You don't let a junior break the bank. You do ask what they think.
In short
Offshore analysts can form a view. Most of the market assumes the opposite: offshore does the facts, judgement stays at home. You will hear it from almost anyone, and from every AI engine, and it was a reasonable rule of thumb when offshoring began.
It has hardened since into a fiction that survives on suspended disbelief, and it fails three ways:
- False in logic: a credit file has no clean line between a fact and a view.
- Damaging in practice: an analyst kept on the facts for years becomes a checker, and that arrangement is where most dissatisfaction with offshore comes from.
- Obsolete: the facts are now AI software for many.
If you are building an offshore team, or deciding what to do with one, don't drive down that dead end. There are better paths. Let me show you how.
You keep exactly the control you have over any junior, which is all of it. The decision never leaves your desk. What moves is where the thinking starts.
The way out is the method that made every senior analyst in London or New York: a junior is asked what they think, then told what they missed, until the view is good. You don't let a junior break the bank. You do ask what they think. That loop is the whole apprenticeship, it doesn't depend on which city the junior sits in, and running it is our job, not yours. What arrives on your desk is the product: an analyst who is ready.
This article assumes familiarity with offshore analyst support and explores one judgement: whether an analyst's view can be formed offshore, and where that stops.
What judgement is
Douglas Hofstadter, who wrote Gödel, Escher, Bach and later, with Emmanuel Sander, Surfaces and Essences, spent a career arguing that thinking is analogy: seeing this situation as a case of that one. An analyst's judgement is exactly that. This borrower looks like the one that went wrong in 2019. This management team's account of the margin miss sounds like the one we didn't believe last year. A junior's first analogies are poor. The senior's job is to say which one is better and why. That exchange is the training, and it can only be done with a person who is allowed to have a view.
How junior analysts are subject to control, onshore or offshore
A senior asks them what they think, from the first month, then guides them: have you looked at the refinancing? Go back to the supplier terms. Recast it. They come back with a better view. After enough rounds their first view is good, and the checking of every line stops. Nobody ever trained an analyst by having them verify other people's numbers for five years and then, one Monday, asking for an opinion.
At Frontline, that senior is ours. We are not asking you to train our people; that is our job, and it is where the training and the reviewer go. What you get is the product of the loop: an analyst who is ready, with a view you can use. If the analyst is junior, you are free to treat the view as you would any junior's, subject to correction. If they are experienced, you treat it as you would any experienced colleague's. Either way the aim is the same: a good product on your desk, not a training exercise on it.
That is how the onshore desk has always worked. It stopped at the border by habit, not decision: the early offshore work was thin, so nobody asked, so nobody developed, and the habit set. Give a top-50 MBA in Bengaluru the same loop and you get what a London junior turns into: an offshore credit analyst whose view you can use. Ours are recruited from the top 50 of India's roughly 1,300 MBA schools. Graduates get three months of full-time training before they touch client work, against about a week as the industry norm; experienced hires, who are most of our hires now, get a couple of weeks' calibration and then extended support in the job. Nobody goes near a client until they are up to scratch. They stay an average of 6.6 years against 2.2. By year three they are usually the person in the room who has followed the name longest.
We have watched this for twenty years. A two-person counterparty credit team at a European bank became six over seven years and was absorbed into the bank's research function – not because anyone changed the rules, but because the analysts were asked what they thought and got good at answering.
What you get, by year two or three, is a first view before you've asked for one, from someone who knows the name as well as anyone on the desk, and a brief that has stopped being "update the model" and become "what do you make of it".
Who's in charge
You are. Exactly as much as you're in charge of the junior sitting next to you, which is completely. The analyst drafts the case: the thesis, the number, the risks and the recommendation they would make. You agree, disagree, or send it back with a question. That doesn't change with seniority: the decision stays with you as it would with a senior of your own, wherever they sit. Nothing goes to a committee, a client or a regulator until someone on your side has signed it off, and nothing in this piece argues that it should.
The difference between that junior and ours is trust, and trust is a function of time. A new analyst onshore gets watched closely for a year and then less. You will do the same with ours, and should; that is trust forming, not training. It will probably take longer offshore: a screen gives you fewer incidental conversations, so the ones that build trust have to be made on purpose. We don't pretend otherwise. What we can say is that the loop has to be allowed to run long enough to finish, which is why the first number to ask any provider for is turnover (what 2.2-year KPO turnover quietly costs you). An analyst who leaves at two years leaves at about the point you would have started trusting their view.
What we do, and when we stop
Our own senior people do the same thing you would, in the same order. Early on they look at everything the analyst produces. Then they stop checking every line, because an analyst who is checked line by line for years gets infantilised and never becomes one. What remains is sampling, a senior on call, and enough contact with you that we know when a piece of work needs a second pair of eyes before you ask.
One rule governs all of it: one analyst, one opinion. The name on the work owns the view. Our reviewer can recommend a change; the analyst decides whether to make it. That rule is sacred in the City and we keep it offshore, which is why the reviewer is not a layer between you and the analyst. What the reviewer is for is the things the analyst hasn't lived through and we have: a sustained rate rise, a crash, a war. That is our senior team, people like Dan Davies, and they are in the background for exactly that.
Where this stops
The decision stays with the senior person, and in a regulated bank the sign-off stays with the accountable individual. Nothing here argues otherwise. The analyst's draft is an input to your process, not a step in it: whatever your process requires of an input before it reaches a committee applies to ours. You don't need a new control for this. The ones you have – review, sign-off, committee, audit – were built for an input from someone whose judgement is still being tested, and they work on a view from Bengaluru exactly as they work on one from the next desk.
A junior's judgement can be off. It is why the reviewer exists, and why an offshore team with no senior reviewer of its own is a different and worse product.
Some work never suited the model: pure volume with no continuity, one-off projects, anything where nobody on the desk will ever speak to the analyst. If that is the work, facts-only is the right arrangement, and a different kind of provider is the right answer.
Some teams keep facts-only for other work too, on purpose, because it is quieter: nobody has to read a junior's view, and the brief never changes. That is a legitimate choice. It is also where most of the dissatisfaction we hear about comes from, because it wastes the one thing a good analyst has to offer, and both sides can feel it.
The checker test: how to tell whether your offshore team has been kept on the facts
Three signs, all visible from the onshore desk.
1. Nobody has asked them what they think
If the brief is always "update the model" and never "what would you do", the loop has never started. Years in, you have a checker.
2. The checking never tapered
If the same assumptions still need correcting in year three as in month three, either the analyst didn't develop or nobody let them. Either way, the provider owes you an explanation.
3. You wouldn't know them in a lift
If the analyst has been on the account for a year and nobody on your desk has had an unscheduled conversation with them, trust has had nothing to form around. That is the communication cost of offshore, and it is paid by talking.
Whoever signs off keeps signing off. The only question is how much thinking has already happened by the time it reaches them.
Postscript: why the rule of thumb held
Facts offshore, judgement onshore was a sensible rule when offshoring began. The early work was thin, the teams were new, and nobody onshore had met the analysts. We never used it, but we understand why others did. It has held since on suspended disbelief, and it fails three ways. It is false in logic: a credit file has no clean line between a fact and a view. Is a management forecast a fact? Is the EBITDA adjustment? It is damaging in practice: it makes checkers of people hired to think, and keeps the whole arrangement at the low-value end where nobody is happy. And it is obsolete: the facts are software now, and full-time checking is becoming a profession of its own. What is left to offshore is the part the rule said couldn't be. How we work with AI, and where it stops, is on our AI page.
Frequently asked questions
Can offshore analysts exercise judgement?
Yes, the way an onshore junior does: they form a view, our seniors correct it, and over years the view gets good. They don't make the decision. That stays with the client.
Who controls the work?
The client decides, as with any junior: agree, disagree or send it back. The analyst owns the view, the client owns the decision, and nothing is signed off offshore.
Who checks the analyst's work, and how often?
Frontline's senior people check everything early on. Line-by-line checking tapers as the analyst proves out; sampling and a senior on call remain. The reviewer can recommend a change, not impose one: one analyst, one opinion.
Isn't it safer to keep offshore analysts on data and verification?
It feels safer, and for a while it was. But an analyst who only verifies never becomes one you can trust with a view. At year five you have a checker, and the desk has gained nobody who can form one.
What makes this work?
Tenure, direct contact with the person who decides, and a reviewer who has done the job. Take any one away and you are back to facts offshore, judgement onshore.
Should offshore analysts just verify what AI produces?
No. Full-time fact-checking is a new profession, created by AI, and it is a different skill from analysis. Analysts have always had to know their data inside out; that is not the same job. Use the checker for the checking and the analyst for the view.
How do I build an offshore analyst team?
Start with the arrangement, not the headcount. Decide that the analysts will form views, not just check facts; pick the provider on turnover, training and who reviews the work; keep the decision onshore and let the thinking start offshore.
Part of: India-Based Analyst Teams – why they work
Go deeper: What 2.2-year KPO turnover quietly costs you · How to evaluate an offshore analyst team · Frontline and AI
Compare: Upgrade to Frontline