A quote tells you the price. A proposal tells you what the provider promises. Neither tells you how the work actually gets done at three in the afternoon on an ordinary Tuesday. That only shows on the floor. So before signing with a BPO, it's worth visiting the operation —or asking for a remote walkthrough if you can't go in person— and seeing with your own eyes what up to then lived only in a PDF.
The visit isn't a courtesy step or a tour of nice facilities. It's the last chance to confirm that what was promised at the table exists in practice. Whoever uses it well arrives with prepared questions and leaves knowing whether the provider operates the way it says. Whoever wastes it walks a floor, drinks a coffee, and signs with the same information they arrived with.
Go in with your own process decided
The most common mistake is showing up to see "how they work" without being clear on what you need. That lets the provider run the visit while you just nod along. Before you go, it's worth having your own process written down: what you'll outsource, what volume, on what hours, at what service level. With that in hand, the questions stop being generic and become specific to your operation.
This connects to how to read a BPO quote: the visit is where you verify what paper can't prove. An hourly rate doesn't tell you how many agents each supervisor handles, or who answers when the person who knows your account gets sick. That gets asked in person.
What to ask about the operation
The useful questions aren't about company size; they're about how quality holds up every day. Four that almost always reveal something:
- How many agents does each supervisor handle? A supervisor with too many people doesn't coach, they fight fires. The healthy number depends on the process, but the question matters more than the figure: if nobody can answer it, nobody is tracking supervision load.
- Who trains, and how long does training last? That training exists is the bare minimum. What's telling is who delivers it —a dedicated trainer or the supervisor in spare moments— and what happens when a replacement joins mid-operation.
- How do you measure quality, and how often? If the answer is "we listen to some calls," there's no method. A serious provider explains how it audits, with what sample, and what it does with the result. It's worth cross-checking against how quality control by audited sampling works.
- What happens when volume spikes? Every operation has peaks. The question is whether the provider has a concrete plan —shifts, reinforcements, prioritization— or improvises.
What to observe on the floor
Questions get rehearsed answers; the floor gives you the truth. As you walk through, it's worth checking whether what was said matches what you see. If the provider claims every agent has a knowledge base at hand, look at an agent's screen and see if they actually use it. If it says supervision coaches, see whether supervisors are on the floor or shut in an office.
Other things you see without asking: whether the mood is focused work or tension, whether the seats are full or there are many empty chairs mid-shift, whether the screens show operational dashboards or sit dark. None of this is conclusive on its own, but together it sketches the real operation, not the one in the slideshow.
Questions about people
A BPO's quality is, to a large degree, the quality of the people who stay. So there are two uncomfortable questions worth asking. First: how much attrition do you have, and what do you do to lower it? An honest provider won't say "nobody leaves here"; it will say what causes it and what it's doing about it. It's worth understanding what causes attrition and what actually cuts it before this conversation, to tell a serious answer from a brochure line.
Second: who is assigned to my account, and what happens if that person leaves? Many operations depend on one or two people who know the process. If there's no documented backup, that person leaving is your operational knowledge leaving.
Questions about data and security
If the process touches personal data —and almost all do— the visit is the moment to see how it's handled. "We comply with the law" isn't enough. It's worth asking concrete things: who has access to what, how what an agent can see and copy is controlled, what happens to the information when the contract ends. In Colombia, personal-data processing is governed by the habeas data regime, and the client–provider relationship has implications for who is answerable for that data. This is general framing and not legal advice; each contract should be reviewed under its own framework. Even so, seeing the floor helps: an operation that leaves sensitive data in plain sight tells you more than any clause.
Red flags
Some answers should turn on a light. A provider that promises an exact result without knowing your operation is selling, not operating. One that can't explain how it measures quality doesn't measure it. One that answers every question with "we'll sort that out after signing" is postponing exactly what needs to be resolved before. And one that won't let you see the floor —or shows only a staged area— probably has something it doesn't want to show.
No single flag disqualifies, but several together do. The visit exists precisely to gather them before signing, not to discover them in the first month of operation, when it already costs dearly. Whatever you agree afterward should be measured from the start: that's why it helps to settle what to measure in the first month using what the visit revealed.
Where smartBPO fits
When someone evaluates us, we'd rather they come with their process written down and their uncomfortable questions ready. We show the floor as it operates, not a showroom: who supervises, how we train, how we audit quality, and what we do when volume rises. We explain our attrition without dressing it up, and who stands as backup on each account. On data, we show concrete controls, not just clauses. We don't promise numbers we don't control; we'd rather the visit confirm we operate the way we say, so the decision gets made by seeing, not believing.