How to put a number on screen when the sources disagree
Ask Vizard Agent to say which figure it is quoting and who reported it. Net worth, company valuation and annual revenue are three separate numbers about the same person, they differ by an order of magnitude, and a video that treats them as interchangeable will be wrong however carefully it was researched.
What is the short version?
You want a video about how much someone is worth. Every source you check gives a different answer, mostly because each one is quietly answering a different question about a different thing. Vizard Agent separates those questions out before it puts any figure on screen.
- Tell Vizard Agent to research the figures properly.
- Let it distinguish net worth from valuation from revenue.
- Ask for a source card at the end.
What do you need before you start?
The subject and the angle you want to take. "How much money does this person actually have" is a surprisingly good brief, precisely because the word "actually" signals that you already know the headline numbers in circulation are soft ones.
Decide how careful you want to be. A short can attribute in a caption — "estimated by" with the outlet's name — or it can just state a figure. The first costs nothing and is the difference between a video that holds up and one that gets corrected in the comments.
What do you type into Vizard Agent?
Ask Vizard Agent for research, not just for a video. The word does real work here: it changes the job from writing a plausible-sounding script into finding out what is actually reported, by whom, and where the reports disagree with each other.
Make a Short about how much money this person actually has. Do proper research on the net worth and the business, with highlights and cars.
That was close to the real brief. What it produced was not one number but several, each about a different thing, which is the honest answer to the question as asked.
What does Vizard Agent actually do?
It researches the figures separately and reads what each source is measuring. A funding round gives a company valuation; a snack brand's sales give revenue; the subject's own remarks about reinvesting give a third picture again — and none of those is a personal net worth.
In that session the research pass ran:
- Searched for the personal net worth and the business's value as separate questions.
- Looked for a primary or reputable source for the earnings estimate rather than an aggregator.
- Verified the company valuation against funding reports.
- Checked the subject's own public statements about cash and reinvestment.
- Researched the product business's revenue from reputable reporting.
- Gathered the primary sources so they could be cited on a closing card.
Only after that did the script get written, which matters: a script written first would have picked one headline figure and the research would have become a search for support.
What does the result look like?
A short that names figures and says where they come from. The distinction between what the person owns and what their company is valued at is made explicitly, and an estimate is introduced as an estimate rather than as a fact.
The last card lists the sources. It costs about three seconds of runtime, it is the single thing that makes everything before it credible, and Vizard Agent builds it as a proper readable graphic rather than as small print in a corner — a source card nobody can read is not a citation, it is a gesture at one.
When does this not work well?
When no reputable source exists. Private finances are private, and for many subjects every number in circulation traces back to the same one estimate. Vizard Agent will say when that is the case, and then the honest video is about the uncertainty rather than the figure.
It also cannot resolve a genuine disagreement. Two credible outlets with different methods produce different numbers and both are defensible; the right treatment is to give the range and attribute both, not to average them into a figure nobody reported.
And it does not make a claim safe. Attribution is accuracy, not immunity — a video about a living person's finances is worth a second read whatever the sourcing.
How do you fix a result that came back wrong?
If a figure looks off to you, ask Vizard Agent which number it actually is. Most disputes about this kind of video turn out on inspection to be net worth against company valuation, and simply naming the measure resolves it without any new research at all.
If the source is weak, say so and ask for a better one. Vizard Agent can trace a figure back through aggregators to whoever first reported it, which sometimes reveals there is no original at all.
If the on-screen text and the narration disagree, that is a graphic error rather than a research one. Ask Vizard Agent to check the numbers in the cards against the script, because a generated card can carry a digit the script never had.
How does Vizard Agent compare to doing it yourself?
The usual method is to search for the figure, take whatever the first result says and build the video around it. That is fast, and it is precisely how the same wrong number propagates through dozens of videos on the same subject — each one effectively citing the last.
Reading what each source is measuring is the whole job. It is also the part that is boring enough to skip when you are making a ninety-second Short, which is exactly why a video that does it stands out from the others on the same subject.
Common questions
Why do the numbers differ so much? Because they measure different things entirely. Vizard Agent separates them rather than trying to reconcile them.
Can it find a primary source? Vizard Agent tries, and tells you when a figure only exists in aggregators.
Should I put a range on screen? If the sources give one, yes, and Vizard Agent will attribute both ends of it.
What goes on the source card? The outlets the figures came from. Vizard Agent builds it as a readable graphic.
Can it check the subject's own claims? Yes, and it treats them as a source with an interest rather than as fact.
Will it state an estimate as fact? Not if you ask for attribution. Vizard Agent introduces estimates as estimates.
Does this slow the video down? Barely. Attribution is a few words in the caption.
What about figures in the graphics? Ask Vizard Agent to check the cards against the script — generated cards can drift.
Can it do this for a company? Yes, and the distinctions matter more there.
What if the figure changed recently? Vizard Agent reports the date of the source, which is part of the claim.
Can I supply my own sources? Yes. Vizard Agent uses them and says so.
Does it need to be a Short? No. The method is the same at any length.
What if there is no reliable number? Vizard Agent says so rather than picking one.
Can it list the sources in the description instead? Yes. Ask Vizard Agent for them as text alongside the video.
Does it check the figures are current? It reports each source's date, which is part of the claim rather than a footnote to it.