy. 200 calls is about 33 a day and about 6 in the busiest hour. One person answering alone leaves about 2 in 7 of those busiest-hour callers unable to get through; a second line cuts that to about 1 in 18.
The total workload is small, about 58 hours a month, which is a quarter of one person's time. Timing is the problem: calls bunch up while that person is packing, at lunch or on another call. The steps below show the sum, so you can redo it with your own call log.
Step by step: from 200 calls a week to lines needed
We assume a store open 10 am to 7 pm, Monday to Saturday. Steps 2 to 4 are our assumptions and the rest is our calculation, so change any number to match yours.
- Calls a day: 200 ÷ 6 days = 33.3, so about 33.
- Busiest day (assumption): it carries 25% more than average, such as the day after a sale: 33.3 × 1.25 = 41.7, about 42 calls.
- Busiest hour (assumption): 15% of that day's calls land in one hour: 42 × 0.15 = 6.3, about 6 calls.
- Handle time (assumption): 3 minutes of talk plus 1 minute of notes = 4 minutes a call.
- Traffic in erlangs: calls in the hour × minutes each ÷ 60 = 6 × 4 ÷ 60 = 0.4. The line is busy 40% of that hour.
- One line (Erlang B): busy chance = 0.4 ÷ (1 + 0.4) = 0.286. About 29%, or 2 in 7 callers can't get through.
- Two lines: (0.4² ÷ 2) ÷ (1 + 0.4 + 0.4² ÷ 2) = 0.08 ÷ 1.48 = 0.054. About 5%, or 1 in 18.
- Three lines: 0.7%, about 1 in 140.
A "line" is anything that can hold one conversation: a person, or one concurrent call on an AI plan. The busy chance also assumes the person is at the phone whenever they are not already talking. If they step away to pack orders, more calls go unanswered.
Why Erlang B and not Erlang C?
Call-centre Erlang calculators usually use Erlang C, which assumes callers wait in a queue with hold music until someone is free. Erlang B assumes there is no queue, so a call that finds every line busy is lost (Erlang formulas explained). A store owner's mobile or a small team's shared number behaves like Erlang B: the caller hears a busy tone or rings out, and some will not call back.
If you do add a queue, Erlang C at the same 0.4 erlangs says 40% of callers wait with one person, an average of 160 seconds. With two people, 7% wait and the average wait is 10 seconds.
Busy-line lookup table for small stores
Find your busiest-hour call count, then read across. Figures are Erlang B with 4-minute calls: the share of callers in that hour who can't get through.
| Calls in your busiest hour | 1 line | 2 lines | 3 lines |
|---|---|---|---|
| 2 | 12% | 0.8% | Under 0.1% |
| 4 | 21% | 2.7% | 0.2% |
| 6 | 29% | 5.4% | 0.7% |
| 8 | 35% | 8.5% | 1.5% |
| 10 | 40% | 12% | 2.5% |
| 12 | 44% | 15% | 3.9% |
We suggest aiming for about 5% or less in the busiest hour. At 200 calls a week, two lines get you there. From about 7 calls in the busiest hour, roughly 225 calls a week on our assumptions, two lines miss 7% or more, so plan for three.
The same maths applies to AI plans. Voice Nimble's Basic plan allows 1 concurrent call and VaniAgent's Unlimited Calls plan has 1 active calling lane, so each acts like one person on this table. Vomyra's Developer plan allows 2 concurrent calls and Voice Nimble's Advanced plan allows 3, so check concurrency before you compare prices.
What 200 calls a week adds up to each month
- Calls: 200 × 52 weeks ÷ 12 months = 866.7, about 867 calls.
- Person time: 867 × 4 minutes = 3,467 minutes, about 58 hours.
- One person's month: 9 hours × 26 days = 234 hours, so the phone takes 58 ÷ 234 = 25% of it.
- AI minutes if the AI answers every call: 867 × 3 minutes of talk = 2,600 minutes.
How we compared: prices from each provider's official pricing page, checked in September 2026. Prices exclude 18% GST unless stated.
What those 2,600 minutes cost on plans with a public price, before GST (our calculation):
- VaniAgent: 2,600 × ₹4.5 = ₹11,700 + 18% GST.
- Superfone AI Receptionist: ₹9,000 per six months includes 1,800 minutes, about 300 a month. The other 2,300 minutes a month at its ₹5 top-up rate add ₹11,500, so about ₹13,000 a month + 18% GST.
- Vomyra Developer: ₹15,000 a month on an annual plan covers 2,500 minutes, just short of 2,600.
- Ringg AI: 2,600 × ₹6 = ₹15,600, plus ₹499 for the number = ₹16,099, subject to Ringg's monthly minimum commitment.
If the AI only takes the calls the person misses while busy, it handles far fewer:
- On an average day this store gets about 3.7 calls an hour (33.3 ÷ 9). That is 0.25 erlangs, so one person misses about 20% of calls.
- 867 × 20% = 173 calls × 3 minutes = 520 AI minutes a month.
- On Superfone, the 300 minutes a month its six-month plan averages leave 220 at ₹5, so about ₹2,600 + GST a month. After-hours calls add minutes on top.
Which calls to hand to AI first
Tag one week of calls by reason. The mix below is an example to show the method, not a benchmark; replace the shares with your own.
| Call type | Example share | Calls a month | Hand to AI? |
|---|---|---|---|
| Where is my order | 40% | 347 | First: a read-only order lookup |
| Return or exchange question | 20% | 173 | Second: policy answer, then the returns link |
| Product or size question | 15% | 130 | Third: answers you approve |
| Refund status | 10% | 87 | Yes, if it can read refund status |
| Complaint or damaged item | 10% | 87 | No: straight to a person |
| Bulk order or other | 5% | 43 | No: take details, call back |
Start with order status because the answer comes straight from your order data and nothing changes on the order. Our guide to Shopify AI voice agents for order and return calls covers the data an assistant needs for that lookup. Keep complaints with a person from day one, and never let the AI take card numbers, UPI PINs or OTPs.
Three setups that work at 200 calls a week
- Owner plus a second person in the busiest hours. It works if someone else can pick up during the two or three hours your call log shows as the peak. You get no AI cost, but evenings and Sundays go unanswered.
- One person plus AI for overflow and after-hours. This gives two lines in the busiest hour and cover at night, for about ₹2,600 a month on Superfone's pricing (billed as ₹9,000 per six months) plus any after-hours minutes. It is the best value for most stores this size.
- AI answers every call, and a person takes hand-offs. With 2 concurrent AI calls, about 5% of busiest-hour callers still hit a busy line; with 3, under 1%. At ₹11,700 to ₹16,099 a month for 2,600 minutes, it is worth it when order-status calls dominate, but check each plan's concurrency before you buy.
To see how these compare with hiring or outsourcing at a larger volume, read our cost comparison of AI answering, a call centre and in-house agents, and our list of AI answering services for online stores compares vendors. If you only need to know who called so you can ring back, a missed-call alert service is cheaper still.
Where TalkEasy by GrowEasy fits
You need two numbers for this sum: calls in your busiest hour and average talk time. TalkEasy by GrowEasy's call analytics track total calls, answer rate, missed calls, average talk-time and peak calling hours, with daily, monthly and per-agent reports. A whole team can share one business number, and incoming calls can be routed by availability with separate rules for office hours and after-hours.
The AI receptionist answers calls after hours, at weekends and on holidays, captures the caller's details, and then transfers the call or books a callback. It is not on the public price list; the Pro plan is ₹999 + GST a month with unlimited users (pricing). For how AI receptionists compare on price and languages, see our AI receptionist guide for India.
See how TalkEasy's AI receptionist covers your busy hours
Last reviewed 24 September 2026.
