One Claim A Judge Cannot Believe Costs You The Whole Award Entry
By Gabriel Tan | June 2026
I judged this year's International Business Awards, in artificial intelligence (AI), technology and sustainability. Across the entries, the same line kept separating the winners from the rest. This is the third of three posts on where that line falls, drawn from what I actually scored. The first post was about proof. The second was about how you write it down. This one is about the difference between a claim a judge believes and one they do not.
You made a bold claim in your entry. The biggest one you had. You wanted it to land, so you led with it, and you reached a little past what you could actually prove.
The judge reads it and does not believe it. Not because they know your business, but because they have read enough entries to know the shape of a claim that has been stretched.
Here is the cost. They do not just set that one claim aside. They go back through the rest of your entry reading every other number with the same doubt, waiting to catch you again. One claim they cannot believe taxes every claim that follows, including the true ones.
A judge gives belief by default and takes it away once. Three things spend it fastest. Claiming a result that has not happened yet. Putting up numbers nobody outside your company can check. And selling once the proof runs out.
A plan is not a result
The clearest way to lose a judge is to describe a future as though it were already true. The entry talks about the impact the product will have, the savings it is expected to deliver, the markets it is about to enter. Read closely, none of it has happened.
I scored an entry from a company founded after the start of the award period it was entering. The product had launched weeks before the submission. There were demo links and bold projections, and not one customer result, because there had not been time for one. The work may turn out well. It had no record to judge.
A young result is the same problem in a smaller form. A tool six months live, written up as proven, lasting change. Six months shows speed. It does not show that the change holds, and a judge knows the difference. Claim what has already happened. If the strong part is still ahead of you, the entry is a year early, and saying so honestly costs you less than a projection dressed as a fact.
A number with no source is a guess
Strong entries put up strong numbers. The judge's next question is always the same. Says who.
If the only source for a figure is the company claiming it, the judge has to take it on trust, and trust is the thing you are trying to earn, not assume. A self-reported number with no outside check sits one notch below the same number confirmed by an auditor, a regulator, a named third party, or a published benchmark. Same figure, very different weight.
The worse version is the number you cannot see at all. I have read entries that pointed to a supporting document for the real proof, where the document was never attached. The evidence existed only as a promise, and a judge cannot score a promise. Name a source a judge could check for every figure that matters. Where there is no outside source, write self-reported plainly. An honest label holds up better than a number that looks too clean to question.
When the facts stop, the adjectives start
This is the tell a judge learns to read. An entry opens with figures and specifics, then somewhere in the second half it changes register. The numbers thin out and the language swells. World-leading. A new era. A turning point for the industry.
That shift is not decoration. It is a signal that the proof ran out and the entry started selling to cover the gap. A judge reads the swell as an admission. If you had more evidence, you would have led with it, not reached for the big words.
I scored an entry that did exactly this. It made a fair case early, then spent its closing on grand language tied to nothing measured. The grand language did not lift the entry. It pulled down the parts that were real, because it made the judge wonder what else had been inflated. Cut every claim you cannot put a fact behind. End on your strongest proof, not your biggest adjective. A plain true sentence does more for you than a grand empty one.
What to do before you submit
Three checks, before the entry leaves your desk.
Mark every claim that is a projection or a plan rather than something that has already happened. Cut it, or label it honestly as a target. Fifteen minutes.
Take your biggest numbers and write the source beside each one. A third party, an auditor, a regulator, a published figure. Where there is none, write self-reported, and leave it visible. Twenty minutes.
Read your final third on its own. Mark every sentence that reaches for a grand word instead of a fact. Delete it, and end on your strongest proof. Fifteen minutes.
The entry a judge believes all the way down
Go back to that one stretched claim, and the doubt it spread over everything after it. The problem was never that the claim was ambitious. It was that it made the judge stop trusting the entry, including the parts that were true.
Now picture the other entry. Every claim is something that happened. Every number has a source. The language stays plain to the last line. The judge reads it once, believes it all, and scores it on the work, because nothing in it asked them to doubt.
That is the line between a claim a judge believes and one they do not. Stay on the right side of it, and the rest of the entry is yours to win on.
Gabriel Tan is the founder of Mekong Bridge Advisory. He builds structured execution systems for PR and communications firms.
