Short answer: to get a reliable call center quote you need five inputs before asking for prices: contact volume by channel and time interval, average handle time (AHT), coverage hours, target service level and pricing model. With those, any serious provider can size the team and give you a price you can compare. Without them, every proposal guesses differently and you end up comparing apples to oranges.
Last updated: November 2026.
If a term is new to you, keep our call center and BPO glossary open while you read.
A customer service line is not priced like a sales campaign. Before talking prices, clarify whether your operation is inbound (you receive contacts: service, support, complaints), outbound (you make contacts: sales, collections, surveys) or back office (work customers never see, such as document validation or email case handling). Many programs mix all three, and each block is best priced separately.
Also write down the business goal in one sentence: cut wait times, raise conversion, recover receivables or free up your in-house team. That goal defines the metrics you will hold the provider to.
This is the most skipped step and the one that inflates costs the most. For each channel (voice, WhatsApp, chat, email, social) you need:
If the program is new and you have no history, use reasonable assumptions and tell the provider. The worst case is a made-up number nobody questions.
Coverage hours drive cost more than any other variable. With a 40-hour work week, one seat staffed 8 hours Monday to Friday takes a little more than one person, while a 24x7 seat needs four times the hours before counting absences.
| Coverage | Hours per week | People per seat | Example: 10 seats |
|---|---|---|---|
| 8x5 (Mon-Fri) | 40 | 1.15 | 12 people |
| 10x6 (Mon-Sat) | 60 | 1.7 | 18 people |
| 12x7 (every day) | 84 | 2.4 | 25 people |
| 24x7 | 168 | 4.8 | 49 people |
Languages also move the price: bilingual agents cost more and the talent pool is smaller, so decide whether you need English or Spanish across all hours or only in some intervals. Nearshore hubs like Colombia are strong for bilingual English and Spanish support in US time zones.
Service level defines how fast contacts are answered: 80/20 means 80% answered within 20 seconds. Moving from 80/20 to 90/10 can require 15% to 25% more agents, so do not ask for more than your business needs.
Beyond service level, define 4 to 6 KPIs for the contract: abandon rate, first contact resolution (FCR), quality, CSAT and, for outbound, conversion or recovery. Anything not in the SLA is hard to enforce later.
With volume, AHT and service level, the provider calculates agents with a model such as Erlang C, then adds shrinkage: paid time when agents are not available (breaks, training, meetings, absences), usually 25% to 35%. Check that the quote states it explicitly. A proposal that looks cheap because it ignores shrinkage ends up missing the SLA or billing you overtime.
For a first number before talking to providers, use our call center cost calculator: it estimates what an in-house team costs you today, with payroll burden, supervision, technology and turnover, and compares it with nearshore outsourcing or adding AI agents.
| Model | How it works | Best for | Watch out for |
|---|---|---|---|
| Per seat or dedicated agent (FTE) | A monthly fee for each assigned agent | Steady customer service, specialized support | Who covers absences and how many hours are included |
| Per hour | You pay for each agent hour | Variable volume, seasonal campaigns | Whether logged or productive hours are billed |
| Per transaction | You pay per contact, minute or resolved case | High, stable volume, back office | A clear definition of what counts as a transaction |
| Pay for performance or hybrid | Fixed base plus a fee per sale, appointment or collection | Sales, telemarketing, collections | Clear attribution rules and sale quality |
Ask who provides and who pays for each piece: telephony and numbers, omnichannel platform, CRM, recording, WFM, quality tools and licenses. Decide whether agents work in your CRM or the provider's and how data flows between them. Poorly defined integrations are the most common cause of launch delays.
If agents handle card data, require PCI DSS compliance, and make sure the contract covers the data protection rules that apply to your customers, such as CCPA or GDPR, plus Colombian Law 1581 when the team operates from Colombia.
Pricing a call center today without considering AI leaves money on the table. Review your contact reasons and flag the repetitive ones: order status, hours, appointments, balances, documents, FAQs. An AI agent can resolve part of them over WhatsApp, chat or voice and hand the rest to a human with full context.
Typical containment ranges from 20% to 60% depending on the type of requests. Ask the provider to separate AI-handled volume from human-handled volume in the proposal, with different prices for each. If you are torn between outsourcing, automating or upgrading your team, read call center, AI or an AI-powered in-house team: which one fits.
Send every provider the same package: scope, volumes, hours, channels, languages, SLAs, preferred pricing model and technology. Ask them to answer in the same format, including the number of agents, supervisors and the shrinkage they assumed. That shows whether a low price comes from efficiency or from an optimistic assumption.
| Item | What to check | Priority |
|---|---|---|
| Capacity plan | Agents, supervisors, shrinkage and AHT assumptions | Key |
| Price and model | What is included, what is billed separately and how volume changes are handled | Key |
| SLA and credits | KPIs, targets, measurement and the range of bonuses or penalties | Key |
| Launch plan | Hiring, training, nesting and go-live date | High |
| Technology and AI | Platform, integrations, licenses and automation scope | High |
| Quality | Scorecard, share of interactions monitored and calibrations | High |
| Continuity and security | Contingency plan, redundancy, data protection | Medium |
Before moving the whole operation, start with a 60 to 90-day pilot with reduced targets at first. Agree on the ramp-up curve, who bears its cost and which results trigger scaling. That way savings and quality are proven with data, not promises.
It depends on hours, channels, language, service level and pricing model. For US companies, nearshore outsourcing to Colombia typically costs 40% to 60% less than a comparable in-house team, because the provider absorbs hiring, attrition, supervision and technology. You can estimate your case with the calculator.
Volume by channel and interval, AHT, coverage hours, languages, service level, contact reasons, the systems agents will use and your preferred pricing model. With that, the quote is accurate from the first version.
A mid-size program usually goes live 3 to 6 weeks after signing, depending on hiring, training and integrations. AI agents can be ready sooner if processes are documented.
Per seat gives stability and dedicated teams; per hour gives flexibility when volume swings. Either way, clarify whether logged or productive time is billed and who covers absences.
Calculate your current cost, review the key terms or talk to an Intelsa expert who can help you size your program.
Use the calculatorSee the glossaryTalk to an expert
