Checks rule things out. They do not rule them in.

Checks work in one direction

The asymmetry at the centre of this subject, stated once properly.

A check can come back saying something is wrong. No check comes back saying everything is right. The best a clean check can report is that it found nothing — which is a different statement, and the difference is not a quibble.

Why you cannot get to "real"

Three reasons, and none of them is about effort or expense.

1. Only the issuer knows. A document is genuine if the body named on it actually issued it. Nobody else is in a position to assert that, which means every check short of the issuer is indirect by construction.
2. A good forgery is defined by passing checks. That is what makes it good. So passing them is weak evidence precisely where the risk is highest.
3. Some things are not recorded anywhere. If an assertion is nobody's record, there is no source against which it could ever be confirmed, however much you spend.

So the useful goal is not certainty. It is knowing precisely how much uncertainty you still have, and whether you are content to carry it.

The two lists

Which is why a verification should end in two lists rather than one verdict.

Almost nobody asks for list two. It is the one you are actually paying for, because it is the only part that lets you decide whether to rely on the document.

The check that cannot fail

Now the test to apply to every check, and it disqualifies the most popular one in existence.

Look at a certificate. Consider what the possible outcomes are. It looks fine. Or it looks slightly odd and you have no idea what that signifies. There is no outcome in which it clearly fails.

The same test eliminates several other popular rituals:

• Asking the person who gave it to you whether it is genuine. No failure mode.
• Checking that a document is on the right letterhead. The easiest thing in the world to get right.
• Confirming it has a signature and a stamp. So does every forgery.
• Reading it carefully for spelling mistakes. Catches the laziest forgery and nothing else.
• Having a third party look at it and say it seems fine. You have outsourced a check that could not fail.

Reassurance is not information

Worth separating cleanly, because the two feel identical from inside.

Information changes what you would do. Reassurance changes how you feel about what you were going to do anyway. A check that could not fail delivers the second with the texture of the first, which is why it is so popular and so useless.

What would this rule out

One question, applied to everything. Not "is this a good check" but: what specifically would this fail on?

Ranked by what they rule out

So the ranking falls out, and the striking thing about it is the order in which people actually proceed.

One: the issuer

The only check that reaches the thing you actually want to know, because it is the only one that goes to a party whose records are not controlled by the person who handed you the paper.

• What it can rule out: that the document was issued at all. Nothing else reaches this.
• What it still cannot touch: whether what the document says was true when issued — see genuine on false particulars.
• Where it is possible: institutions and authorities that respond to such enquiries, generally with a reason and often only with the holder's consent.
• Where it is not: most private issuers, bodies that do not answer third parties, and anything where the record no longer exists.

For educational qualifications specifically, this is a defined process with its own timelines and its own service — degree verification — because the enquiry goes to the institution and that is not a same-day matter.

Why nobody asks the issuer

Not laziness. Four real obstacles, and they are worth naming so the decision to skip it is a decision.

1. Time. It takes days or weeks, and the thing you are verifying for usually has a deadline.
2. Standing. Many issuers will only respond to the holder, or to somebody with a reason they accept.
3. Consent. Where the subject's consent is needed, asking for it is awkward — and refusal is itself information you may not want to have to act on.
4. Habit. Nobody around you does it, so not doing it feels normal.

Two: the look-up

Where the issuer maintains a record that can be queried, this is close to the first in strength and far cheaper in time.

We are deliberately not naming systems, portals or applications on this page, for two reasons: they are the fastest-changing thing in this subject, and which one applies is specific to the document. What generalises is the question to ask of any look-up, and that is the next section.

• What it can rule out: that a matching entry exists in the issuer's own record.
• What it cannot: anything the record does not contain, and anything about whether the entry itself was correctly made.
• When it is strongest: where the record is maintained by the issuer and reached independently of the holder.

Does the code resolve to somebody else

The single most useful question to ask about any reference number, code or scannable mark on a document, and it is not whether the mark exists.

Three outcomes, and they are very different:

1. It resolves against the issuer's own record, and matches. That is a genuine check with a real failure mode. Strong.
2. It leads nowhere, or produces a page that merely repeats what the document already said. That is appearance with extra steps.
3. It leads somewhere the holder could have arranged. Worse than nothing, because it manufactures confidence.

Three: internal consistency

The check that costs nothing, needs no permission, requires no system, and finds more problems in practice than anything else available to an ordinary person.

The method is physical and dull. Lay the document beside every other document you have and every fact you already know, and look for things that cannot all be true at the same time.

Things that cannot all be true

What to look for, stated as categories rather than as a checklist, because categories transfer and checklists do not.

• Order of events. A date that precedes something it should follow — an authority given after the act it authorises, a certificate dated before the thing it certifies.
• The same fact, twice. A name, a parent's name, a date of birth or an address that differs between two papers. Usually innocent, always worth asking about.
• Geography against time. A document issued somewhere the person was not, in a period you know about from something else.
• Figures that disagree. An amount, a duration, a period of service stated differently in two places.
• The wrong issuer for the thing. A document issued by a body that would not have been the one to issue it — for that subject, in that place, in that year.
• Absence where there should be presence. A thing that, if true, would have left a trace somewhere else in your file and has not.

Two practical notes on how to actually do this, because "look for inconsistencies" is easy to agree with and easy to do badly.

• Read the same field across all documents, rather than reading each document in turn. Take "date of birth" and read it on every paper, then take the name, then the address. Reading documents one at a time is how a difference between the first and the fourth goes unnoticed, because by the fourth you are remembering rather than comparing.
• Write down the facts you already know before you start — the timeline, where somebody was living, where they were working. Those are checks too, and they are the ones a false document is least likely to have been built against, because the person making it did not know what you know.

Why consistency works so well

Worth understanding, because the explanation tells you how hard to look.

Which also tells you where to put the effort: not in staring harder at one document, but in getting more documents onto the table. Two papers to compare is worth more than an hour with one.

Four: appearance, last

Last, and the reasons are worth stating because being told to ignore your eyes is counter-intuitive.

1. It is the easiest thing to get right if somebody is making an effort, which means it is weakest exactly where the risk is greatest.
2. You are not equipped to judge it. You have probably never seen a genuine example of this document from this issuer in this year, so you have no baseline.
3. Genuine documents look wrong all the time. Older documents, documents from small offices, documents that were handwritten, reprinted, folded or photocopied badly. Judging on appearance produces false alarms as readily as false comfort.

Why there is no checklist here

You will not find a list of forgery indicators on this page, and we would rather explain that than appear evasive.

The substitute is the whole of this page: pick checks that can fail, and put your effort above appearance rather than inside it.

There is a version of this worth naming because it is the most consequential single misreading in ordinary practice. People treat a stamp — any stamp, from anybody — as having moved a document up the strength ranking. It has not. An attestation sits beside the ranking rather than inside it: it adds a fact about an act that was performed, and that fact can be perfectly true while the document underneath it remains entirely unchecked. A stamped document and an unstamped one are, for authenticity purposes, in exactly the same position unless the stamp came from the body that holds the record.

Which gives a simple discipline when you are handed something stamped. Ask what the stamp says, rather than noticing that there is one. If it says a copy corresponds to an original produced, the question "was that original genuine" is untouched and still yours. If it says the issuing authority confirms the entry in its own record, that is a different animal entirely and belongs at the top of the ranking. Both look like stamps from across a desk.

Public or private

Before anything else, one classification decides whether verification is even possible.

The private document problem

Agreements, letters, receipts, employer certificates, acknowledgements, experience letters, salary slips, no-objection letters. Enormously common, and mostly unverifiable in the sense people mean.

There is no register behind them. There is only the person or organisation that made them. So:

• Verification means contacting the maker. There is no substitute, and no system to query.
• Consistency becomes the main check, and it is genuinely useful here — a letter that disagrees with a bank statement or a timeline is the usual way these fail.
• If somebody offers to "verify" a purely private document without contacting its maker, ask exactly what they will check. The honest answer is consistency and appearance, and one of those is worthless.

Attestation is not verification

A misreading that does real damage, because it converts a weak position into a confident one.

An attestation generally speaks to the act of attesting — that something was signed or produced before the person attesting, or that a copy corresponds to what was shown to them. It does not speak to whether the contents are true, or whether what was shown was genuine.

The same caution applies, in its own way, to legalisation for use abroad: that process attends to the chain of authentication of signatures and seals rather than to the truth of contents, and it is a different subject with its own guide.

If you only have a scan

A smaller loss than it feels, and worth saying because people treat it as disqualifying.

A scan removes exactly one check — appearance — and appearance was the bottom of the ranking. Everything above it is unaffected:

• Consistency works identically. You are comparing contents, not paper.
• A look-up works identically. The reference is the reference.
• Issuer confirmation works identically. The issuer is being asked about an entry, not about your copy.

The old document

An important distinction that people collapse, usually to somebody's unfair disadvantage.

An old document is not more suspicious. It is less checkable. Records get reorganised. Institutions merge, close or rename. The people who could have confirmed something have retired. Systems that hold recent entries do not reach back.

• Practically, "less checkable" has the same effect on your unchecked list as "suspicious" does — the risk is unresolved either way.
• Fairly, it is a completely different statement about the person, and treating the two as one is how perfectly honest people get refused.
• So say which it is. "We could not check this, because the record does not reach that far" is honest. "This could not be verified" implies something it should not.

One more thing about old documents, and it is the part that gets honest people into trouble with institutions rather than with individuals. Where a check comes back empty, the wording of what is reported matters enormously, and the person writing it rarely thinks about that. "We could not check this" and "this could not be verified" describe the same event and land completely differently — the first is a statement about the reach of a check, the second sounds like a statement about the document. Anybody writing up a verification should use the first, and anybody reading one should mentally rewrite the second into it before drawing a conclusion.

Genuine is not sufficient

Now the second half of the subject, which has nothing to do with authenticity and which catches more people than forgery ever does.

Assume every check comes back clean. The document was issued, by the body named, exactly as it appears. It can still be completely useless to you.

Two paimane, not one: real, and sufficient. People spend everything on the first and never state the second.

Superseded and cancelled

The quiet ones, because nothing about the document itself changes when they happen.

• Superseded. A later entry in the same record replaced it — a correction, a fresh issue, an amended particular. The document in your hand is a true copy of a position that is no longer the position.
• Cancelled. The thing it evidenced was withdrawn, revoked or surrendered. The paper is unaffected and entirely genuine.
• Expired. The clearest case and still frequently missed, because a date on a document is read as issue rather than as validity.

Genuine, on false particulars

The case that shows exactly how far authenticity reaches, and it is less far than people assume.

A document can be issued entirely properly, by the right body, following the right process, on the basis of information that was false. The issuer did nothing wrong. The document is genuine. And the fact it asserts is not true.

What does reach it, partially: consistency against independent facts. If a document asserts something that disagrees with a bank record, a timeline, or another institution's paper, the disagreement is visible even though the document is genuine. This is the one place where the third-ranked check outperforms the first.

The real and wrong document

And the commonest failure in this entire subject, which nobody calls a failure because nothing went wrong.

Somebody needed to establish a fact. They were given a document. They verified it, properly, and it was genuine. And it does not establish the fact.

• An address proof that proves an address — at a date years before the one that mattered.
• A certificate that evidences one thing, where the requirement was about an adjacent thing.
• A document in the name of a person, where what was needed was the status of an entity.
• A letter from somebody who was not the person whose confirmation the requirement called for.
• A copy, properly attested, where the requirement needed the record-holder's own issue.

Start from the requirement

The cure, and it is one habit, applied before any checking at all.

People start from the document in their hand, because that is what arrived. The right starting point is the fact you need established, which is somewhere else entirely.

1. Write down the fact. In a sentence. "That this person lived at this address on this date." "That this entity existed on that date." "That this qualification was awarded."
2. Ask who is entitled to establish it. Which body's record is the one that settles it.
3. Then ask whether the document you have — assuming it is entirely genuine — establishes that fact.
4. Only then verify.

The question before the check

Step three above deserves isolating, because it is thirty seconds long and it is skipped essentially always.

• If the answer is no, stop. Verification is wasted effort and you need a different document. This is the cheapest saving available in the whole subject.
• If the answer is "partly", name the part it does not cover. That part belongs on the unchecked list regardless of what any check returns.
• If the answer is yes, now the checking is worth doing — and you know what you are protecting.

Two further habits make that question sharper, and both take seconds.

• Find out who imposed the requirement, and in what words. People work from a summary that has passed through two or three hands, and the summary has usually lost the part that mattered — a date, a form, whose confirmation. Asking for the original wording routinely changes which document is needed.
• Ask what would satisfy them instead. Requirements are frequently stated in terms of one document because that is the usual document, not because it is the only acceptable one. The alternative is often easier to get and easier to verify, which makes this question the best value in the subject.

The unchecked list

The deliverable, and the reason this page exists in the shape it does.

And the reason it is the output rather than a footnote to the output: you cannot act on "verified". You can act on "the issuer confirmed the award; the record shows it current; the name differs by an initial across two documents and is unexplained; and whether the particulars given to the institution were correct is not checkable by anybody."

The three answers, per item

What a verification should say about each thing that matters. Three possible answers, and exactly one of them is usually missing from what people are handed.

If what you are handed instead is a document saying "verified", you have been given reassurance. That is the one product in this subject which is always available and never worth anything.

Why something cannot be checked

The reasons, stated plainly, because "unchecked" needs its cause attached or it reads as an accusation.

• The issuer does not respond to third parties. Common and entirely proper.
• The record is accessible only to the holder. So the check exists but is not yours to run.
• There is no register behind the document at all. True of most private documents.
• The assertion is not a matter of record anywhere. Nothing to check against, at any price.
• Confirming it needs the consent of the person who gave it to you. Which they may give, and may not.
• The record does not reach that far back. The old document problem.

Pricing your own risk

What the unchecked list is actually for, once you have it.

For each item on it, three options and a cost against each:

1. Remove it. Do the stronger check. Cost: time, usually, and sometimes the subject's consent.
2. Carry it. Decide the exposure is acceptable for what you are relying on the document for. Cost: the exposure, knowingly held.
3. Ask for a different document. Frequently the cheapest route, and almost never considered — a document with a record behind it instead of one without.

If you find something

Stop, and be careful about the next move, because this is where people create problems for themselves.

1. Write down exactly what you found, factually, before it becomes a conclusion in your memory.
2. Do not rely on the document meanwhile. That part is straightforward.
3. Ask about it, neutrally, of the person who gave it to you — and note that most discrepancies have an explanation that arrives within a sentence.
4. Do not characterise it, to them or to anybody else, until you know.

A discrepancy is not dishonesty

Worth its own section because the inference is automatic and usually wrong.

What you have found is a reason to ask. It is not a finding about a person, and treating it as one is unfair where you are wrong and only accidentally right where you are not.

What not to do next

Four things, and the last is the one that most often turns a checking exercise into a dispute.

• Do not accuse. Ask. The difference costs you nothing and is the whole difference.
• Do not tell third parties what you think you have found. If you are wrong, that is your problem; if you are right, it is still not yours to publish.
• Do not confront with "proof" you do not have. A discrepancy is not proof and presenting it as one ends the conversation in which the explanation would have arrived.
• Do not decide what to report, or whether to. Whether anything should be reported, to whom, and what exposure attaches either way is an advocate's question on your facts, and worth asking quickly rather than acting on instinct. It is not ours and we will not advise on it.

When your document is the one checked

This page has two readers, and the second half of the audience is the person whose paper is being examined. Almost everything above reads differently from that side, and the practical advice inverts.

Which makes the passive path the dangerous one: do nothing, answer only what is asked, and let the person checking spend three weeks on the weak band forming an impression.

Disclose it first

The single most valuable thing on this page for the person being checked.

So before handing anything over:

1. Read your own set the way a checker would — side by side, looking for things that do not agree. You will find what they would find.
2. Say it first, plainly, with the reason. "My name is spelled two ways across these two documents, because the older one transliterated it differently."
3. Bring the fix if there is one — a correction already made, or a one-and-same-person position already documented. Produced in advance it is an answer; offered afterwards it is an excuse.
4. Give what makes the strong check easy — the reference, the date, the issuing office, and consent where consent is what is blocking it.

There is a second reason to volunteer the strong route, and it is about how impressions form rather than about fairness. Somebody checking a document is making a judgement under uncertainty, and they will resolve the uncertainty one way or another — if not with a record, then with an impression. Impressions are built from exactly the material this page calls weakest: how the paper looked, how long the answers took, whether you seemed comfortable. You cannot prevent that process; you can make it unnecessary by handing them something that settles the question.

And one piece of cold comfort worth having: if your document is old or from a body that no longer responds, the honest result is going to be "unchecked", and that is not a verdict on you. The useful move there is to say so yourself, early, in the terms this page sets out — less checkable, not more suspicious — and to offer whatever independent facts are checkable instead. A person who explains their own unchecked list is in a far better position than one who leaves somebody else to discover it and name it.

What not to hand over

Helping is not the same as emptying your file, and two limits are worth holding.

• Give what is needed, not everything. A checker with your entire file will check your entire file, including things nobody required, and every additional document is an additional chance of an unexplained inconsistency about something irrelevant.
• Be careful about originals leaving your hands. A clear copy plus the ability to produce the original for inspection is almost always sufficient. Somebody who insists on retaining your originals in order to verify them is asking for something verification does not require — and the risks of that are a separate subject with its own rules.

Documents with their own route

Some things are not a general document check at all; they are a defined process with a defined counterparty, and treating them as a quick check wastes the time.

When checking is not worth it

Said plainly, because we would rather lose the order than take it for nothing.

• A purely private document with no record behind it. There is nothing we can check that you cannot, by asking its maker. Save the fee and make the call.
• A single document, where you are the only source of comparison. Consistency needs more than one paper. If there is one document and no other facts, the strong checks are either available to you directly or not available at all.
• Where the answer to the question before the check is no. Do not verify a document that would not help you even if genuine.
• Where you are going to proceed regardless. If no result would change your decision, you are buying reassurance. We would rather tell you that.

Who brings us a document

Four callers, and each needs a different first question.

1. "Is this real?" The commonest, and the first thing we do is reframe it — what fact do you need, and what would each available check rule out.
2. "Something about this feels off." Useful instinct, weak evidence. We move it off appearance and onto consistency, which is usually where the feeling came from anyway.
3. "I have five documents and they do not seem to agree." The best case for this service, because consistency is the check we can actually run well.
4. "My documents are being verified and I am worried." The other side of the page, and the work is disclosing first.

What we ask before we start

• What fact you need established. Asked first, because it occasionally ends the matter.
• Who is asking you for it, and in what terms, if anybody is.
• Whether the document is public or private — which decides whether any strong check exists.
• What else you have. Every additional document makes the consistency work more powerful.
• Whether you have the original, and whether the subject's consent is available if a strong check needs it.
• What you will do with the answer. If nothing would change, we will say so.

What we check, and how

1. The requirement first — whether this document, if genuine, does the job at all.
2. Classification — public or private, and therefore which checks exist.
3. Any available look-up, run against the issuer's own record where there is one, and reported with its reach stated.
4. The consistency work — your documents read against each other and against the facts you give us, which is where most findings come from.
5. Currency — not only whether it was issued, but whether it is still the position in that record, where that is checkable.
6. What an issuer enquiry would require, where one is the only remaining route — who, how long, and whether consent is needed.

What you get from us

A written note, per item, in three columns — and the third is the point:

• Checked and consistent, with which check and what its reach was.
• Checked and inconsistent, stated factually, with no characterisation of anybody.
• ★ Unchecked, with the reason, and what it would take to check it.

Plus a digitised set of exactly what you gave us, so the position as at today is on record — which matters if anybody later asks what you had in front of you when you decided.

What we cannot certify

The first and last cost us work, and the sixth is the one that distinguishes this from most of what is sold under the word "verification".

Our fee on a check

Our part for a document check — the requirement question, the classification, any available look-up, the consistency work across everything you give us, the currency question, and the written three-column note ending in the unchecked list — is ₹399, with a turnaround of Same day.

That turnaround covers what can be done quickly plus the honest statement of what cannot. Where a check depends on an issuer replying, that is not a same-day matter and we will say so at the start rather than let the stated time imply otherwise.

What is separate:

• An issuer enquiry, where that is the route — for a qualification, degree verification; for a marriage, marriage verification.
• A police process, where antecedents are the question — police clearance or a character certificate.
• A full file read rather than a document check — scrutiny, or both sets read against each other.
• Correcting a mismatch that the check surfaces in your own documents — that work.
• Legalisation for use abroad, which is a different thing entirely — apostille assistance.
• Collecting or returning originals — recorded pickup.
• An advocate's fee, wherever the question becomes what to do about what was found.

And the framing. This is a small service and its value is concentrated in two things nobody charges separately for — asking what you actually need established, and writing down what remains unknown. The checking in the middle is sometimes substantial and sometimes, honestly, unavailable. If your document is private with no record behind it, we will tell you there is nothing here for us to do, and if you were going to proceed whatever we found, we will tell you that too. A service that returns "verified" on every document it is given is selling the one thing in this subject that is always available and never worth anything.

Court work is for your advocate, whose fee is engaged and paid by you directly; we do not quote, collect or share it.

Questions people actually ask