Ask five BPO companies for a quote and you will get five numbers that cannot be compared, because each one is priced on a different model with different assumptions buried inside. This guide explains the models, the realistic ranges, and the costs that never appear on the first quote.
One caveat before the numbers: pricing moves with salaries, city, language mix and volumes. Treat everything below as directional and get a scoped quote for your actual process. That said, directional numbers are exactly what you need at the shortlisting stage, and most vendors will not give you any.
Per FTE (full-time equivalent). You pay a monthly rate per dedicated agent. It is the most common model for domestic voice work in India and the easiest to budget. The catch: the vendor’s incentive is to keep headcount high, so pair it with productivity metrics you review monthly.
Per transaction. You pay per call handled, ticket resolved or document processed. Good once volumes are predictable and the process is stable. Risky in year one, before anyone knows what “normal volume” looks like.
Per hour or per minute. Common for shared (non-dedicated) support and overflow lines. You only pay for usage, but you also give up dedicated agents who know your product deeply.
Outcome-based. Payment tied to collections recovered, sales closed or CSAT achieved. The most aligned model and the hardest to negotiate, because it requires trust and clean data on both sides. Usually the second contract, not the first.
Most mature engagements blend these: a per-FTE core team, transaction pricing for overflow, and outcome incentives on top.
For domestic India support (Hindi, English and regional languages), fully loaded per-agent monthly pricing typically falls between 25,000 and 50,000 rupees depending on channel, complexity and city. Voice sits at the top of that range, chat and email lower, and specialised work such as insurance or credit card servicing higher still.
For international English support delivered from India, per-hour pricing generally runs 6 to 12 US dollars, against 25 to 35 dollars for the same seat onshore in the US. That gap is why India still handles a large share of global support volume even as automation grows.
Three factors move you within these ranges more than anything else. Channel mix, because one chat agent can hold two to three concurrent conversations while voice is one at a time. Complexity, because a level-one FAQ line and a retention desk are different jobs at different salaries. And delivery city, because a seat in a smaller city can cost meaningfully less than the same seat in a metro, often with better attrition. This is one reason we spread delivery across ten centres, from Thane and Bengaluru to Bhubaneswar and Sawantwadi, and let clients choose the mix.
Transition cost is the big one: training, knowledge transfer, parallel running, and the productivity dip in the first 60 to 90 days. Budget roughly one to two months of run cost for a clean transition.
Then there is technology. Clarify who pays for telephony, CRM licences, dialers and QA tooling. Vendors with their own stack can absorb much of this; we deploy our own platforms, including the Aurexion CX suite and the Vaani dialer, which keeps third-party licence costs out of the client’s bill for those layers.
Management overhead sits on your side of the table: someone in your team needs to own the vendor relationship. Plan for a fraction of an FTE, not zero.
And attrition cost hides inside quality. Every agent who leaves takes product knowledge with them. When you compare two quotes, a vendor with 30 percent annual attrition at a slightly higher price is usually cheaper than one at 60 percent attrition with a lower rate card.
The old outsourcing equation was labour arbitrage: same work, cheaper geography. That equation is dead, even if plenty of vendors still quote against it. The current one is different: automation removes work rather than relocating it, which means the right partner is not the one with the cheapest labour but the one whose AI takes the most work off the table entirely.
In our operating model, AI leads up to 90 percent of the workflow and people own the last 10 percent, where empathy and judgment actually change outcomes. In practice that looks like bots and intelligent document processing clearing routine volume end to end, agent assist and knowledge retrieval cutting handle time on the contacts that do reach a person, and automated QA scoring 100 percent of interactions instead of the 2 percent sample a human QA team can manage. On QA alone, that shift can take out most of the cost of a traditional quality function while covering fifty times the volume.
The practical consequence for buyers: stop comparing per-seat rates in isolation. A vendor whose automation removes 30 percent of your contact volume at a higher per-seat rate will beat a cheaper vendor who answers every contact with a human. Ask every bidder one question: what will my total cost per resolved contact be in month twelve? That number, not the rate card, is the real price.
Say you run 60,000 support contacts a month across voice and chat with a 40-person in-house team. A conventional outsourcing quote might put 35 agents against that volume. An AI-led approach would typically start with process mapping, deflect a quarter to a third of contacts through self-service and automation within two to three quarters, and staff perhaps 22 to 26 people against what remains, with QA and reporting automated. The second approach can cost more per seat and still come in 25 to 40 percent lower on total cost, with faster answers on the contacts that matter.
Generic calculators will not price your process, because the honest answer depends on your volumes, languages, complexity and targets. What we can do is run AI-driven process mapping on your current operation and give you a scoped quote with the month-twelve cost per resolved contact made explicit, so you can compare it against any other bid on equal terms. That is also the fastest way to see the difference between a transformation partner and a staffing vendor: one prices the process after redesigning it, the other prices your current process and adds margin.
Write to enquiry@eosglobe.com with a rough description of volumes and channels, and we will come back with a range before you ever sit through a sales deck.