Writing a call center outsourcing RFP
A template structure, the questions that genuinely differentiate providers, and the ones that guarantee identical answers from everybody.

Most call center RFPs produce a stack of proposals that are impossible to compare, because they asked questions every competent provider answers identically. A useful RFP is designed to create differences, not to collect assurances.
Below is a workable section structure, then the part that matters more: which questions separate providers and which ones waste a page. If you take nothing else, take the principle — ask for evidence and specifics, never for adjectives.
The section structure
Eight sections cover it. Anything longer tends to reduce response quality rather than increase it, because providers start delegating the writing.
- 1. Company and program context — who you are, what you sell, and why you're outsourcing now.
- 2. Scope of work — channels, contact types, hours, languages, and expected volume with its seasonal shape.
- 3. Volume data — actual historical contacts by reason, channel, and hour. This is the section that makes proposals comparable.
- 4. Technology and integration — your stack, what agents need access to, and who provides the telephony.
- 5. Compliance and security — the frameworks you're subject to and the evidence you require.
- 6. Service levels and quality — the targets you'll measure and how you expect them reported.
- 7. Commercial structure — pricing model, term, and how volume changes are handled.
- 8. Response requirements — format, deadline, and who to contact with questions.
Include real volume data — this is the whole game
The most common RFP failure is describing volume vaguely: 'approximately 5,000 contacts per month'. Every provider then prices a different set of assumptions and you compare numbers that mean different things.
Give them contacts by reason, by channel, by hour of day, and by month across a full year. Include average handle time and current resolution rates. A provider who can see that your volume triples in November and that 40% of contacts are one billing question will propose something specific — and the ones who ignore that detail have told you something useful about how they'll operate.
Questions that differentiate
These produce genuinely different answers from different providers.
- Describe how an agent handles a situation your script doesn't cover. Walk through a specific example.
- What is agent attrition on accounts of our size and type, over the last twelve months?
- Will our agents be dedicated or shared across accounts? If shared, how many?
- Describe a program that went badly and what you changed as a result.
- What does your QA rubric measure, and what percentage of contacts are scored?
- Who is the named account lead, and will they still be on this account in year two?
- Give a specific example of absorbing an unplanned volume spike, with numbers.
- What internal time will you need from us during transition, in hours per week?
Questions that waste a page
Every provider answers these the same way, so they consume effort and produce nothing comparable.
'Are you committed to quality?' 'Describe your company culture.' 'Do you have experience in our industry?' — this last one is nearly always yes; ask instead for two named reference clients in your vertical at your size. And 'are you compliant with HIPAA / PCI / GDPR?' is a yes from everyone. Ask for the training records, access-control policy, and audit history instead.
Score before you read
Decide your weighting before the responses arrive, or you will unconsciously weight toward whichever proposal reads best.
A workable starting split: delivery-model and size fit 30%, compliance evidence 20%, quality and QA approach 20%, commercial terms 20%, references 10%. Adjust to your situation — a regulated program should weight compliance far higher — but fix it in advance and score against it.
End with a paid pilot, not an award
The RFP should shortlist two or three providers, not select one. No written response predicts delivery quality. Structure the process so the finalists run a paid pilot on a defined slice of volume with success criteria agreed up front, and make the award decision on that evidence instead.
“A useful RFP is designed to create differences, not to collect assurances. Ask for evidence and specifics, never for adjectives.”
The bottom line
Give providers real volume data so their proposals are comparable, ask questions whose answers can actually differ, fix your scoring weights before responses arrive, and finish with a paid pilot rather than an award. An RFP that shortlists well is doing its job; one that picks a winner on paper is overreaching.

